2 Towards an Agent-based Negotiation Scheme for Scheduling Electric Vehicle Charging 2 13th European Conference on Multi-Agent Systems 2015 15 We consider the problem of scheduling Electric Vehicle (EV) charging within a single charging station aiming to maximize the number of charged EVs, as well as the amount of charged energy. In so doing, we propose one online optimal solution using Mixed Integer Programming (MIP) techniques, and two online solutions which incrementally execute the MIP algorithm each time an EV arrives to the charging station. More- over, we apply agent based negotiation techniques between the station and the EVs in order to service EVs when the MIP problem is initially unsolvable due to insaficient resources (i.e., requested energy, charging time window). We evaluate our solutions in a setting partially using real data, and we show that when applying negotiation techniques, the num- ber of EVs charged increases on average by 7%, energy utilization by 6:5%, while there is only a small deficit (about 10%) on average agent utility which is unavoidable due to the fact that the initial incremental demand-response problem is unsolvable. EVsSeitaridisFinal.pdf Athens,Greece A Heuristic for Planning based on Action Evaluation 2 Proc. 10th International Conference on Artificial Intelligence: Methodology, Systems, Applications (AIMSA '02) Springer-Verlag LNAI 2443 2002 10 61-70 This paper proposes a domain independent heuristic for state space planning, which is based on action evaluation. The heuristic obtains estimates for the cost of applying each action of the domain by performing a forward search in a relaxed version of the initial problem. The estimates for the actions are then utilized in a backward search on the original problem. The heuristic, which has been further refined by a goal-ordering technique, has been implemented in AcE (Action Evaluation), a state space heuristic planner, and thoroughly tested on a variety of toy problems. aimsa.pdf D. Vrakas and I. Vlahavas, "A Heuristic for Planning based on Action Evaluation", Proc. 10th International Conference on Artificial Intelligence: Methodology, Systems, Applications, AIMSA 2002, Varna, Bulgaria, September 2002, Springer-Verlag, LNAI, Vol. 2443, pp. 61-70. On the Discovery of Weak Periodicities in Large Time Series 2 Proc. 6th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD '02) Springer-Verlag LNAI 2431 2002 10 51-61 The search for weak periodic signals in time series data is an active topic of research. Given the fact that rarely a real world dataset is perfectly periodic, this paper approaches this problem in terms of data mining, trying to discover weak periodic signals in time series databases, when no period length is known in advance. In existing time series mining algorithms, the period length is user-specified. We propose an algorithm for finding approximate periodicities in large time series data, utilizing autocorrelation function and FFT. This algorithm is an extension to the partial periodicity detection algorithm pre-sented in a previous paper of ours. We provide some mathematical background as well as experimental results. Berberidis-PKDD02-Procs.pdf C. Berberidis, I. Vlahavas, W. G. Aref, M. Atallah and A. K. Elmagarmid, "On the Discovery of Weak Periodicities in Large Time Series", Proc. 6th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD'02), Helsinki, Finland, August 2002, Springer-Verlag, LNAI, vol. 2431, pp.51-61. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Fchristos%5Fpersonal%2FBerberidis%2DPKDD02%2DProcs%2Epdf eLPA: An e-Learner's Personal Assistant 2 Proc. ICCS'02 Workshop on Applications of Conceptual Graphs (electronic proceedings) 2002 6 Intelligent hypertext is a promising approach to information systems, because it combines the power of inference of expert systems and the intuitive power of hypertext. In this paper we propose the "COMFRESH", a common framework for expert systems and hypertext. It is based on a Prolog interpreter and uses the conceptual graph knowledge representation formalism for browsing and reasoning. COMFRESH can be used as a knowledge based hypertext (intelligent hypertext) or as an expert system with hypertext capabilities. kokkoras-elpa-iccs2002.pdf F. Kokkoras and I. Vlahavas, "eLPA: An e-Learner’s Personal Assistant", Proc. Workshop on Applications of Conceptual Graphs, ICCS2002, Borovets, Bulgaria, July 2002. Fdfd Borovets, Bulgaria hypertext information retrieval conceptual graphs An Educational Metadata Management System using a deductive object-oriented database approach 2 Proc. ED-MEDIA World Conference on Educational Multimedia, Hypermedia and Telecommunications 2002 0 1724-1728 The traditional use of large search engines for retrieving educational information on the Internet is rather inaccurate and provides irrelevant search results in most cases. Recently metadata are widely used to semantically describe educational resources and a number of educational metadata specifications have been proposed aiming at defining the set of elements that can better describe an educational resource. This paper proposes an architecture for an educational metadata management system, which facilitates both metadata storage and data retrieval, by using a deductive object-oriented database. The proposed system provides not only tolls for creation, validation and modification of educational metadata documents, but also an efficient information retrieval mechanism based on user queries. edmedia2002.pdf D. Sampson,V. Papaioannou, N. Bassiliades, I. Vlahavas, "An Educational Metadata Management System using a deductive object-oriented database approach", accepted for presentation at ED-MEDIA 2002 - World Conference on Educational Multimedia, Hypermedia and Telecommunications, Denver Colorado, USA, June 24-29, 2002. Distributed Data Mining of Large Classifier Ensembles 2 Proc. (companion volume) 2nd Hellenic Conference on AI (SETN '02) I. Vlahavas, C. Spyropoulos 2002 6 249-255 Nowadays, classifier ensembles are often used for distributed data mining in order to discover knowledge from inherently distributed information sources and scale up learning to very large databases. One of the most successful methods used for combining multiple classifiers is Stacking. However, this method suffers from very high computational cost in the case of large number of distributed nodes. This paper presents a new classifier combination strategy that scales up efficiently and achieves both high predictive accuracy and tractability of problems with high complexity. It induces a global model by learning from the averages of the local classifiers' output. This way, fast and effective combination of large number of classifiers is achieved. tsoumakas-setn02.pdf G. Tsoumakas and I. Vlahavas, "Distributed Data Mining of Large Classifier Ensembles", Proc. Companion Volume, 2nd Hellenic Conference on AI, SETN 2002, pp 249-255, Thessaloniki, April 2002. Thessaloniki, Greece http%3A%2F%2Fwww%2Ecsd%2Eauth%2Egr%2F%7Esetn02%2Fposter%5Fpapers%2Findex%2Ehtml CSPCONS: A Communicating Sequential Prolog with Constraints 2 Proc. 2nd Hellenic Conference on AI (SETN '02) Springer-Verlag LNAI 2308 2002 12 72-84 I. Vlahavas, I. Sakellariou, I. Futo, Z. Pasztor, J. Szeredi: "CSPCONS: A Communicating Sequential Prolog with Constraints", Proc. 2nd Hellenic Conference on AI, SETN 2002, Thessaloniki, Greece, April 2002, Springer-Verlag, LNAI, vol. 2308, pp. 72-84. Effective Stacking of Distributed Classifiers 2 Proc. 15th European Conference on Artificial Intelligence (ECAI '02) IOS Press Frank van Harmelen 2002 4 340-344 One of the most promising lines of research towards discovering global predictive models from physically distributed data sets is local learning and model integration. Local learning avoids moving raw data around the distributed nodes and minimizes communication, coordination and synchronization cost. However, the integration of local models is not a straightforward process. Majority Voting is a simple solution that works well in some domains, but it does not always offer the best predictive performance. Stacking on the other hand, offers flexibility in modelling, but brings along the problem of how to train on sufficient and at the same time independent data without the cost of moving raw data around the distributed nodes. In addition, the scalability of Stacking with respect to the number of distributed nodes is another important issue that has not yet been substantially investigated. This paper presents a framework for constructing a global predictive model from local classifiers that does not require moving raw data around, achieves high predictive accuracy and scales up efficiently with respect to large numbers of distributed data sets. tsoumakas-ecai02.pdf G. Tsoumakas and I. Vlahavas, "Effective Stacking of Distributed Classifiers", Proc. 15th European Conference on Artificial Intelligence,ECAI 2002, Lyon, France, IOS Press, pp. 340-344 Multiple and Partial Periodicity Mining in Time Series Databases 2 Proc. 15th European Conference on Artificial Intelligence (ECAI '02) IOS Press Frank Van Harmelen 2002 4 370-374 Periodicity search in time series is a problem that has been investigated by mathematicians in various areas, such as statistics, economics, and digital signal processing. For large databases of time series data, scalability becomes an issue that traditional techniques fail to address. In existing time series mining algorithms for detecting periodic patterns, the period length is user-specified. This is a drawback especially for datasets where no period length is known in advance. We propose an algorithm that extracts a set of candidate periods featured in a time series that satisfy a minimum confidence threshold, by utilizing the autocorrelation function and FFT as a filter. We provide some mathematical background as well as experimental results. Berberidis-ECAI02-Proceedings.pdf C. Berberidis, A. G. Walid, M. Atallah, I. Vlahavas and A. K. Elmagarmid, "Multiple and Partial Periodicity Mining in Time Series Databases", Proc. 15th European Conference on Artificial Intelligence, ECAI 2002, pp. 370-374, Lyon, France, 2002, IOS Press, pp.370-374. Lyon, France The MO-GRT System: Heuristic Planning with Multiple Criteria 2 Proc. AIPS 2002 Workshop on Planning and Scheduling with Multiple Criteria 2002 0 I. Refanidis and I. Vlahavas, "The MO-GRT System: Heuristic Planning with Multiple Criteria", Proc. AIPS 2002 Workshop on Planning and Scheduling with Multiple Criteria, Toulouse, France, April 23, 2002. Multiobjective Heuristic State-Space Planning 1 Artificial Intelligence Elsevier Vol. 145, Issues 1-2 2003 0 1-32 I. Refanidis and I. Vlahavas, "Multiobjective Heuristic State-Space Planning", Artificial Intelligence, Volume 145, Issues 1-2, April 2003, Pages 1-32, Elsevier. Smart Video Text: A Conceptual Graph based Video Data Model 1 ACM Multimedia Systems Journal ACM - Springer Vol. 8, No 4 2002 10 328-338 An intelligent annotation-based video data model called Smart VideoText is introduced. It utilizes the Conceptual Graph knowledge representation formalism to capture the semantic associations among the concepts described in text annotations of the video data. The aim is to achieve more effective query, retrieval and browsing capabilities based on video data’s semantic content. Finally, a generic and modular video database architecture based on Smart VideoText data model is described. kokkoras-svt.pdf F. Kokkoras, H. Jiang, I. Vlahavas, A. Elmargarmid, E. Houstis and W. Aref, "Smart Video Text: A Conceptual Graph based Video Data Model", .ACM Multimedia Systems Journal, vol. 8, pp. 328-338, 2002. http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Flink%2Easp%3Fid%3D8te7u25vnygev4vp Video Databases Conceptual Graphs Content based Video Retrieval MACLP: Multi Agent Constraint Logic Programming 1 Information Sciences Elsevier Vol. 144 (1-4) 2002 15 127-142 I. Vlahavas, "MACLP: Multi agent constraint logic programming", Information Sciences, Elsevier, Vol. 144(1-4), pp. 127-142, 2002. ExperNet: An Intelligent Multi-Agent System for WAN Managemen 1 IEEE Intelligent Systems Vol. 17, No. 1 2002 10 62-72 This paper describes ExperNet, an intelligent multi-agent system that was developed under an EU funded project to assist in the management of a large-scale data network. ExperNet assists network operators at various nodes of a WAN to detect and diagnose hardware failures and network traffic problems and suggests the most feasible solution, through a web-based interface. ExperNet is composed by intelligent agents, capable of both local problem solving and social interaction among them for coordinating problem diagnosis and repair. The current network state is captured and maintained by conventional network management and monitoring software components, which have been smoothly integrated into the system through sophisticated information exchange interfaces. For the implementation of the agents, a distributed Prolog system enhanced with networking facilities was developed. The agent's knowledge base is developed in an extensible and reactive knowledge base system capable of handling multiple types of knowledge representation. ExperNet has been developed, installed and tested successfully in an experimental network zone of Ukraine. ieee-intelligent.pdf I.Vlahavas, N. Bassiliades, I. Sakellariou, et al., “ExperNet: An Intelligent Multi-Agent System for WAN Management”, IEEE Intelligent Systems, Vol. 17, No. 1, pp. 62-72, 2002. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fprojects%2Finco%2FInco%2Ehtml ExperNet%3A+A+Distributed+Expert+System+for+the+Management+of+a+National+Network http%3A%2F%2Fwww%2Ecomputer%2Eorg%2Fintelligent%2Fex2002%2Fx1062abs%2Ehtm Distributed Artificial Intelligence Expert System Intelligent Agents Network Management Active Knowledge Base System Distributed Prolog Intelligent Querying of Web Documents Using a Deductive XML Repository 2 Proc. 2nd Hellenic Conference on AI (SETN '02) Springer-Verlag I. Vlahavas, C. Spyrolpoulos LNAI 2308 2002 11 437-448 In this paper, we present a deductive object-oriented database system, called X DEVICE, which is used as a repository for XML documents. X DEVICE employs a powerful rule-based query language for intelligently querying stored Web documents and data and publishing the results. XML documents are stored into the OODB by automatically mapping the DTD to an object schema. XML elements are treated either as objects or attributes based on their complexity, without loosing the relative order of elements in the original document. The rule-based language features second-order logic syntax, generalized path and ordering expressions, which greatly facilitate the querying of recursive, tree-structured XML data and the construction of XML trees as query results. All the extended features of the rule language are translated through the use of object metadata into a set of first-order deductive rules that are efficiently executed against the object database using the system’s basic inference engine. setn02-bassiliades.pdf N. Bassiliades and I. Vlahavas, "Intelligent Querying of Web Documents Using a Deductive XML Repository", Proc. 2nd Hellenic Conference on AI, SETN 2002, Thessaloniki, Greece, April 2002, Springer-Verlag, LNAI, vol. 2308, pp. 437-448. http%3A%2F%2Fwww%2Ecsd%2Eauth%2Egr%2F%7Elpis%2Fsystems%2Fx%2Ddevice%2Ehtml X%2DDEVICE http%3A%2F%2Flink%2Espringer%2Ede%2Flink%2Fservice%2Fseries%2F0558%2Ftocs%2Ft2308%2Ehtm XML Object Database Deductive Database Educating ICT Tutors: Conclusions from the Operation of ICT Tutor's School at Aristotle University of Thessaloniki 2 Proc. Panhellenic Conference: Connecting Tertiary to Secondary Education 2001 0 S. Demetriadis, G. Palaigeorgiou, A. Barbas, A. Molohides, I. Tsoukalas, D. Psillos and I. Vlahavas, "Educating ICT Tutors: Conclusions from the Operation of ICT Tutor's School at Aristotle University of Thessaloniki", (in Greek), Proc. Panhellenic Conference: Connecting Tertiary to Secondary Education, Thessaloniki, April 2001. Intelligent Industrial Applications Using Constraint Logic Programming 2 Education of Informaticians and Industrial Mathematicians: New Challenges and Needs 2001 23 193-216 I. Vlahavas, "Intelligent Industrial Applications Using Constraint Logic Programming", Proc. Workshop for "Education of Informaticians and Industrial Mathematicians: New Challenges and Needs", Ohrid, FYROM, pp. 193-216, 2001. The GRT Planner 1 AI Magazine Vol. 22 (3) 2001 2 63-65 I. Refanidis and I. Vlahavas, "The GRT Planner", AI Magazine, Vol. 22(3), pp. 63-65, 2001. The GRT Planning System: Backward heuristic construction in Forward State - Space Planning 1 Journal of Artificial Intelligence Research vol. 15 2001 46 115-161 I. Refanidis and I. Vlahavas, "The GRT Planner: Backward Heuristic Construction in Forward State-Space Planning", Journal of Artificial Intelligence Research, vol.15, pp. 115-161, 2001. An Open Learning Environment for Thermal Phenomena 2 Proc. 5th International Conference on Computer Based Learning in Science (CBLIS '01) 2001 0 E. Hatzikraniotis, I. Lefkos, G. Bisdikian, D. Psillos, I. Refanidis and I. Vlahavas, "An open learning environment for thermal phenomena", 5th International Conference on Computer Based Learning in Science (CBLIS 2001). Brno, Czech Republic, July 2001. Parallel Planning via the Distribution of Operators 1 Journal of Experimental and Theoretical Artificial Intelligence Vol. 13 (3) 2001 15 211-226 This paper describes ODMP (Operator Distribution Method for Parallel Planning), a parallelization method for efficient heuristic planning. The method innovates in that it parallelizes the application of the available operators to the current state and the evaluation of the successor states using the heuristic function. In order to achieve better load balancing and a lift in the scalability of the algorithm, the operator set is initially enlarged, by grounding the first argument of each operator. Additional load balancing is achieved through the reordering of the operator set, based on the expected amount of imposed work. ODMP is effective for heuristic planners, but it can be applied to planners that embody other search strategies as well. It has been applied to GRT, a domain–independent heuristic planner, and CL, a heuristic planner for simple Logistics problems, and has been thoroughly tested on a set of Logistics problems adopted from the AIPS-98 planning competition, giving quite promising results. php.pdf D. Vrakas, I. Refanidis and I. Vlahavas, "Parallel Planning via the Distribution of Operators", Journal of Experimental and Theoretical Artificial Intelligence, Vol. 13 (3), pp. 211-226, Jul-Sep 2001. Fuzzy Meta-Learning: Preliminary Results 2 Proc. UK Workshop on Computational Intelligence (UKCI '01) 2001 5 43-48 G. Tsoumakas and I. Vlahavas, "Fuzzy Meta-Learning: Preliminary Results", Proc. UK Workshop on Computational Intelligence, pp. 43-48, Edinburgh, 2001. Combining Progression and Regression in State-Space Heuristic Planning 2 Proc. 6th European Conference on Planning (ECP '01) 2001 12 One of the most promising trends in Domain Independent AI Planning, nowadays, is state-space heuristic planning. The planners of this category construct general but efficient heuristic functions, which are used as a guide to traverse the state space either in a forward or in a backward direction. Although specific problems may favor one or the other direction, there is no clear evidence why any of them should be generally preferred. This paper proposes a hybrid search strategy that combines search in both directions. The search begins from the Initial State in a forward direction and proceeds with a weighted A* search until no further improving states can be found. At that point, the algorithm changes direction and starts regressing the Goals trying to reach the best state found at the previous search. The direction of the search may change several times before a solution can be found. Two domain-independent heuristic functions based on ASP/HSP planners enhanced with a Goal Ordering technique have been implemented. The whole bidirectional planning system, named BP, was tested on a variety of problems adopted from the recent AIPS-00 planning competition with quite promising results. The paper also discusses the subject of domain analysis for state-space planning and proposes two methods for the elimination of redundant information from the problem definition and for the identification of independent sub-problems. bp.pdf D. Vrakas and I. Vlahavas, "Combining Progression and Regression in State-Space Heuristic Planning", Presented at 6th European Conference on Planning, Toledo, Spain, Sep 2001. Bi-Directional Heuristic Planning in State-Spaces 2 Proc. 8th Panhellenic Conference on Informatics Vol. 1 2001 0 511-520 One of the most promising trends in Domain Independent AI Planning, nowadays, is state – space heuristic planning. The planners of this category construct general but efficient heuristic functions, which are used as a guide to traverse the state space either in a forward or a backward direction. Although specific problems may favor one or the other direction, there is no clear evidence why any of them should be generally preferred. This paper proposes a hybrid search strategy that combines search in both directions. The search begins from the Initial State in a forward direction and proceeds with a weighted A* search until no further improving states can be found. At that point, the algorithm changes direction and starts regressing the Goals trying to reach the best state found at the previous step. The direction of the search may change several times before a solution can be found. Two domain-independent heuristic functions based on ASP/HSP planners enhanced with a Goal Ordering technique have been implemented. The whole bi-directional planning system, named BP, was tested on a variety of problems adopted from the recent AIPS-00 planning competition with quite promising results. bp-pci.pdf D. Vrakas and I. Vlahavas, "Bi-Directional Heuristic Planning in State-Spaces", Proc. 8th Panhellenic Conference on Informatics, Nicosia, Cyprus, November 2001. CG-PerLS: Conceptual Graphs for Personalized Learning Systems 2 Proc. 8th Panhellenic Conference on Informatics Livani Publishing Organization Vol. 2 2001 10 531-540 Two of the most important standardization efforts for e-learning technologies are related to the definition of metadata describing educational resources and metadata describing the learner's profile. The internal details of systems that utilize these metadata is still an open issue since these efforts are primarily dealing with "what" and not "how". Under the light of these emerging efforts, we present CG-PerLS, a knowledge based approach for organizing and accessing educational resources. CG-PerLS is a model of a WWW portal for learning objects that encodes the learning technologies metadata in the Conceptual Graph knowledge representation formalism, and uses related inference techniques to provide advanced, personalized functionality. CG-PerLS allows learning resource creators to manifest their material, client-side learners to access these resources in a way tailored to their individual profile and educational needs, and dynamic course generation based on fine or coarse grained educational resources. kokkoras-cgperls-epy8.pdf F. Kokkoras, D.G. Sampson and I. Vlahavas, "CG - PerLS: Conceptual Graphs for Personalized Learning Systems", Proc. 8th Panhellenic Conference on Informatics, Nicosia, Cyprus, November2001. Nicosia, Cyprus Educational Metadata Conceptual Graphs e-Learning FUNAGES: An Educational Expert System for Fundus Fluorescein Angiography 1 Health Informatics Journal Vol. 7, No. 3-4 2001 7 214-221 FUNAGES is an expert system that deals with the interpretation of fundus fluorescein angiography. Fluorescein angiography is an extremely valuable clinical test that provides information about the circulatory system of the ocular fundus (the back of the eye) not attainable with a routine examination. The different, in place and time, appearance of fluorescein and the classification of the fundus diseases render angiography a dynamic, cinematographic and deductive diagnostic method. Therefore, the knowledge for interpreting fundus fluorescein angiograms allows an ophthalmologist specialized in ocular fundus diseases to follow a systematic, orderly and logical line of reasoning that leads to a proper diagnosis. FUNAGES was developed to simulate the above logical reasoning, in order to facilitate the inexperienced ophthalmologists in the interpretation of the angiograms. The system achieved its purposes in an adequate way via a graphical user interface and a thorough knowledge base. shimr2001.pdf V. Dimitroula, N. Bassiliades, I. Vlahavas, S. Dimitrakos, "FUNAGES: An Expert System for Fundus Fluorescein Angiography", Health Informatics Journal, vol. 7, pp. 214-221, 2001, (also in Proc. 6th Int. Symposium on Health Information Management Research, pp. 75 – 89, May 2001). Expert Systems Artificial Intelligence in Medicine AI Planning for Transportation Logistics 2 Proc. 17th International Logistics Conference 2001 7 241-248 In the last decade the efficiency of the Artificial Intelligence Planning Systems has been increased significantly. New systems appeared that are able to cope with planning problems being orders of magnitude more complex than the ones solvable in early 90's. This vast improvement increase was made possible mainly by three new approaches in plan generation: planning graphs, satisfiability planning and heuristic state-space planning. The latter approach, which is the most powerful one, derives a heuristic function from the specification of a planning problem, independently of its domain, and uses it for guiding the search through the space of the states. During the last years appeared many heuristic state-space planners, such as ASP, HSP, GRT and FF, which were able to solve large transportation logistics problems, with numerous locations, trucks and objects that have to be transferred, very efficiently, as it has been shown in the recent international planning competitions. This paper briefly presents the current status in domain-independent heuristic state-space planning and concentrates on the GRT and MO-GRT planners, where the latter is a recent extension of GRT being able to consider multiple criteria in the plan generation and evaluation process. Finally, the paper outlines results of running MO-GRT in some transportation logistics problems and poses directions for future research. I. Refanidis, N. Bassiliades, I. Vlahavas, "AI Planning for Transportation Logistics", Logistics from α to Ω: Strategies and Applications, Proc. 17th Int. Logistics Conf., October 2001, pp. 241-248. Thessaloniki, October 2001 Transportation Logistics Artificial Intelligence Planning Heuristic Search Multiple Criteria Knowledge Management and Artificial Intelligence 2 Proc. Workshop on Education of Informaticians and Industrial Mathematicians: New Challenges and Needs 2001 8 31-49 km_ai.pdf N. Bassiliades, I. Vlahavas, F. Kokkoras, "Knowledge Management and Artificial Intelligence", in Proc. Workshop for "Education of Informaticians and Industrial Mathematicians: New Challenges and Needs", Ohrid, FYROM, 2001, pp. 31-49. Land Evaluation: An Artificial Intelligence Approach 3 Environmental Information Systems in Industry and Public Administration IDEA Group Publishing C. Rautenstrauch and S. Patig Ch. 9 2001 8 158-166 G. Tsoumakas and I. Vlahavas, "Land Evaluation: An Artificial Intelligence Approach", in C. Rautenstrauch and S. Patig (Eds.), Environmental Information Systems in Industry and Public Administration, Chapter 9, pp 158-166, Idea Group Publishing, 2001. A Multimedia System for teaching Prolog (in Greek) 2 Proc. 1st National Conference on Education and Informatics 2000 0 I. Vlahavas, I. Refanidis and E. Sakellariou, "A Multimedia System for teaching Prolog", in 1st National Conference on Education and Informatics, Thessaloniki, Greece, November 2000 (in Greek). A virtual laboratory on Heat Phenomena 2 Proc. 5th Workshop on Multimedia in Physics Teaching and Learning 2000 0 E. Hatzikraniotis, G. Bisdikian, I. Refanidis, I. Vlahavas and D. Psillos, "A virtual laboratory on Heat Phenomena", 5th Workshop on Multimedia in Physics Teaching and Learning, Vienna, October 2000. Composite Virtual-Lab Environment for Teaching Heat and Thermodynamics (in Greek) 2 Proc. 2nd National Conference on Information and Communication Technologies in Education 2000 0 D. Psillos, P. Argirakis, I. Vlahavas, E. Hatzikraniotis, I. Refanidis et. all, "Composite Virtual-Lab Environment for teaching Heat and Thermodynamics", in Proc. of 2nd National Conference on Information and Communication Technologies in Education, Patras, Greece, October 2000 (in Greek). Virtual Lab-Environment for Thermal Phenomena (in Greek) 2 Proc. 1st National Conference on Education and Informatics 2000 0 I. Lefkos, I. Refanidis et. all, "Virtual Lab - Environment for thermal Phenomena", in 1st National Conference on Education and Informatics, Thessaloniki, Greece, November 2000 (in Greek). Virtual Lab-Environment for Thermodynamics (in Greek) 2 Proc. 1st National Conference on Education and Informatics 2000 0 I. Refanidis, K. Korobilis et. all, "Virtual Lab-Environment for thermodynamics", in 1st National Conference on Education and Informatics, Thessaloniki, Greece, November 2000 (in Greek). E-DEVICE: An Extensible Knowledge Base System with Multiple Rule Support 1 IEEE Transactions on Knowledge and Data Engineering Vol. 12, No. 5 2000 20 824-844 This paper describes E-DEVICE, an extensible active knowledge base system (KBS) that supports the processing of event-driven, production, and deductive rules into the same active OODB system. E-DEVICE provides the infrastructure for the smooth integration of various declarative rule types, such as production and deductive rules, into an active OODB system that supports low-level event-driven rules only by a) mapping each declarative rule into one event-driven rule, offering centralized rule selection control for correct run-time behavior and conflict resolution, and b) using complex events to map the conditions of declarative rules and monitor the database to incrementally match those conditions. E-DEVICE provides the infrastructure for easily extending the system by adding a) new rule types as subtypes of existing ones and b) transparent optimizations to the rule matching network. The resulting system is a flexible, yet efficient, KBS that gives the user the ability to express knowledge in a variety of high-level forms for advanced problem solving in data intensive applications. TKDE-12.ps.gz N. Bassiliades, I. Vlahavas, and A. Elmagarmid, "E-DEVICE: An Extensible Knowledge Base System with Multiple Rule Support", IEEE Transactions on Knowledge and Data Engineering, Vol. 12 (5), pp. 824-844, 2000. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdevice%2Ehtml DEVICE http%3A%2F%2Fwww%2Ecomputer%2Eorg%2Ftkde%2Ftk2000%2Fk0824abs%2Ehtm Knowledge base system Production rules Deductive rules Derived attributes Aggregation Negation Active object-oriented database Heuristic Planning with Resources 2 Proc. 14th European Conference on AI (ECAI '00) 2000 0 I. Refanidis and I. Vlahavas, "Heuristic Planning with Resources", 14th European Conference on AI (ECAI-2000), Berlin, August 2000. An Operator Distribution Method for Parallel Planning 2 Proc. AAAI-2000 Workshop on Parallel and Distributed Search for Reasoning 2000 4 17-21 This paper presents the Operator Distribution Method for Parallel Planning (ODMP), a parallelization method especially suitable for heuristic planners. ODMP distributes the process of finding and applying the ground applicable actions to a given state, to the set of the available processors. The operator schemas of the domain are distributed to the available processors in a dynamic manner. In order to utilize a larger number of processors and to achieve better load balancing, the set of the domain’s operators is initially expanded by considering all the possible instantiations of their first argument. The proposed method, ODMP, is an effective parallelization method for heuristic planners, but it can also be applied to planners that embody other search strategies as well. We implemented ODMP in a best first planner that uses a domain specific heuristic for logistics problems and tested its efficiency on a variety of problems, adopted from the AIPS-98 planning competition. odmp.pdf D. Vrakas, I. Refanidis and I. Vlahavas, "An Operator Distribution Method for Parallel Planning", Proc. AAAI Workshop on Parallel and Distributed Search for Reasoning, pp. 17-21, Austin, Texas, Jul 2000. An Adaptable Framework for Educational Software Evaluation 3 Recent Developments and Applications in Decision Making Kluwer Academic Publishers S.H. Zanakis, C. Zopounidis and G. Doukidis 2000 0 I. Stamelos, I. Refanidis, P. Katsaros, A. Tsoukias, I. Vlahavas and A. Pombortsis, "An Adaptable Framework for Educational Software Evaluation", in Recent Developments and Applications in Decision Making, ed. S.H. Zanakis, C. Zopounidis, G. Doukidis, Kluwer Academic Publishers (to appear in 2000). Exploiting State Constraints in Heuristic State-Space Planning 2 Proc. 5th International Conference on Artificial Intelligence Planning and Scheduling Systems 2000 0 363-370 I. Refanidis and I. Vlahavas, "Exploiting State Constraints in Heuristic State-Space Planning ", 5th International Conference on Artificial Intelligence Planning and Scheduling Systems (AIPS-2000), Breckenridge, Colorado, USA, April 2000. Knowledge Based Evaluation of Software Systems: A Case Study 1 Information and Software Technology Elsevier Vol. 42(5) 2000 12 333-345 I. Stamelos, I. Vlahavas, I. Refanidis and A. Tsoukias, "Knowledge Based Evaluation of Software Systems: A Case Study", Information and Software Technology, Elsevier, vol. 42(5), pp. 333-345, 2000. Active Knowledge-Based Systems: Techniques and Applications 3 Knowledge Engineering: Systems Techniques and Applications Academic Press C.T. Leondes Vol. 1 2000 36 1-36 This chapter focuses on active knowledge base systems; more specifically, it presents various implementation techniques that are used by the numerous systems found in the literature and on applications made based on such systems. Systems are compared based on the basis of the different techniques and of their efficiency in various applications. Finally, the active object-oriented knowledge base system DEVICE is thoroughly described, giving emphasis to its advantages over similar systems. Furthermore, two applications based on the DEVICE system are described: deductive databases and data warehouses. academic-chapter1.pdf N. Bassiliades and I. Vlahavas, "Active Knowledge-Based Systems", in "Knowledge Based Systems: Techniques and Applications", C.T. Leondes, ed., Vol. 1, pp. 1-36, Academic Press, May 2000. http%3A%2F%2Fwww%2Eapnet%2Ecom%2Fknowledgesystems%2F Knowledge Base Systems Object-Oriented Database Systems Active Database Systems Deductive Databases Data Warehouses On the Parallelization of Greedy Regression Tables 2 Proc. 18th Workshop of the UK Planning and Scheduling Special Interest Group (PlanSIG '99) 1999 9 180-189 This paper presents PGRT, a parallel version of a best first planner based on the Greedy Regression Tables approach. The parallelization method of PGRT distributes the task of extracting applicable actions to a given state among the available processors. Although the number of operators limits the scalability of PGRT, it has proven to be quite efficient for low scale parallelization. A modified Operator Reordering method has been used in order to achieve further increase in the efficiency of the parallel algorithm. We illustrate the speedup of PGRT on a variety of hard logistics problems, adopted from the AIPS-98 planning competition. pgrt.pdf D. Vrakas, I. Refanidis, F. Milcent and I. Vlahavas, "On the Parallelization of Greedy Regression Tables", Proc. 18th Workshop of the UK Planning and Scheduling SIG, Manchester U.K., December, 1999, pp. 180-189. On Determining and Completing Incomplete States in STRIPS Domains 2 Proc. IEEE International Conference on Information, Intelligence and Systems 1999 0 289-296 I. Refanidis, I. Vlahavas and L. Tsoukalas, "On Determining and Completing Incomplete States in STRIPS Domains", IEEE International Conference on Information, Intelligence and Systems, Washington D.C., November 1-3, 1999 (to be presented). GRT: A Domain Independent Heuristic for STRIPS Worlds based on Greedy Regression Tables 2 Proc. 5th European Conference on Planning (ECP '99) Springer-Verlag 1999 12 346-358 I. Refanidis and I. Vlahavas, "GRT: A Domain Independent Heuristic for STRIPS Worlds based on Greedy Regression Tables", 5th European Conference on Planning (ECP-99), Durham, UK, Springer-Verlag, September 8-10, 1999, pp. 346-358. PARCIS: A Robust Parallel VLSI Circuit Simulator 1 Simulation Practice and Theory Journal Elsevier Vol. 7 1999 12 91-103 P.Linardis and I. Vlahavas, "PARCIS: A Robust Parallel VLSI Circuit Simulator", Simulation Practice and Theory Journal, Elsevier, vol. 7, pp. 91-103, 1999. OASys: An AND/OR Parallel Logic Programming System 1 Parallel Computing Elsevier Vol. 25 1999 15 321-336 I.Vlahavas, P. Kefalas and C. Halatsis, "OASys: An AND/OR Parallel Logic Programming System", Parallel Computing, Elsevier, vol. 25, pp. 321-336, 1999. ESSE: An Expert System for Software Evaluation 1 Knowledge-Based Systems Elsevier Vol. 12 (4) 1999 14 183-197 I. Vlahavas, I. Stamelos, I. Refanidis and A. Tsoukias, "ESSE: An Expert System for Software Evaluation", Knowledge-Based Systems, Elsevier, vol. 12(4), pp. 183-197, 1999. Automating the Evaluation of Educational Software 2 Proc. 5th International Conf. of the Decision Sciences Institute (DSI '99) 1999 4 1369-1373 I. Stamelos, I. Refanidis, P. Katsaros, A. Tsoukias, I. Vlahavas and A. Pombortsis, "Automating the Evaluation of Educational Software", 5th International Conf. Of the Decision Sciences Institute, DSI '99, Athens 1999, pp. 1369, 1373. ISLE: An Intelligent System for Land Evaluation 2 Proc. ACAI '99 Workshop on Intelligent Techniques for Spatio-Temporal Data Analysis in Environmental Applications 1999 6 26-32 G. Tsoumakas and I. Vlahavas, "ISLE: An Intelligent System for Land Evaluation", ACAI'99 Workshop on Intelligent Techniques for Spatio-Temporal Data Analysis in Environmental Applications, Chania 1999. Stock Miner: A System for Knowledge Discovery and Financial Data 2 Proc. ACAI '99 Workshop on Data Mining in Economics, Marketing and Finance 1999 0 M. Karamanlidou and I. Vlahavas, "Stock Miner: A System for Knowledge Discovery and Financial Data", ACAI'99 Workshop on Data Mining in Economics, Marketing and Finance, Chania 1999. SSPOP: A State Space Partial-Order Planner 2 Proc. 3rd World Multiconference on Systemics, Cybernetics and Informatics (SCI '99) 1999 0 I. Refanidis, I. Vlahavas, "SSPOP: A State Space Partial-Order Planner", SCI '99: The 3rd World Multiconference on Systemics, Cybernetics and Informatics, and ISAS '99: The 5th International Conference on Information Systems Analysis and Synthesis, Orlando, Florida, 31 July - 4 August 1999 A Heuristic Based Approach to Planning in Strips Domains 3 Advances in Informatics World Scientific D.I. Fotiadis and S.D. Nikolopoulos 2000 7 305-312 I. Refanidis and I. Vlahavas, "A Heuristic Based Approach to Planning in Strips Domains", also appears in Proc. 7th Hellenic Conference on Informatics, Ioannina, Greece, 26-29 August, 1999, pp. 96-103. Parallel, Object-Oriented and Active Knowledge Base Systems 4 Book Kluwer Academic Publishers ISBN 0-7923-8117-3 1998 168 <p>Modern data intensive real-world applications, such as data warehousing, data mining, information retrieval, expert database systems, network management, etc. strive for advanced data representation, querying and programming facilities, in order to capture the increasing demand for efficient, automated, tolerant, intelligent and really useful information systems. Such information systems can only be supported by application developing tools that provide for complex representation and efficient processing of knowledge.</p> <p>Knowledge Base Systems are an integration of conventional database systems with Artificial Intelligence techniques. Knowledge Base Systems provide inference capabilities to the database system by encapsulating the knowledge of the application domain within the database. Furthermore, Knowledge Base Systems provide sharing, ease of maintenance, and reusability of this knowledge which is usually expressed in the form of high-level, declarative rules, such as production and deductive rules.</p> <p>However, the enormous amount and complexity of data and knowledge to be processed by these systems imposes the need for increased performance and expressiveness from the Knowledge Base System. The problems associated with the large volumes of data are mainly due to the sequential data processing and the inevitable input/output bottleneck. In order to avoid this bottleneck parallel database systems have emerged to speed-up data intensive applications.</p> <p>Furthermore, the synchronous, sequential execution of large numbers of rules leads to unnecessarily many inferencing cycles that slow down Knowledge Base Systems. Parallel rule-based systems try to speed-up rule processing by executing asynchronously the various phases of rule evaluation in multiprocessor environments. Finally, the decision about the applicability of a certain piece of knowledge to a certain information state requires a large amount of pattern matching and control synchronization that can be distributed in a multiprocessor environment.</p> <p>On the other hand, the structure complexity of the data and data manipulating programs, along with the impedance mismatch between the programming languages and the relational database management systems led to the advent of Object-Oriented Database systems, an intersection of object-oriented ideas and conventional databases. Object-Oriented Databases reduce the "semantic gap" between real world concepts and data representation models. This one-to-one mapping helps the development of complex applications, such as CAD/CAM, simulation, graphical user interfaces, etc. Object-Oriented Databases encapsulate within the database system both data and programs, with advantages such as program re-use, modularization, and ease of maintenance.</p> <p>The object-oriented model offers a uniform, extensible and re-usable data and program representation that seems a promising solution for the integration of databases and knowledge-based systems. This book presents such an approach to Knowledge Base Systems: A Parallel Knowledge Base System that is built on top of a Parallel Active Object-Oriented Database System.</p> <p>In the first part of the book, we discuss extensively the various attempts to integrate one or more rule types into databases in order to provide inferencing capabilities to the latter. The initial presentation of mostly sequential Knowledge Base Systems gives the reader a feel of the various problems and the proposed solutions for such systems. At the end of this first part, we present in detail one such system which integrates high-level, declarative rules into an active Object-Oriented Database. The resulting system is a flexible and extensible knowledge base system with multiple rule support.</p> <p>In the second part of the book, we move into parallel Knowledge Base Systems by providing initial discussions of related research issues, such as parallel relational and object-oriented database systems. Many aspects of parallel rule execution are discussed including production, deductive, and active rules.</p> <p>Finally, a complete parallel Knowledge Base System is presented. The system is based on the integration of a parallel Object-Oriented Database model with the multiple-rule integration scheme that is presented in the first part. The final system is implemented on a hierarchical multiprocessor architecture.</p> <p>The book is intended as a reference text to the integration of database and knowledge base techniques for the researchers in the field of Knowledge Base Systems. It covers an extensive bibliography on the areas of rule integration in databases, namely active and deductive databases, as well as the unification of various rule types. Furthermore, the topics of parallel production, deductive, and active rule execution, both for databases and main-memory expert systems are reviewed.</p> <p>Several chapters of the book (except probably of chapters 4 and 7 that describe a specific system) analyze in detail, using examples, various techniques for the above topics. Therefore, the book can also be used as a textbook for an advanced course in Knowledge Base Systems. Finally, the book gives an in-depth insight to a specific parallel Knowledge Base System for the researchers that work in the fields of Active Databases, Knowledge Bases, and Object-Oriented Databases, on the one hand, and Parallel Databases, on the other.</p> I. Vlahavas and N. Bassiliades, "Parallel, Object-Oriented, and Active Knowledge Base Systems", Advances in Database Systems, A.K. Elmagarmid (ed.), Vol. 11, Kluwer Academic Publishers, ISBN 0-7923-8117-3, February 1998. http%3A%2F%2Fwww%2Ewkap%2Enl%2Fprod%2Fb%2F0%2D7923%2D8117%2D3 Knowledge Base Systems Object-Oriented Database Systems Active Database Systems Deductive Database Systems Parallel Production Systems Parallel Database Systems Parallel Knowledge Base Systems Exploiting And-Or Parallelism in Prolog: The OASys Computational Model and Abstract Architecture 1 The Journal of Systems and Software Elsevier Vol. 43(1) 1998 11 45-57 I. Vlahavas, "Exploiting And-Or Parallelism in Prolog: The OASys Computational Model and Abstract Architecture", The Journal of Systems and Software, Elsevier, Vol. 43(1), pp. 45-57, 1998. System Architecture of a Distributed Expert System for the management of a National Data Network 2 Proc. 8th International Conference on Artificial Intelligence (AIMSA '98) Springer-Verlag F. Giunchiglia Vol. 1480 1998 13 438-451 The management of large data networks, like a national WAN, is without any doubt a complex task. Taking into account the constantly increasing size and complexity of today's TCP/IP based networks, it becomes obvious that there is a demanding need for better than simple monitoring management tools. Expert system technology seems to be a very promising approach for the development of such tools. This paper describes the system architecture of ExperNet, a distributed expert system for the management of the National Computer Network of Ukraine, and the implementation of the tools used for its development. ExperNet is a multiagent system built in DEVICE, an active OODB enhanced with high level rules, that uses CS-Prolog II to implement the communication facilities required. The system employs HNMS+ and BigBrother, two modified versions of existing network management tools, in order to obtain a complete view of the monitored network. aimsa98.pdf I. Vlahavas, N. Bassiliades, I. Sakellariou, et al., "System Architecture of a Distributed Expert System for the management of a National Data Network", Proc. 8th International Conference on Artificial Intelligence, AIMSA'98, pp. 438-451. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fprojects%2Finco%2FInco%2Ehtml ExperNet%3A+A+Distributed+Expert+System+for+the+Management+of+a+National+Network Sozopol, Bulgaria, September 1998 http%3A%2F%2Flink%2Espringer%2Ede%2Flink%2Fservice%2Fseries%2F0558%2Fbibs%2F1480%2F14800438%2Ehtm Distributed Expert Systems Agents Network Management Distributed Prolog Parallel and Constraint Logic Programming 4 Book Kluwer Academic Publishers ISBN 0-7923-8371-0 1998 0 I. Vlahavas, P. Tsarchopoulos and I. Sakellariou, " Parallel and Constraint Logic Programming", Kluwer Academic Publishers, ISBN 0-7923-8371-0, November 1998. D-WMS: Distributed Workforce Management using CLP 2 Proc. 4th International Conf. on the Practical Applications of Constraint Technology (PACT '98) 1998 18 129-146 We present a distributed CLP-based approach for solving a real workforce management problem (BT's DT-250-118). The problem consists of a set of jobs that we want to assign to engineers in order to serve as many of them as possible at a minimum cost. We first divide the problem into sub-problems and then assign each of them to a solving agent. Each agent works independently to solve its own sub-problem and then co-operates with its peers to optimise further the intermediate results. In the sub-problem solving stage, our agents use a CLP based approach which has been used in the past in a centralised, global way. Our method allows naturally distributed scheduling and resource allocation problems to be solved in a short time with minimal disruption to the quality of solutions when compared against global approaches. kokkoras-dwms-pact98.pdf Kokkoras F. and Gregory S., "D-WMS: Distributed Workforce Management using CLP", Proc. 4th Int. Conf. on the Practical Applications of Constraint Technology (PACT'98), London, UK, 25th-27th March, 1998, pp.129-146. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpaper%5Fdetails%2Easp%3FpublicationID%3D133 D%2DWMS+on+the+CSPCONS+platform%2E constraint logic programming distributed problem solving DEVICE: Compiling Production Rules into Event-Driven Rules Using Complex Events 1 Information and Software Technology Elsevier Vol. 39, No. 5 1997 0 331-342 This paper describes a technique for the smooth integration of production rules into an active Object-Oriented Database (OODB) system that provides Event-Condition-Action (ECA) rules only, called DEVICE. The emphasis is given on the compilation of rule conditions into a discrimination network for incremental matching at run-time. The network consists of primitive, logical and complex events, that save information about partial condition element matching, as in RETE algorithm and triggers one ECA rule that corresponds to the production rule. The DEVICE method re-uses the primitives of active OODB systems, without introducing low-level data structures and provides an infrastructure for the integration of all database rule paradigms into a single knowledge base system. ist-39.ps.gz N. Bassiliades and I. Vlahavas, "DEVICE: Compiling Production Rules into Event-Driven Rules Using Complex Events", Information and Software Technology, Vol. 39(5), pp. 331-342, Elsevier Science, 1997. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdevice%2Ehtml DEVICE+system http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%3F%5Fob%3DArticleURL%26%5Fudi%3DB6V0B%2D3SNV488%2DB%26%5Fuser%3D604493%26%5FcoverDate%3D05%252F31%252F1997%26%5Frdoc%3D3%26%5Ffmt%3Dsummary%26%5Forig%3Dbrowse%26%5Fsrch%3D%2523toc%25235642%25231997%2523999609994%25239386%21%26%5Fcdi%3D5642%26%5Fsort%3Dd%26%5Facct%3DC000006498%26%5Fversion%3D1%26%5FurlVersion%3D0%26%5Fuserid%3D604493%26md5%3D4d3f6507a0f16358f68dc141b0c2f82a Production Rule Active Database Rule Compilation Discrimination Network Processing Production Rules in DEVICE, an Active Knowledge Base System 1 Data and Knowledge Engineering Elsevier Vol. 24, No. 2 1997 38 117-155 Production rules are useful for several tasks of active database systems, such as integrity constraint checking, derived data maintenance, database state monitoring, etc. Furthermore production rules can express knowledge in a high-level form for problem solving in Knowledge Base Systems (KBS). Present active object-oriented database (OODB) systems traditionally provide event-driven rules which are triggered by events, i.e. database modifications. This paper describes DEVICE, a high-level rule integration scheme into an active OODB system, resulting in an active KBS. The paper emphasizes on the run-time processing of production rules, namely the incremental matching of rule conditions, as well as rule selection and firing. The matching of production rules requires special algorithms based on the flow of updated data through a discrimination network, like RETE, TREAT, etc. DEVICE offers a smooth integration of production rules into an active OODB system that only supports event-driven rules, without introducing new data structures, maintaining at the same time the properties of discrimination networks. This is achieved using complex events to map the conditions of production rules and monitor the database to incrementally match those conditions. DEVICE maps each production rule into one event-driven rule that is easy to maintain and offers centralized rule selection control for correct run-time behavior and conflict resolution. Furthermore, DEVICE provides the infrastructure for the integration of various other rule paradigms into a single KBS, like deductive rules and integrity constraints and leaves room for the optimization of the matching process through variations of the basic discrimination network. DKE-24.ps.gz N. Bassiliades and I. Vlahavas, "Processing Production Rules in DEVICE, an Active Knowledge Base System", Data and Knowledge Engineering, Vol. 24(2), pp. 117-155, Elsevier Science, 1997. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdevice%2Ehtml DEVICE http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%3F%5Fob%3DArticleURL%26%5Fudi%3DB6TYX%2D3SYR69Y%2D1%26%5Fuser%3D604493%26%5FcoverDate%3D11%252F30%252F1997%26%5Frdoc%3D1%26%5Ffmt%3Dsummary%26%5Forig%3Dbrowse%26%5Fsrch%3D%2523toc%25235630%25231997%2523999759997%252313805%21%26%5Fcdi%3D5630%26%5Fsort%3Dd%26%5Facct%3DC000006498%26%5Fversion%3D1%26%5FurlVersion%3D0%26%5Fuserid%3D604493%26md5%3D92cb5580674c34c0bcb36d0073d0c9bb Knowledge Base Systems Production Rules Active Object-Oriented Databases Complex Events Discrimination Network IDIS-KS: an Intelligent Drug Information System as a Knowledge Server 1 Proc. 14th International Congress of Medical Informatics Europe (MIE '97) 1997 4 368-372 I. Vlahavas, A. Pombortsis and I. Sakellariou, "IDIS-KS: an Intelligent Drug Information System as a Knowledge Server", 14th Int. Congress of Medical Informatics Europe (MIE-97), May 1997, IOS press, 368-372. A Deductive Object-Oriented Database System based on Active Rules 2 Proc. 6th Hellenic Conference on Informatics 1997 9 180-189 This paper describes a Deductive Object-Oriented Database (DOOD) system that is built on top of an active Object-Oriented Database (OODB) system. The system, named DEVICE, uses the primitives of the latter, like active rules, simple and complex events, to integrate deductive and production rules. The integration is based on the emulation of deductive rules by special purpose if-then-else production rules that have been smoothly integrated into an active OODB. The DEVICE system supports thus multiple rule systems, like active (event-driven), production (data-driven) and deductive (goal-driven) rules into the same OODB system. The core of this multiple rule integration is: a) the mapping of each high-level rule into one event-driven rule, offering centralised rule selection control for correct run-time behaviour and conflict resolution, and b) the use of complex events to map the conditions of high-level rules and monitor the database to incrementally match those conditions. DEVICE is extensible because a) it re-uses the primitives of the host active OODB system to build the integration scheme, without introduc-ing low-level data structures that do not blend well with the OO model and are not easily extensible, and b) the rule managers support general-purpose rule scheduling functions. In conclusion, DEVICE is a flexible Knowledge Base System (KBS) that gives the user the ability to express knowledge in a vari-ety of high-level forms for advanced problem solving in data intensive applications. epy97.ps.gz N. Bassiliades and I. Vlahavas, "A Deductive Object-Oriented Database System based on Active Rules", 6th Hellenic Conference on Informatics, Athens, December 1997, pp. 180-189. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdevice%2Ehtml DEVICE Athens, December 1997 Knowledge Base System Deductive Rule Active Rule Production Rule Derived Class Discrimination Network The Basic OASys Model: Preliminary Results 2 Proc. 6th Hellenic Conference on Informatics 1997 8 723-731 I.Vlahavas, P. Kefalas, I. Sakellariou and C. Halatsis, "The Basic OASys Model: Preliminary Results", 6th Hellenic Conference on Informatics, Athens, December 1997, pp. 723-731. A New Hybrid Neural-Genetic Methodology for Improving Learning 2 Proc. IEEE 9th International Conference on Tools with Artificial Intelligence (ICTAI '97) 1997 0 A. Likartsis, I. Vlahavas and L. Tsoukalas, "A New Hybrid Neural-Genetic Methodology for Improving Learning",Proc. IEEE 9th International Conference on Tools with Artificial Intelligence, ICTAI'97. Path Planning in a 2-D known Space using Neural Networks and Skeletonization 2 Proc. IEEE International Conference on Systems Man and Cybernetics (SMC '97) 1997 0 N. Bourbakis, D. Goldman, R. Fermatt, I. Vlahavas and L. Tsoukalas, "Skeletonization: Neural Network Based Path Planning in a 2-D known Space", Proc. IEEE International conference on Systems Man and Cybernetics, SMC '97, 1997. Hierarchical Query Execution in a Parallel Object-Oriented Database System 1 Parallel Computing Elsevier Vol. 22, No. 7 1996 31 1017-1048 This article presents a hierarchical query execution strategy for a parallel object-oriented database (OODB) system. The system, named PRACTIC, is based on a concurrent active class management model and is mapped to an abstract hierarchical multiprocessor architecture. The proposed strategy is studied analytically and by simulation on a transputer-based machine, verifying the theoretical results. Although the analysis suits both main-memory and disk-based database systems, it becomes significant for main-memory systems where the multiprocessor initialization and communication overheads are comparable to the actual workload. The hierarchical query execution strategy is proved much better than the usual flat strategy of parallel database systems, except some clearly identified extreme cases, where flat processing is better. Furthermore, we propose a declustering scheme for space optimization to improve processor utilization and single-class query performance, by having different classes share memory and computation power of neighboring processing elements. parco-22.ps.gz N. Bassiliades and I. Vlahavas,"Hierarchical Query Execution in a Parallel Object-Oriented Database System", Parallel Computing, Vol. 22 (7), pp. 1017-1048, Elsevier Science, 1996. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fpractic%2Ehtml PRACTIC http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%3F%5Fob%3DArticleURL%26%5Fudi%3DB6V12%2D3WP2DJV%2DC%26%5Fuser%3D604493%26%5FcoverDate%3D10%252F01%252F1996%26%5Frdoc%3D5%26%5Ffmt%3Dsummary%26%5Forig%3Dbrowse%26%5Fsrch%3D%2523toc%25235662%25231996%2523999779992%2523101992%21%26%5Fcdi%3D5662%26%5Fsort%3Dd%26%5Facct%3DC000006498%26%5Fversion%3D1%26%5FurlVersion%3D0%26%5Fuserid%3D604493%26md5%3D8377d220ca4b73e36f780782c4bdef24 Parallel Main-Memory Database System Object-Oriented Databases Multiprocessor Architecture Parallel Query Execution Analytic Performance Model Simulation Multiple OR-Parallel Resolution: Meta-Level Control of Parallel Logic Programs 2 Proc. Parallel Processing (Euro-Par '96) Springer-Verlag Bouge, P. Fraigniaud, A. Mignotte and Y. Robert LNCS 1123 1996 9 694-703 P.Kefalas and I.Vlahavas, "Multiple OR-Parallel Resolution: Meta-Level Control of Parallel Logic Programs", Lecture Notes in Computer Science 1123, L.Bouge, P.Fraigniaud, A.Mignotte, Y.Robert (eds.), Euro-Par'96 Parallel Processing, Lyon, France, Springer Verlag, pp.694-703, 1996. Towards a Drug Information Center: The Aesculapius project 2 Proc. 1st EURO DRUG Conference 1996 0 K. Mamzoridi, M. Pirpasopulos, I. Vlahavas and A. Pomportsis, "Towards a Drug Information Center: The Aesculapius project", 1st EURO DURG conference, Karolinska Institute, Sweden, 1996. CoLan: A Functional Constraint Language and Its Implementation 1 Data and Knowledge Engineering Vol. 14, No. 3 1995 46 203-249 This paper is about the definition of CoLan, a high-level declarative Constraint Descrip-tion Language, for use with an Object-Oriented Database (OODB). CoLan has features of both first-order logic and functional programming and is based on Daplex. CoLan expres-sions are tran-slated into Prolog code that implements the operational semantics of the con-straint. Pieces of generated code are cached inside the class descriptor of the "host" class attached to appropriate slots. The pieces of code are retrieved along an inheritance path when an update on the data-ba-se is attempted. If the update violates any of the re-trieved constraints then it is re-jected with an info-rmative message. Thus constraints are ex-pressed declaratively and they can even be retracted individually. However, they are im-plemented efficiently as code-gene-rated methods, triggered selectively by an update. The implementation is described for the ADAM OODB, which uses meta-classes of the CoLan system to generate class descriptions. DKE-14.pdf N. Bassiliades and P.M.D. Gray, "CoLan: A Functional Constraint Language and Its Implementation", Data and Knowledge Engineering, Vol. 14, No. 3, February 1995, pp. 203-249 http%3A%2F%2Fwww%2Ecsd%2Eauth%2Egr%2F%7Enick%2Fsystems%2Fcolan%2Ehtml CoLan http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%3F%5Fob%3DArticleURL%26%5Fudi%3DB6TYX%2D3Y5FDKX%2D2%26%5Fuser%3D604493%26%5FcoverDate%3D02%252F28%252F1995%26%5Frdoc%3D2%26%5Ffmt%3Dsummary%26%5Forig%3Dbrowse%26%5Fsrch%3D%2523toc%25235630%25231995%2523999859996%2523148850%21%26%5Fcdi%3D5630%26%5Fsort%3Dd%26%5Facct%3DC000006498%26%5Fversion%3D1%26%5FurlVersion%3D0%26%5Fuserid%3D604493%26md5%3D3d3c4b61e431787bc44d3802532d8d46 Semantic Integrity Constraints Object-Oriented Databases Functional Data Model Constraint Compilation Incremental Constraint Checking Numerical Quantifiers Constraint Checking in a Parallel Object-Oriented Database System 1 Journal of Parallel Algorithms and Applications Gordon and Breach Vol. 5, No. 3-4 1995 18 129-147 This paper deals with parallel checking of passive constraints in object-oriented databases. It presents a parallel algorithm for constraint checking based on a master-slave technique and discusses its implementation on a parallel object-oriented database system. The system is named PRACTIC and is based on class concurrency. Passive constraints, unlike active database rules, are independent and can be executed using AND-parallelism. Simulation shows that the proposed algorithm offers considerable speed-up, which mainly depends on the number of constraints and the total constraint execution time, while it is only slightly affected from the distribution of constraints and the constraint scheduling policy. Finally, it is explained how the PRACTIC system enhances the algorithm's performance using features, like nested query parallelism and constraint overlapping. PAA-5.ps.gz N. Bassiliades and I.Vlahavas, "Constraint Checking in a Parallel Object-Oriented Database System", Journal of Parallel Algorithms and Applications, Gordon and Breach, Vol. 5 (3-4), pp. 129-147, 1995. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fpractic%2Ehtml PRACTIC Passive Constraints Parallel Constraint Checking Object-Oriented Databases Parallel Database Systems Independent-Task Parallelism COMFRESH: A Common Framework for Expert Systems and Hypertext 1 Information Processing and Management Pergamon, Elsevier Vol. 31 (4) 1995 17 593-604 Intelligent hypertext is a promising approach to information systems, because it combines the power of inference of expert systems and the intuitive power of hypertext. In this paper we propose the "COMFRESH", a common framework for expert systems and hypertext. It is based on a Prolog interpreter and uses the conceptual graph knowledge representation formalism for browsing and reasoning. COMFRESH can be used as a knowledge based hypertext (intelligent hypertext) or as an expert system with hypertext capabilities. kokkoras-comfresh-ipm.pdf F. Kokkoras and I.Vlahavas, "COMFRESH: A common framework for expert systems and hypertext", Information Processing and Management, Pergamon, Elsevier, Vol. 31 (4), pp. 593-604, 1995. http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%3F%5Fob%3DIssueURL%26%5Ftockey%3D%2523TOC%25235948%25231995%2523999689995%2523163566%2523FLP%2523Volume%5F31%2C%5FIssue%5F4%2C%5FPages%5F455%2D624%5F%28July%5F1995%29%26%5Fauth%3Dy%26view%3Dc%26%5Facct%3DC000006498%26%5Fversion%3D1%26%5FurlVersion%3D0%26%5Fuserid%3D604493%26md5%3D6f8f179326b7923e601d0f28257f77cb hypertext conceptual graphs information retrieval PRACTIC: A Concurrent Object Data Model for a Parallel Object-Oriented Database System 1 Information Sciences Elsevier Vol. 86 (1-3) 1995 29 149-178 In this paper, a concurrent object data model for a parallel object-oriented database system, named PRACTIC, and its abstract machine are presented. PRACTIC means PaRallel ACTIve Classes and is based on the vertical partitioning and concurrent management of the database schema classes and meta-classes, which are collectively called active objects. Active objects are permanent processes in memory that encapsulate their definitions, methods and management procedures. Semi-active and passive objects exist to realise abstract classes and instances (the actual data), respectively. The object model gives rise to a query/method execution model that provides parallelism on all levels of the instantiation hierarchy. The abstract PRACTIC machine directly maps the model to a MIMD machine. The performance of one of the proposed parallel query/method execution schemes is measured by simulation on the abstract machine. IS-86.ps.gz N. Bassiliades and I.Vlahavas, "PRACTIC: A Concurrent Object Data Model for a Parallel Object-Oriented Database System", Information Sciences, Elsevier, Vol. 86 (1-3), pp. 149-178, 1995. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fpractic%2Ehtml PRACTIC http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%3F%5Fob%3DArticleURL%26%5Fudi%3DB6V0C%2D41JTC98%2D7%26%5Fuser%3D604493%26%5FcoverDate%3D09%252F30%252F1995%26%5Frdoc%3D7%26%5Ffmt%3Dsummary%26%5Forig%3Dbrowse%26%5Fsrch%3D%2523toc%25235643%25231995%2523999139998%2523216636%21%26%5Fcdi%3D5643%26%5Fsort%3Dd%26%5Facct%3DC000006498%26%5Fversion%3D1%26%5FurlVersion%3D0%26%5Fuserid%3D604493%26md5%3Df80fabf5a4d2af5776c35293d8595582 Object-Oriented Databases Active Objects Instance Hierarchy Inheritance Parallel Method Execution Vertically Partitioned Systems A Non-Uniform Data Fragmentation Strategy for Parallel Main-Memory Database Systems 2 Proc. 21st International Conference on Very Large Data Bases (VLDB '95) Morgan Kaufmann U. Dayal, P.M.D. Gray, S. Nishio 1995 11 370-381 In multi-processor database systems there are processor initialization and inter-communication overheads that diverge real systems from the ideal linear behaviour as the number of processors in-creases. Main-memory database systems suffer more since the database processing cost is small compared to disk-based database systems and thus comparable to the processor initialization cost. The usual uniform data fragmentation strategy divides a relation into equal data partitions, lead-ing to idleness of single processors after local query execution termination and before global termination. In this paper, we propose a new, non-uniform data fragmentation strategy that re-sults in concurrent termination of query process-ing among all the processors. The proposed fragmentation strategy is analytically modeled, simulated and compared to the uniform strategy. It is proven that the non-uniform fragmentation strat-egy offers inherently better performance for a par-allel database system than the uniform strategy. Furthermore, the non-uniform strategy scales-up perfectly till an upper limit, after which a system re-configuration is needed. vldb95.ps.gz N. Bassiliades and I.Vlahavas, "A Non-Uniform Data Fragmentation Strategy for Parallel Main-Memory Database Systems", Proc. 21st International Conference on Very Large Data Bases, VLDB'95, Zurich, Switzerland, 1995, pp. 370-381. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fpractic%2Ehtml PRACTIC September 11-15, 1995, Zurich, Switzerland http%3A%2F%2Fwww%2Einformatik%2Euni%2Dtrier%2Ede%2F%257Eley%2Fdb%2Fconf%2Fvldb%2FBassiliadesV95%2Ehtml Multi-Processor Database System Parallel Query Execution Main-Memory Database Data Fragmentation Analytic Model Speed-up Scale-up Hashing Function GRS-Prolog: Applying Different Resolution Strategies to Prolog Programs 2 Proc. 5th Hellenic Conference on Information Techology 1995 0 P. Kefalas and E. Tzelalis, "GRS-Prolog: Applying Different Resolution Strategies to Prolog Programs", 5th Hellenic Conference on Information Techology, Athens, 1995. Parallel Management of Large Deductive Databases in a Multi-Processor Environment 2 Proc. IEEE Mediterranean Electrotechnical Conference (Melecon '94) 1994 3 367-370 This paper describes a parallel deductive database system, built on top of Prolog. The system is based on the TOP-DOWN evaluation of logic programs. Parallelism is provided at the rule level, by transforming the query AND/OR tree into Disjunctive Normal Form. The clauses of the transformed formula are executed independently in parallel, on a transputer multi-processor machine, using the processor-farm algorithm. Both main-memory consultation and direct disk access have been implemented and tested. The measurement of the system performance shows speed improvement over the sequential Prolog interpreter, for large rule bases, but also exhibits implementation-dependent drawbacks that cause under-linear speed-up. melecon94.ps.gz C. Maciazek, N. Bassiliades and I. Vlahavas, "Parallel Management of Large Deductive Databases in a Multi-Processor Environment", Proc. 7th IEEE Mediterranean Electrotechnical Conference, Melecon'94, Antalya, Turkey, April 1994, pp. 367-370. Antalya, Turkey, April 1994 Deductive Databases Prolog Transputers OR-Parallelism Modelling Constraints with Exceptions in Object-Oriented Databases 2 Entity-Relationship Approach - ER'94, Business Modelling and Re-Engineering, 13th International Conference on the Entity-Relationship Approach Springer-Verlag P. Loucopoulos LNCS 881 1994 15 189-204 This paper deals with modelling constraints in object-oriented databases, with emphasis given on exceptions. Constraints are restrictions on properties and relations of database objects that ensure the integrity of data. Therefore, they should be obeyed by every object, but as in real-life, there are some exceptions to this rule. Object-oriented databases provide rich semantic constructs, adequate to model real-world relations. Inheritance of constraints in object-oriented databases has been treated in a completely mandatory way, providing no room for exceptions. In this paper, an object-oriented constraint representation scheme is presented, along with a methodology for modelling constraint exceptions. Finally, an algorithm is described that ensures correct run-time resolution of constraint applicability. Since business is not in abstract, but in real-world, business database modelling would be greatly benefited from a tool that allows both a clear definition and an efficient enforcement of constraints with exceptions. ER94.ps.gz N.Bassiliades and I.Vlahavas, "Modelling Constraints with Exceptions in Object-Oriented Databases", Entity-Relationship Approach - ER'94, Business Modelling and Re-Engineering, Lecture Notes in Computer Science, P. Loucopoulos (Ed.), Vol. 881, pp. 189-204, 1994. Manchester, United Kingdom, December 1994 http%3A%2F%2Fwww%2Einformatik%2Euni%2Dtrier%2Ede%2F%257Eley%2Fdb%2Fconf%2Fer%2FBassiliadesV94%2Ehtml Semantic Integrity Constraints Exceptions Multiple Inheritance A Contribution to the Problem of Avoiding Congestion in Multistage Networks in the Presence of Unbalanced Traffic 1 The Journal of Systems and Software Elsevier Vol. 26 1994 11 273-284 A.Pomportsis and I.Vlahavas, "A Contribution to the Problem of Avoiding Congestion in Multistage Networks in the Presence of Unbalanced Traffic", The Journal of Systems and Software, Elsevier vol. 26, pp. 273-284, 1994 Flow Control and Switching Strategy for Preventing Congestion in Multistage Networks 3 Architecture and Protocols for High-speed Networks Kluwer Academic Publishers O. Spaniol, A. Dantine and W. Effelsberg 1994 0 A.Pomportsis and I.Vlahavas, "Flow Control and Switching Strategy for Preventing Congestion in Multistage Networks", Architecture and Protocols for High-speed Networks, eds. O. Spaniol, A. Dantine and W. Effelsberg, KLUWER Academic Publishers, 1994. Constraint Maintenance Using Generated Methods in the P/FDM Object-Oriented Database 2 Proc. 1st International Workshop on Rules in Database Systems Springer-Verlag N.W. Paton and M.H. Williams 1994 17 364-381 We discuss the use of code-generated methods in Prolog as a flexible and efficient way to implement complex semantic constraints in an OODB. We introduce a high-level constraint language CoLan, based on functions and sets and including range quantifiers, from which fragments of code are generated to check the constraints. These fragments are attached to slots in class descriptors, and are also inherited (constraints cannot be overridden). Thus many fragments can come from one constraint and one slot may have attached fragments from many constraints. Constraints can be selectively disabled or removed which causes inhibition or disabling of corresponding fragments. This overcomes many objections to implementing constraints through methods. We have prototyped it by using the metaclass facilities of ADAM to initiate code generation. We are now re-implementing it in P/FDM, using changes to metadata (P/FDM does not have full metaclasses). This will incorporate a transaction mechanism and also provide queries on constraints. This approach opens a number of interesting future directions. RIDS93.ps.gz S.M. Embury, P.M.D. Gray and N. Bassiliades, "Constraint Maintenance Using Generated Methods in the P/FDM Object-Oriented Database", in N.W. Paton and M.H. Williams (Eds.), Rules in Database Systems: Proceedings of the 1st International Workshop on Rules in Database Systems, Edinburgh, Scotland, 30 August- 1 September 1993, Workshops in Computing Series, Springer-Verlag, 1994, pp. 364-381. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Fnbassili%2Fsystems%2Fcolan%2Ehtml CoLan Edinburgh, Scotland, 30 August- 1 September 1993 Semantic Integrity Constraints Object-Oriented Databases Functional Data Model Constraint Compilation Incremental Constraint Checking Meta-Data The AND/OR Parallel Prolog Machine APIM: Execution Model and Abstract Design 1 Journal of Programming Languages Chapman and Hall Vol. 1 1993 16 245-261 I.Vlahavas and P.Kefalas, "The AND/OR Parallel Prolog Machine APIM: Execution Model and Abstract Design", Journal of Programming Languages, Chapman and Hall, Vol. 1 pp.245-261, 1993. A Novel Flow Control and Switching Strategy for Preventing Hot Spot Congestion in Multistage Networks 1 Microprocessors and Microsystems Butterworth Vol. 17 (7) 1993 0 A.Pomportsis and I.Vlahavas, "A Novel Flow Control and Switching Strategy for Preventing Hot Spot Congestion in Multistage Networks", Microprocessors and Microsystems, Butterworth, vol. 17, No 7, 1993. A Multiprocessor Machine for the Parallel Management of an Active Object-Oriented DataBase (in Greek) 2 Proc. 4th Hellenic Conference on Informatics 1993 11 501-512 H εργασία περιγράφει ένα πλαίσιο παράλληλης εκτέλεσης των διαδικασιών μιας αντικειμενοστραφούς ενεργής βάσης δεδομένων πάνω σε μία μηχανή με πολλούς επεξεργαστές. Στην εργασία περιγράφεται σύντομα το μοντέλο διαχείρισης της βάσης δεδομένων CLASP (Class Parrallel), το οποίο βασίζεται στον παραλληλισμό των κλάσεων, καθώς επίσης και μια μηχανή πολλών επεξεργαστών (CLASP μηχανή) πάνω στην οποία εκτελείται το μοντέλο. Στη συνέχεια παρουσιάζεται ένας μηχανισμός εκτέλεσης των ενεργών κανόνων μιας αντικειμενοστραφούς βάσης και η υλοποίησή του στην CLASP μηχανή. Τέλος, αναφέρονται στοιχεία της απόδοσης του προτεινόμενου μηχανισμού σε σύγκριση με αντίστοιχους σειριακούς. N. Bassiliades and I. Vlahavas, "A Multiprocessor Machine for the Parallel Management of an Active Object-Oriented DataBase", Proc. 4th Hellenic Conference on Informatics, Patra, Greece, December 1993, pp. 501-512 (in Greek). Patra, Greece, December 1993 A Multiprocessor System for the Efficient Management of Large Knowledge Bases (in Greek) 2 Proc. 6th Hellenic Physics Conference 1993 5 217-221 Στην εργασία περιγράφεται ένα σύστημα πολυ-επεξεργαστών για τη διαχείριση μιας αντικειμενοστραφούς βάσης γνώσης. Στην αρχή αναπτύσσεται η ανάγκη ενοποίησης των βάσεων δεδομένων με τις βάσεις γνώσης. Στη συνέχεια περιγράφεται ένα μοντέλο διαχείρισης μιας αντικειμενοστραφούς βάσης γνώσης και η αρχιτεκτονική στην οποία υλοποιείται. N.Bassiliades and I.Vlahavas, "A Multiprocessor System for the Efficient Management of Large Knowledge Bases" (in Greek), Proc. 6th Hellenic Physics Conference, Greece, 1993. Xanthi, Greece, March 1993 An Expert System for Land Evaluation 2 Proc. 6th Hellenic Physics Conference 1993 0 E.Sakelariou and I.Vlahavas, "An Expert System for Land Evaluation" (in Greek), Proc. 6th Hellenic Physics Conference, Greece, 1993. A Parallel Prolog Resolution Based on Multiple Unifications 1 Parallel Computing Elsevier Vol. 18 1992 0 1275-1283 I.Vlahavas and P.Kefalas, "A Parallel Prolog Resolution Based on Multiple Unifications", Parallel Computing, Elsevier, vol.18, 1275-1283, 1992. An Abstract Prolog Machine Based on Parallel Resolution Principle 2 Proc. Microprocessing and Microprogramming (Euromicro '92) North-Holland No. 35 1992 7 755-762 I. Vlahavas and P. Kefalas, "An Abstract Prolog Machine Based on Parallel Resolution Principle", Proc. Euromicro 92, Microprocessing and Microprogramming, The Euromicro Journal, North-Holland, no. 35, pp. 755-762, 1992. Flow Control in Packet-Switched Multistaged Interconnection Networks 2 Proc. CompEuro '92 1992 0 A.Pomportsis and I.Vlahavas, "Flow Control in Packet-Switched Multistaged Interconnection Networks", Proc. CompEuro '92. Preventing Performance Degradation in Packet-Switched Multistage Interconnection Networks Under Nonuniform Traffic 2 Proc. Microcomputer and Microprocessor Applications Conference 1992 0 A.Pomportsis and I.Vlahavas, "Preventing Performance Degradation in Packet-Switched Multistage Interconnection Networks Under Nonuniform Traffic", Proc. Microcomputer and Microprocessor Applications Conf., Budapest 1992. Hill-Climbing and Genetic Algorithms coded using OR-parallel Logic plus Meta-Control 2 Proc. IJCAI'91 1st International Workshop on Parallel Processing for AI 1991 0 P.Kefalas and T.J.Reynolds, "Hill-Climbing and Genetic Algorithms coded using OR-parallel Logic plus Meta-Control", Proc. of the 1st Intern. Workshop on Parallel Processing for AI, held in association with Intern. Conference in AI, IJCAI-91, Sydney, 1991. Controlling Search with meta-BRAVE 2 Proc. ICLP'91 Workshop in Parallel Execution of Logic Programs Springer-Verlag A. Beaumont and G. Gupta LNCS 569 1991 0 P.Kefalas and T.J.Reynolds, "Controlling Search with meta-BRAVE", In Parallel Execution of Logic Programs, ICLP'91 Pre-conference Workshop, Lecture Notes in Computer Science, vol.569, A.Beaumont and G.Gupta (eds.), Springer-Verlag, Paris 1991. BRAVE: An OR-Parallel dialect of Prolog and its Application to AI 2 Proc. 1st and 2nd International Conference in Logic Programming in Soviet Union Springer-Verlag LNAI 592 1991 0 T.J.Reynolds and P.Kefalas, "BRAVE: An OR-Parallel dialect of Prolog and its Application to AI", Proc. of the 1st and 2nd Inter. Conf. in Logic Programming in Soviet Union, Lecture Notes in Artificial Intelligence, vol. 592, Springer-Verlag, Irkutsk 1990- Leningrad 1991. A Communication Analysis For Optimizing Hierarchical Multicomputer Systems 2 Proc. CompEuro '91 1991 0 A.Pomportsis and I.Vlahavas, "A Communication Analysis For Optimizing Hierarchical Multicomputer Systems", Proc. CompEuro '91. LBASE: A Logical Database Management System 2 General Conference of the Balkan Physical Union 1991 3 640-642 Logic offers a uniform environment both for data description and program execution and provides a powerful programming language with the use of "headed" queries. Relational data model proved to be the best model of the "conventional" data-base theory. This paper describes the advantages of connecting a relational data-base management system (RDBMS) (DBASE III PLUS) with logic programming (Arity Prolog), building a logical DBMS (LDBMS), called LBASE, that extends the power of the relational data model. Relations are described both by facts (assertions) and rules (deductions), so data become more meaningful, and complex queries can be answered. Memory organisation also provides space saving and efficient indexing using b-trees instead of sequential searching. lbase.ps.gz N.Vasiliades and I.Vlahavas, "LBASE: A Logical Database Management System", Proc. General Conference of the Balkan Physical Union, 1991. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Flbase%2Ehtml LBASE Thessaloniki, Greece, September 1991 Prolog Databases Dbase Towards a Parallel Inference Machine: the APIM project 2 Proc. Microprocessing and Microprogramming (Euromicro '90) North-Holland No. 30 1990 5 201-206 I. Vlahavas, A. Pomportsis and D. Stamatis, "Towards a Parallel Inference Machine: the APIM project", Proc. Euromicro 90, Microprocessing and Microprogramming, The Euromicro Journal, North-Holland, no. 30, pp. 201-206, 1990. OR-Parallel Prolog and Search Problems in AI 2 Proc. 7th International Conference in Logic Programming MIT Press D.H.D. Warren and P. Szeredi 1990 0 T.J.Reynolds and P.Kefalas, "OR-Parallel Prolog and Search Problems in AI", Proc. of the 7th Intern. Conf. in Logic Programming, D.H.D.Warren and P.Szeredi (eds.), MIT Press, Jerusalem, 1990. OR-Parallel Logic Languages in Law Applications 2 Proc. 3rd International Cong. in Law and Expert Systems A.A. Martino 1989 15 579-594 T.J.Reynolds and P.Kefalas, "OR-Parallel Logic Languages in Law Applications", Proc. of the 3rd Intern. Cong. in Law and Expert Systems, A.A.Martino (ed.), pp. 579-594, Florence, Italy, 1989. L-Machine: A Low Cost Personal Sequential Inference Machine 1 The Journal of Systems and Software Elsevier Vol. 9 1989 11 209-223 I.Vlahavas and C.Halatsis, "L-Machine: A Low Cost Personal Sequential Inference Machine", The Journal of Systems and Software, Elsevier, Vol. 9, pp. 209-223, 1989. A RISC Prolog Machine Architecture 2 Proc. Microprocessing and Microprogramming (Euromicro '87) North-Holland No. 21 1987 17 259-266 I. Vlahavas and C. Halatsis, "A RISC Prolog Machine Architecture", Proc. Euromicro 87, Microprocessing and Microprogramming, The Euromicro Journal, North-Holland, no. 21, pp. 259-266, 1987. On the Performance of Packet Switching Interconnection Networks for Multiprocessor Systems 2 Proc. Parallel Processing and Applications 1987 0 A.Pomportsis, I.Vlahavas and C.Halatsis, "On the Performance of Packet Switching Interconnection Networks for Multiprocessor Systems", Proc. Parallel Processing and Applications, L'Aquila 1987. A New Abstract Prolog Instruction Set 2 Proc. Avignon 87, Expert systems and their applications 1987 0 I.Vlahavas and C.Halatsis, "A New Abstract Prolog Instruction Set", Proc. Avignon 87, Expert systems and their applications, Avignon, May 1987 A Prolog Compiler for the L-Machine (in Greek) 2 Proc. 4th Hellenic Physics Conference 1986 0 C.Konidaris and I.Vlahavas, "A Prolog Compiler for the L-Machine", (in Greek), Proc. 4th Hellenic Physics Conference, Greece, 1986. Artificial Intelligence (in Greek - Τεχνητή Νοημοσύνη) 4 Book Gartaganis Publications ISBN 960-7013-28-X 2002 600 <p>Τεχνητή Νοημοσύνη είναι ο τομέας της επιστήμης των υπολογιστών που ασχολείται με τη σχεδίαση ευφυών (νοημόνων) υπολογιστικών συστημάτων, δηλαδή συστημάτων που επιδεικνύουν χαρακτηριστικά που σχετίζουμε με τη νοημοσύνη στην ανθρώπινη συμπεριφορά.</p> <p>Το βιβλίο αυτό προσεγγίζει το θέμα της Τεχνητής Νοημοσύνης οριοθετώντας τα προβλήματα που αυτή αντιμετωπίζει, περιγράφοντας τρόπους αναπαράστασης της γνώσης των προβλημάτων και παρουσιάζοντας αλγορίθμους αναζήτησης των λύσεων τους.</p> <p>Το περιεχόμενο συμπληρώνεται με τη μελέτη κλασσικών εφαρμογών που ιστορικά ανήκουν στην περιοχή, όπως ο σχεδιασμός ενεργειών, η μηχανική μάθηση, τα έμπειρα συστήματα, κ.α. ενώ για λόγους πληρότητας γίνεται συνοπτική παρουσίαση και άλλων δημοφιλών εφαρμογών, όπως για παράδειγμα η ρομποτική και η μηχανική όραση.</p> <p>Σκοπός του βιβλίου είναι να αποτελέσει διδακτικό εγχειρίδιο για τη διδασκαλία του μαθήματος της Τεχνητής Νοημοσύνης, καθώς και συναφών μαθημάτων όπως Έμπειρα Συστήματα και Συστήματα Λήψης Αποφάσεων. Κάποια θέματα αναπτύσσονται σε μεγαλύτερο βάθος και απευθύνονται σε μεταπτυχιακούς φοιτητές ή ερευνητές, αποτυπώνοντας την πολυετή ερευνητική εμπειρία της συγγραφικής ομάδας σε αυτά. </p> <p>Με την προσπάθεια αυτή επιχειρείται να καλυφθεί ένα κενό της Ελληνικής βιβλιογραφίας για ένα τόσο σημαντικό θέμα και να προαχθεί κατά το δυνατό η διάδοση και εφαρμογή της Τεχνητής Νοημοσύνης στον Ελληνικό εκπαιδευτικό και ερευνητικό χώρο.</p> I. Vlahavas, P. Kefalas, N. Bassiliades, I. Refanidis, F. Kokkoras, I. Sakellariou, "Artificial Intelligence", (in Greek), Gartaganis Publications, ISBN 960-7013-28-X, 2002. http%3A%2F%2Faibook%2Ecsd%2Eauth%2Egr%2Fv1%2Findex%2Ehtml +Home+Page+of+1st+edition http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2FAIbook%2F Artificial Intelligence Knowledge Systems An Intelligent Educational Metadata Repository 3 Intelligent Systems, Techniques and Applications CRC Press C.T. Leondes Vol. 4, Ch.12 2003 40 297-337 Recently, several standardization efforts for e-learning technologies gave rise to various specifications for educational metadata, that is, data describing all the "entities" involved in an educational procedure. The internal details of systems that utilize these metadata are still an open issue since these efforts are primarily dealing with "what" and not "how". In this chapter, under the light of these emerging standardization efforts, we present X-DEVICE, an intelligent XML repository system for educational metadata. X-DEVICE can be used as the intelligent back-end of a WWW portal on which "learning objects" are supplied by educational service providers and accessed by learners according to their individual profiles and educational needs. X-DEVICE transforms the widely adopted XML binding for educational metadata into a flexible, object-oriented representation and uses intelligent second-order logic querying facilities to provide advanced, personalized functionality. Furthermore, a case study is presented, in which learning object metadata and learner's profile metadata are combined under certain X-DEVICE rules in order to dynamically infer customized courses for the learner. crc-chapter1.pdf N. Bassiliades, F. Kokkoras, I. Vlahavas and D. Sampson, "An Intelligent Educational Metadata Repository", Intelligent Systems, Techniques and Applications, C.T. Leondes (ed), vol. 4, Ch. 12, pp. 297-337, CRC Press, 2003. http%3A%2F%2Fwww%2Ecsd%2Eauth%2Egr%2F%257Elpis%2Fsystems%2Fx%2Ddevice%2Ehtml X%2DDEVICE http%3A%2F%2Fwww%2Ecrcpress%2Ecom%2Fshopping%5Fcart%2Fproducts%2Fproduct%5Fdetail%2Easp%3Fsku%3D1121%26parent%5Fid%3D%26pc%3D e-learning XML educational metadata Using Logic for Querying XML Data 3 Web Powered Databases IDEA Group Publishing D. Taniar and W. Rahayu Ch. 1 2003 35 1-35 In this chapter, we propose the use of first-order logic, in the form of deductive database rules, as a query language for XML data and we present X-DEVICE, an extension of the deductive object-oriented database system DEVICE for storing and querying XML data. XML documents are stored into the OODB by automatically mapping the DTD to an object schema. XML elements are treated either as classes or attributes based on their complexity, without loosing the relative order of elements in the original document. Furthermore, this chapter describes the extension of the system's deductive rule query language with second-order variables, general path and ordering expressions, for querying over the stored, tree-structured XML data and constructing XML documents as a result. The extensions were implemented by translating all the extended features into the basic, first-order deductive rule language of DEVICE using meta-data about stored XML objects. idea-chapter1.pdf N. Bassiliades, I. Vlahavas and D. Sampson, "Using Logic for Querying XML Data", Web Powered Databases, D. Taniar and W. Rahayu (eds.), Ch. 1, pp. 1-35, IDEA-Group Publishing, 2003. http%3A%2F%2Fwww%2Ecsd%2Eauth%2Egr%2F%257Elpis%2Fsystems%2Fx%2Ddevice%2Ehtml X%2DDEVICE http%3A%2F%2Fwww%2Eidea%2Dgroup%2Ecom%2Fbooks%2Fdetails%2Easp%3Fid%3D4059 XML query language XML repository Deductive Object-Oriented Databases A Knowledge Based Approach on Educational Metadata Use 3 8th Panhellenic Conference on Informatics, PCI 2001. Nicosia, Cyprus, November 8-10, 2001, Revised Selected Papers Springer-Verlag Y. Manolopoulos, S. Evripidou and A. Kakas LNCS 2563 2003 16 201-216 <p>One of the most rapidly evolving e-services is e-Learning, that is, the creation of advanced educational resources that are accessible on-line and, potentially, offer numerous advantages over the traditional ones like intelligent access, interoperability between two or more educational resources and adaptation to the user. The driving force behind these approaches is the definition of the various standards about educational metadata, that is, data describing learning resources, the learner, assessment results, etc. The internal details of systems that utilize these metadata is still an open issue since these efforts are primarily dealing with "what" and not "how".</p> <p>Under the light of these emerging efforts, we present CG-PerLS, a knowledge based approach for organizing and accessing educational resources. CG-PerLS is a model of a web portal for learning objects that encodes the educational metadata in the Conceptual Graph knowledge representation formalism, and uses related inference techniques to provide advanced, personalized functionality. The model allows learning resource creators to manifest their material, client-side learners to access these resources in a way tailored to their individual profile and educational needs, and dynamic course generation based on fine or coarse grained educational resources.</p> kokkoras-epy8post.pdf Ratio: 31/104 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Flink%2Easp%3Fid%3Djg1kkdf5w46hjm8q e-Learning educational metadata Conceptual Graphs The GRT Planner: New Results 2 Proc. ECAI 2000 Workshop on Local Search Techniques for Planning and Scheduling 2000 0 I. Refanidis and I. Vlahavas, "The GRT Planner: New Results", Proc. Workshop on Local Search Techniques for Planning and Scheduling, ECAI 2000, Berlin, 2000. Framework for Multi - Criteria Plan Evaluation in Heuristic State-Space Planning 2 Proc. IJCAI'01 Workshop on planning with Resourses 2001 0 I. Refanidis and I. Vlahavas, "Framework for Multi - Criteria Plan Evaluation in Heuristic State-Space Planning", Proc. IJCAI-01 Workshop on planning with Resourses, Seatle, Washington, 2001. Towards Adaptive Heuristic Planning through Machine Learning 2 Proc. 21st Workshop of the UK Planning and Scheduling Special Interest Group (PlanSIG '02) 2002 0 In domain independent heuristic planning there is a number of planning systems with very good performance on some problems and very poor on others. Few attempts have been made in the past to explain this phenomenon. In this paper we use machine learning techniques to discover knowledge hidden in the dynamics of the planning process that would relate specific characteristics of a planning problem with specific properties of a planning system that lead to good or bad performance. By this, we aim at shedding light to some of the dark areas of heuristic planning and develop an adaptive planner that would be able to optimize its configuration according to the problem at hand. hap.pdf D. Vrakas, G. Tsoumakas and I. Vlahavas, "Towards adaptive Heuristic Planning through Machine Learning", Proc. 21st Workshop of the UK Planning and Scheduling Special Interest Group, PlanSIG 2002, Delft, The Netherlands, November 2002. Cultures in Negotiation: Teachers' Acceptance/Resistance Attitudes Considering the Infusion of Technology into Schools 1 Computers and Education Elsevier 41 (1) 2003 0 19-37 A teachers’ training project, employing teacher-mentored in-school training approach, has been recently initiated in Greek secondary education for the introduction of Information and Communication Technology (ICT) into the classroom. Data resulting from this project indicate that although teachers express considerable interest in learning how to use technology they need consistent support and extensive training in order to consider themselves able for integrating it into their instructional practice. Teachers are interested in using ICT (1) to attain a better professional profile, and (2) to take advantage of any possible learning benefits offered by ICT but always in the context of the school culture. They are willing to explore open and communicative modes of ICT-based teaching whenever school objectives permit, otherwise they appear to cautiously adapt the use of ICT to the traditional teacher-centered mode of teaching (strongly connected to the established student examination system). Teachers’ attitude to adapt ICT mode of use is supported by research evidence that emphasize the situational character of knowledge and expertise. Authors discuss the view that introducing ICT into schools can be understood as initiating a “negotiation” process between cultures and the way that technological tools are used reflects school “single context” epistemological stance. S. Demetriadis, A. Barbas, A. Molohides, G. Palaigeorgiou, D. Psillos, I. Vlahavas, I. Tsoukalas, and A. Pomportsis, "Cultures in Negotiation: Teachers' Acceptance/Resistance Attitudes Considering the Infusion of Technology into Schools", Computers and Education, Elsevier (to appear). (Impact Factor 0.571). Learning Rules for Adaptive Planning 2 Proc. 13th International Conference on Automated Planning and Scheduling (ICAPS '03) 2003 10 82-91 This paper presents a novel idea, which combines Planning, Machine Learning and Knowledge-Based techniques. It is concerned with the development of an adaptive planning system that can fine-tune its planning parameters based on the values of specific measurable characteristics of the given planning problem. Adaptation is guided by a rule-based system, whose knowledge has been acquired through machine learning techniques. Specifically, the algorithm of classification based on association rules was applied to a large dataset produced by results from experiments on a large number of problems used in the three AIPS Planning competitions. The paper presents experimental results with the adaptive planner, which demonstrate the boost in performance of the planning system. icaps03.pdf D. Vrakas, G Tsoumakas, N. Bassiliades and I. Vlahavas, "Learning Rules for Adaptive Planning", submitted for presentation on the 13th International Conference on Automated Planning and Scheduling, AICAPS 03, Trento, Italy, June 2003 (accepted for presentation). (acceptance ratio 30/98, 1 : 3.3) Trento, Italy, June 2003 planning and learning domain-independent classical planning machine learning knowledge based systems Controlling Performance Degradation of Multistage Interconnection Networks with Non-uniform Traffic 1 International Journal of Modeling and Simulation Vol. 19 (3) 1999 0 A. Pomportsis and I. Vlahavas, "Controlling Performance Degradation of Multistage Interconnection Networks with Non-uniform Traffic", International Journal of Modeling and Simulation, vol. 19(3). 1999. Interbase-KB: A Knowledge-based Multidatabase Sustem for Data Warehousing 1 IEEE Transactions on Knowledge and Data Engineering Vol. 15, No. 5 2003 18 1188-1205 This paper describes the integration of a multidatabase system and a knowledge-base system to support the data-integration component of a Data Warehouse. The multidatabase system integrates various component databases with a common query language, however it does not provide capability for schema integration and other utilities necessary for Data Warehousing. The knowledge base system offers in addition a declarative logic language with second-order syntax but first-order semantics for integrating the schemes of the data sources into the warehouse and for defining complex, recursively defined materialized views. Furthermore, deductive rules are also used for cleaning, checking the integrity and summarizing the data imported into the Data Warehouse. The Knowledge Base System features an efficient incremental view maintenance mechanism that is used for refreshing the Data Warehouse, without querying the data sources. interbase-kb.pdf N. Bassiliades, I Vlahavas, A. Elmagarmid and E. Houstis, "Interbase(KB) : A Knowledge-based Multidatabase Sustem for Data Warehousing", IEEE Transactions on Knowledge and Data Engineering, vol. 15 (3), 2003. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdevice%2Ehtml DEVICE http%3A%2F%2Fcsdl%2Ecomputer%2Eorg%2Fcomp%2Ftrans%2Ftk%2F2003%2F05%2Fk1188abs%2Ehtm Multidatabase Schema Integration Data Warehouse Materialized View Knowledge Base System Deductive Rule Active Rule Applying a Distributed CLP Platform to a Workforce Management Problem 2 Proc. 12th Conference on Intelligent Systems Application to Power Systems (ISAP '03) (electronic proceedings) 2003 6 The work presented in this paper concerns the application of CSPCONS, a distributed constraint logic programming platform to a workforce management problem, namely the BT-250-118 problem instance. The latter is a well-studied problem instance in which the requirement is to create sequences of job locations for the technicians to visit (tours), so as to serve as many jobs as possible, minimizing at the same time the travel duration. CSPCONS is a logic programming platform that supports program execution over multiple Prolog processes with constraints. It offers channel-based communicating processes and TCP/IP communication and is based on the CSP model introduced by Hoare. This paper demonstrates its applicability to such complex Distributed Constraint Satisfaction problems. sakellariou_cspwom_isap03.pdf Το call λέει ότι θα δημοσιευτούν selected papers στο International Journal of Engineering Intelligent Systems. Οπότε δεν θα έχει proceedings ????????? http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpaper%5Fdetails%2Easp%3FpublicationID%3D75 D%2DWMS Greece distributed constraint logic programming traveling salesman problem Bi-Directional Heuristic Planning in State-Spaces 2 Proc. Workshop on Education of Informaticians and Industrial Mathematicians: New Challenges and Needs 2001 0 217-228 Clustering Classifiers for Knowledge Discovery from Physically Distributed Databases 1 Data and Knowledge Engineering Elsevier 49(3) 2004 20 223-242 Most distributed classification approaches view data distribution as a technical issue and combine local models aiming at a single global model. This however, is unsuitable for inherently distributed databases, which are often described by more than one classification models that might differ conceptually. In this paper we present an approach for clustering distributed classifiers in order to discover groups of similar classifiers and thus similar databases with respect to a specific classification task. We also show that clustering distributed classifiers as a pre-processing step for classifier combination enhances the achieved predictive performance of the ensemble. tsoumakas-dke-49.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Edatak%2E2003%2E09%2E002 Multi DBs Knowledge discovery Machine learning Metadata Aware Peer-to-Peer Agents for the e-Learner 2 A "Hercma03" Symposium on "AI Techniques in e-Learning" accepted for publication 2003 10 Metadata, being the first building block of the emerging semantic web, will enable computers to understand what the accessed information is all about, allowing in that way the building of advanced web services. The e-Learning domain is one of the first that is benefited by the definition, among others, of the Learning Object Metadata (LOM). From another perspective, the distributed nature of the web suggests that agent technologies will play a key role towards the use of these metadata. In this paper, we detail a Conceptual Graph (CG) binding of LOM (CG/LOM) and present the eLPA, a knowledge based, client side, metadata aware, peer-to-peer agent, that relies solely on the CG knowledge representation formalism. eLPA serves primarily as a personal memory agent for the e-learner. kokkoras-hercma03.pdf F. Kokkoras and I. Vlahavas, "Metadata Aware Peer-to-Peer Agents for the e-Learner", A "Hercma03" Symposium on "AI Techniques in e-Learning" Athens, Greece Educational Metadata Conceptual Graphs e-Learning LOM Aggregator: A Knowledge based Comparison Chart Builder for e-Shopping 3 Intelligent Knowledge-Based Systems: Business and Technology in the New Millennium Kluwer Academic Publishers C.T. Leondes Vol. 1: Knowledge-Based Systems, Ch. 6 2005 24 140-163 <p>Although there are many on-line stores, where potential e-shoppers can purchase desired products by visiting a few web pages, finding the right product to purchase is quite a tedious task. On-line stores usually offer a limited set of ill-described products, presented in a way that prevents side-by-side comparison shopping.</p> <p> In this chapter, we present a comparison chart building approach that is based on information extraction wrappers. The novelty of our approach consists of the usage of the Conceptual Graphs knowledge representation and reasoning formalism, which naturally supports both the wrapper induction and the wrapper evaluation tasks through the generalization, specialization and projection operations. In addition, the graphical representation of Conceptual Graphs makes them a proper technology for creating visual, wrapper management programming environments.</p> <p>Finally, we present the Aggregator, a prototype system based on our approach, that allows the user to quickly and visually train information extraction wrappers and use them to built comparison charts of arbitrary detail, gathering information about similar products from multiple known sites.</p> http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpaper%5Fdetails%2Easp%3FpublicationID%3D152 CG%2DWrappers http%3A%2F%2Fwww%2Ewkap%2Enl%2Fprod%2Fb%2F1%2D4020%2D7746%2D7 wrappers information extraction conceptual graphs A Knowledge-based Framework for Building Web Service Domains 2 9th Panhellenic Conference on Informatics (PCI'2003) 2003 15 331-345 This paper describes a knowledge-based framework, called SWIM, for building Web Service Domains, which are collections or communities of related Web Services that are mediated and/or aggregated by a single Web Service, called the Mediator Service that functions as a proxy for them. When a requestor sends a message to the Mediator Service our system will select one or more of the Web Services to dispatch the message and will fuse the results returned by the selected services. The selection of Web services and the algorithm for fusing the results is defined by the administrator of the Service Domain using a declarative rule language, called X-DEVICE. SWIM system offers services for registering new Web Services and Service Domains. The main advantage of the SWIM system, compared to similar proposed approaches is that it allows the easy definition of arbitrary service selection strategies using a logic-based language. Furthermore, it goes beyond the mere conditional re-routing of Web Service requests by allowing combination of results of multiple Web Services leading to a simple logic-based form for Web Service composition. pci9-bassiliades.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fswim%2Ehtml SWIM Thessaloniki, Greece, November 2003 http%3A%2F%2Fepy9%2Ecsd%2Eauth%2Egr Web Service Service Domain Deductive Object Oriented Database XML Query Language Information Fusion A Knowledge-based Web Information System for the Federation of Distributed Classifiers 3 Web Information Systems Idea-Group Publishing D. Taniar and W. Rahayu Chapter 8 2004 38 271-308 This chapter presents the design and development of WebDisC, a knowledge-based Web information system for the fusion of classifiers induced at geographically distributed databases. The main features of our system are: i) a declarative rule language for classifier selection that allows the combination of syntactically heterogeneous distributed classifiers, ii) a variety of standard methods for fusing the output of distributed classifiers, iii) a new approach for clustering classifiers in order to deal with the semantic heterogeneity of distributed classifiers, detect their interesting similarities and differences and enhance their fusion and iv) an architecture based on the Web services paradigm that utilizes the open and scalable standards of XML and SOAP. tsoumakas-wis-chapter8.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fwebdisc%2Ehtml WebDisC http%3A%2F%2Fwww%2Eidea%2Dgroup%2Ecom%2Fbooks%2Fdetails%2Easp%3Fid%3D4307 Capturing RDF Descriptive Semantics in an Object Oriented Knowledge Base System 2 Proc. 12th Int. WWW Conf. (WWW2003) ACM CDROM (Poster) 2003 0 In this paper, we present a deductive object-oriented knowledge base system, called R-DEVICE, which imports RDF data into the CLIPS production rule system as objects and uses a deductive rule language for querying and reasoning about them. In our model properties of resources are not scattered across several triples as in most other RDF storage and querying systems, resulting in increased query per-formance due to less joins. R-DEVICE features a powerful deductive rule language which is able to express arbitrary queries both on the RDF schema and data, including generalized path expressions, strati-fied negation, aggregate, grouping, and sorting, functions, mainly due to the second-order syntax of the rule language which is efficiently translated into sets of first-order logic rules using metadata. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fr%2Ddevice%2Ehtml +R%2DDEVICE Budapest, Hungary, May 2003 http%3A%2F%2Fwww2003%2Eorg%2Fcdrom%2Fpapers%2Fposter%2Fp277%2Fp277%2Dnbassili%2Ehtml RDF Object Model CLIPS Descriptive Semantics Deductive Rules Production Rules R-DEVICE: An Object-Oriented Knowledge Base System for RDF Metadata 1 International Journal on Semantic Web and Information Systems Idea Group Amit Sheth , Miltiadis D. Lytras Vol. 2, No. 2 2006 61 24-90 In this paper we present R-DEVICE, a deductive object-oriented knowledge base system for reasoning over RDF metadata. R-DEVICE imports RDF documents into the CLIPS production rule system by transforming RDF triples into COOL objects and uses a deductive rule language for reasoning about them. R-DEVICE is based on an OO RDF data model, different than the established triple-based model, which maps resources to objects and encapsulates properties inside resource objects, as traditional OO attributes. In this way, fewer joins are required to access the properties of a single resource resulting in better inferencing/querying performance, as it is experimentally shown in the paper. Furthermore, RDF can interoperate seamlessly with other web data models and languages. The descriptive semantics of RDF may call for dynamic redefinitions of resource classes, which are handled by R-DEVICE effectively. Furthermore, R-DEVICE features a pow-erful deductive rule language for reasoning on top of RDF metadata. The rule language includes features such as normal and generalized path expressions, stratified negation, aggregate, grouping, and sorting, func-tions. The rule language supports a second-order syntax, which is efficiently translated into sets of first-order logic rules using metadata, where variables can range over classes and properties, so that reasoning over the RDF schema can be made. Users can define views which are materialized and incrementally main-tained by translating deductive rules into CLIPS production rules that preserve truth. Users can choose be-tween an OPS5/CLIPS-like and a RuleML-like syntax. Finally, users can define and use functions through the CLIPS host language. ijswis-r-device.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fr%2Ddevice%2Ehtml R%2DDEVICE http%3A%2F%2Fwww%2Eidea%2Dgroup%2Ecom%2Farticles%2Fdetails%2Easp%3Fid%3D6106 RDF Object Data Model CLIPS Descriptive Semantics Deductive Rules Production Rules Generalized Path Expressions Aggregation Materialized Views RuleML Web Service Composition Using a Deductive XML Rule Language 1 Distributed and Parallel Databases Kluwer Ahmed K. Elmagarmid Vol. 17, No. 2 2005 35 135-178 This paper describes a knowledge-based Web Service composition system, called SWIM, which is based on the Service Domain model. Service Domains are communities of related Web Services that are mediated by a single Web Service, called the Mediator Service, which functions as a proxy for them. When a requestor sends a message to the Mediator Service one or more of the related Web Services are selected to dispatch the message and the results returned are aggregated to a single answer to the requestor. Mediator Services can be further composed to more complex Mediator Services that combine several selection and aggregation algorithms among many heterogeneous web services. The system utilizes the X DEVICE deductive XML rule language for defining complex algorithms for selecting registered web services, combining the results, and synchronizing the workflow of information among the combined web services in a declarative way. In the paper, we demonstrate the flexibility and expressibility of our approach for composing Web Services using several e-business examples, covering most of the workflow patterns found in a comprehensive workflow management system [2]. DAPD789-03.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fswim%2Ehtml SWIM http%3A%2F%2Fspringerlink%2Emetapress%2Ecom%2Fopenurl%2Easp%3Fgenre%3Darticle%26id%3Ddoi%3A10%2E1007%2Fs10619%2D004%2D0087%2Dz Web Service Composition Service Domain Deductive Object Oriented Database XML Rule Language Periodicity Mining in Industrial Data: A Real World Example on Power Data 2 First International Conference for Mathematics and Informatics for Industry (MII2003) 2003 12 This article presents an approach for detecting weak periodicities in power demand databases. The method attacks the problems in terms of data mining, where the size of the data along with the inherent notion of noise and randomness are crucial factors in the data manipulation process. We target the issue of pe-riodicity, aiming to provide the user with useful time-oriented knowledge about the nature of the data. We conducted experiments over real world power demand data and we exhibit the results. Thessaloniki Using the k-nearest problems for adaptive multicriteria planning 2 Proc. 3rd Hellenic Conference on Artificial Intellligence (SETN '04) Springer-Verlag G. Vouros and T. Panayiotopoulos LNAI 3025 2004 10 132-141 This paper concerns the design and development of an adaptive planner that is able to adjust its parameters to the characteristics of a given problem and to the priorities set by the user concerning plan length and planning time. This is accomplished through the implementation of the k nearest neighbor machine learning algorithm on top of a highly adjustable planner, called HAP. Learning data are produced by running HAP offline on several problems from multiple domains using all value combinations of its parameters. When the adaptive planner HAPNN is faced with a new problem, it locates the k nearest problems, using a set of measurable problem characteristics, retrieves the performance data for all parameter configurations on these problems and performs a multicriteria combination, with user-specified weights for plan length and planning time. Based on this combination, the configuration with the best performance is then used in order to solve the new problem. Comparative experiments with the statistically best static configurations of the planner show that HAPNN manages to adapt successfully to unseen problems, leading to an increased planning performance. tsoumakas-setn04.pdf Samos, Greece http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fopenurl%2Easp%3Fgenre%3Darticle%26issn%3D0302%2D9743%26volume%3D3025%26spage%3D132 Machine Learning Planning Prediction A Graphical Interface for Adaptive Planning 2 Proceedings of the Doctoral Consortium of the 13th International Conference on Automated Planning and Scheduling AAAI 2003 0 This paper describes a friendly graphical interface for HAP, a rule-configurable planning system, which automatically adapts to each problem, in order to achieve best performance. HAP analyzes the problem and uses a rule system in order to make the most appropriate choices for planning. The graphical interface enables the user to use and even interfere in the process of fine tuning the planning system. Furthermore, the interface allows the user to manually configure and experiment with the system and also to define its own domains and problems through an easy-to-use visual tool. dc.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fvitaplan%2F ViTAPlan+web+site Trento, Italy Planning PDDL Tools Graphical Interfaces Communicating Sequential Processes for Distributed Constraint Solving 1 Information Sciences Elsevier 176(5) 2006 31 490-521 CSPCONS is a programming language that supports program execution over multiple Prolog processes with constraints. The language is an extended version of CSP-II, a version of Prolog that supports channel-based communicating processes and TCP/IP communication, that is based on the CSP model introduced by Hoare. CSPCONS inherits all the advanced features of CSP-II and extends it by introducing constraint solving capabilities to the processes. In CSPCONS each Prolog process has one or more solvers attached and each solver is independent from the others, following the original CSP-II model, thus resulting to a communicating sequential constraint logic programming system. Such a model can facilitate greatly the implementation of distributed CLP applications. This paper describes the original CSP-II system along with details of the extensions that resulted to the CSPCONS system and presents an example demonstrating the applicability of the system to distributed constraint satisfaction problems. Distributed Singleton Consistency 1 Journal of Experimental and Theoretical Artificial Intelligence Taylor and Francis 16(2) 2004 0 107-124 Distributed constraint satisfaction has drawn much attention in the past years, with a number of algorithms proposed to tackle the problem. Research in the area has followed two directions: distributed search techniques and distributed filtering techniques. This paper presents a new distributed filtering algorithm, named Distributed Singleton Arc Consistency (DSAC), which is based on the singleton consistency algorithm. DSAC is a parallel coarse grain filtering algorithm aiming at improving the performance of singleton consistency by distributing the work to be done to a number of agents. The current paper presents the basic idea behind the algorithm and two versions of it that employ different communication policies along with experimental results obtained on a set of random binary CSP problems. ViTAPlan: A Visual Tool for Adaptive Planning 2 Proceedings of the 9th Panhellenic Conference on Informatics 2003 0 This paper presents a friendly visual tool for HAP, a rule-configurable planning system, which automatically adapts to each problem, in order to achieve best performance. HAP analyzes the problem and uses a rule system in order to configure the planning parameters in a way that best suites the morphology of the problem. The visual tool enables the user to use the planning system, get advice from the built-in rule system and even interfere with it. ViTAPlan also contains a visual designer, based on the Planning Domain Definition Language, that enables the user to create new planning domains and problems in a graphical way and get visual representations of existing ones. Furthermore the tool contains a module that simulates the execution of the plan and illustrates the changes in the world, which follow the application of each action in the plan. vitaplan.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fvitaplan%2F ViTAPlan+web+site Thessaloniki, Greece Visual Tools Planning Modelling Information Extraction Wrappers with Conceptual Graphs 2 Proc. (2nd Volume) 3rd Panhellenic Conference on Artificial Intelligence (SETN'04) Zitis Publications ISBN 960-431-910-8 2004 10 In this paper, we propose the use of the Conceptual Graphs knowledge representation and reasoning formalism to model information extraction wrappers (CG-Wrappers). An information extraction wrapper is a mapping that populates a data repository with implicit objects that exist inside a given web page. Creating a wrapper, usually involves some training by which the wrapper learns to identify the desired information based, mainly, on the surrounding HTML elements. In the paper, we demonstrate how the generalization, specialization and projection operations of the Conceptual Graph theory naturally support both the wrapper induction and the wrapper evaluation tasks. The proposed modeling approach is flexible enough to support wrapper reuse, enabling us in that way to create more complex wrappers. kokkoras-setn2004.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpaper%5Fdetails%2Easp%3FpublicationID%3D137 Aggregator 5-8 May 2004, Samos, Greece information extraction conceptual graphs personalization wrapper induction Document Object Model Lazy Adaptive Multicriteria Planning 2 Proc. 16th European Conference on Artificial Intelligence, ECAI 2004 IOS Press R. Lopez de Mantaras and L. Saitta 2004 5 693-697 This paper describes a learning system for the automatic configuration of domain independent planning systems, based on measurable features of planning problems. The purpose of the Lazy Adaptive Multicriteria Planning (LAMP) system is to configure a planner in an optimal way, concerning two quality metrics (i.e. execution speed and plan quality), for a given problem according to user-specified preferences. The training data are produced by running the planner under consideration on a set of problems using all possible parameter configurations and recording the planning time and the plan length. When a new problem arises, (LAMP) extracts the values for a number of domain-expert specified problem features and uses them to identify the k nearest problems solved in the past. The system then performs a multicriteria combination of the performances of the retrieved problems according to user-specified weights that specify the relative importance of the quality metrics and selects the configuration with the best score. Experimental results show that LAMP improves the performance of the default configuration of two already well-performing planning systems in a variety of planning problems. tsoumakas-ecai04.pdf Valencia, Spain Institutional efforts in tertiary education for promoting the integration of technology in instruction: a case study 2 Proc. 5th International Conference on Information and Communication Technologies in Education (ICICTE) 2004 0 July 1-3, 2004, Samos, Greece Machine Learning for Adaptive Planning 3 Intelligent Techniques for Planning IDEA GROUP PUBLISHING I. Vlahavas and D. Vrakas Ch. 3 2005 49 90-120 This chapter is concerned with the enhancement of planning systems using techniques from Machine Learning in order to automatically configure their planning parameters according to the morphology of the problem in hand. It presents two different adaptive systems that set the planning parameters of a highly adjustable planner based on measurable characteristics of the problem instance. The planners have acquired their knowledge from a large data set produced by results from experiments on many problems from various domains. The first planner is a rule-based system that employs propositional rule learning to induce knowledge that suggests effective configuration of planning parameters based on the problem's characteristics. The second planner employs instance-based learning in order to find problems with similar structure and adopt the planner configuration that has proved in the past to be effective on these problems. The validity of the two adaptive systems is assessed through experimental results that demonstrate the boost in performance in problems of both known and unknown domains. Comparative experimental results for the two planning systems are presented along with a discussion of their advantages and disadvantages. itp.pdf http%3A%2F%2Fwww%2Eidea%2Dgroup%2Ecom%2Fbooks%2Fdetails%2Easp%3Fid%3D4496 Domain Independent Planning Machine Learning Knowledge Engineering Inter-Transaction Association Rules Mining for Rare Events Prediction 2 Proc. 3rd Hellenic Conference on Artificial Intellligence (SETN '04) 2004 10 Rare events prediction is a very interesting and critical issue that has been approached within various contexts by research areas, such as statistics and machine learning. Data mining has provided a set of tools to treat this prob-lem when the size as well as the inherent features of the data, such as noise, randomness and special data types, become an issue for the traditional methods. Transaction databases that contain sets of events require special approaches in order to extract valuable temporal knowledge. Sequential analysis aims to dis-cover patterns or rules describing the temporal structure of data. In this paper we propose an approach that extends sequential analysis to predict rare events in transaction databases. We utilize the framework of inter-transaction associa-tion rules, which associate events across a window of transactions. The pro-posed algorithm produces rules for the accurate and timely prediction of a user-specified rare event, such as a network intrusion or an engine failure. 076-Berberidis-Angelis-Vlahavas-SETN04.pdf Samos, Greece Data Mining Prediction Effective Voting of Heterogeneous Classifiers 2 Proc. European Conference on Machine Learning, ECML 04 Jean-Francois Boulicaut, Floriana Esposito, Fosca Giannoti, Dino Pedreschi LNAI 3201 2004 12 465-476 This paper deals with the combination of classification models that have been derived from running different (heterogeneous) learning algorithms on the same data set. We focus on the Classifier Evaluation and Selection (ES) method, that evaluates each of the models (typically using 10-fold cross-validation) and selects the best one.We examine the performance of this method in comparison with the Oracle selecting the best classifier for the test set and show that 10-fold cross-validation has problems in detecting the best classifier. We then extend ES by applying a statistical test to the 10-fold accuracies of the models and combining through voting the most significant ones. Experimental results show that the proposed method, Effective Voting, performs comparably with the state-of-the-art method of Stacking with Multi-Response Model Trees without the additional computational cost of meta-training. tsoumakas-ecml2004.pdf Pisa, Italy Voting Multiple Classifier Systems Ensemble Methods Classification Web Services for Adaptive Planning 2 Proc. ECAI04 Workshop on Planning and Scheduling: Bridging Theory to Practice IOS Press D. Borrajo 2004 10 This paper presents the design and development of an adaptive planning system using the technology of Web services. The Web-based adaptive planning system consists of two modules that can work independently. The first one is called HAP-WS and is the Web service interface to the domain independent planner HAP (Highly Adjustable Planner) that can be customized through the adjustment of several parameters, either manually or automatically. In the manual mode, the user itself adjusts planner parameters giving explicitly the values. In the automatic mode, the second subsystem, called LAMP-WS, computes the values of the planning parameters of HAP. LAMP-WS is the Web service interface to the learning system LAMP (Lazy Adaptive Multicriteria Planning) that can automatically configure a planning system using instance-based learning on past performance data of that system. The two subsystems are implemented as independent Web services, which can be used stand-alone and reside in different servers in potentially different geographical locations. tsoumakas-ecai04workshop.pdf Valencia Spain Mining for weak periodic signals in time series databases 1 Intelligent Data Analysis IOS Press Dr. F. Famili 9(1) 2005 0 Periodicity is a particularly interesting feature, which is often inherent in real world time series data sets. In this article we propose a data mining technique for detecting multiple partial and approximate periodicities. Our approach is exploratory and follows a fil-ter/refine paradigm. In the filter phase we introduce an autocorrelation-based algorithm that produces a set of candidate partial periodicities. The algorithm is extended to capture ap-proximate periodicities. In the refine phase we effectively prune invalid periodicities. We conducted a series of experiments with various real-world data sets to test the performance and verify the quality of the results. (to appear) Time Series Mining Periodicity Detection DR-DEVICE: A Defeasible Logic System for the Semantic Web 2 Second International Workshop on Principles and Practice of Semantic Web Reasoning (PPSWR04) Springer-Verlag Sebastian Schaffert LNCS 3208 2004 15 134-148 This paper presents DR-DEVICE, a system for defeasible reasoning on the Web. Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is, among others, useful for ontology integration, where conflicting information arises naturally; and for the modeling of business rules and policies, where rules with exceptions are often used. In this paper we describe these scenarios in more detail along with the implementation of the DR-DEVICE system, which is capable of reasoning about RDF data over multiple Web sources using defeasible logic rules. The system is implemented on top of CLIPS production rule system and builds upon R-DEVICE, an earlier deductive rule system over RDF data that also supports derived attribute and aggregate attribute rules. Rules can be expressed either in a native CLIPS-like language, or in an extension of the OORuleML syntax. The operational semantics of defeasible logic are implemented through compilation into the generic rule language of R-DEVICE. The paper includes a use case of a semantic web broker that reasons defeasibly about renting apartments based on buyer's requirements (expressed RuleML defeasible logic rules) and seller's advertisements (expressed in RDF). PPSWR04-bassiliades.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE St Malo, France, 6-10 Sept. http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Flink%2Easp%3Fid%3Dk0r9w04mg0uh8h88 rules reasoning defeasible logic rule markup languages semantic brokering R-DEVICE: A Deductive RDF Rule Language 2 Third International Workshop on Rules and Rule Markup Languages for the Semantic Web (RuleML 2004) Springer-Verlag G. Antoniou, H. Boley LNCS 3323 2004 16 65-80 In this paper we present R-DEVICE, a deductive rule language for reasoning about RDF metadata. R-DEVICE includes features such as normal and generalized path expressions, stratified negation, aggregate, grouping, and sorting, functions. The rule language supports a second-order syntax, where variables can range over classes and properties. Users can define views which are materialized and incrementally maintained by translating deductive rules into a couple of CLIPS production rules. Users can choose between an OPS5/CLIPS-like or a RuleML-like syntax. R-DEVICE is based on a OO RDF data model, different than the established graph model, which maps resources to objects and encapsulates properties inside resource objects, as traditional OO attributes. In this way, less joins are required to access the properties of a single resource resulting in better inferencing/querying performance. The descriptive semantics of RDF may call for dynamic re-definitions of resource classes and objects, which are handled by R-DEVICE effectively. RuleML04-bassiliades1.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fr%2Ddevice%2Ehtml R%2DDEVICE Hiroshima, Japan, 8 Nov. 2004 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Flink%2Easp%3Fid%3Dj5bfdafdeddyngtb RDF Deductive Rules Production Rules RuleML A Defeasible Logic Reasoner for the Semantic Web 2 Third International Workshop on Rules and Rule Markup Languages for the Semantic Web (RuleML 2004) Springer-Verlag G. Antoniou, H. Boley LNCS 3323 2004 16 49-64 Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is, among others, useful for ontology integration, where conflicting information arises naturally; and for the modeling of business rules and policies, where rules with exceptions are often used. This paper describes these scenarios in more detail, and reports on the implementation of a system for defeasible reasoning on the Web. The system is called DR-DEVICE and is capable of reasoning about RDF metadata over multiple Web sources using defeasible logic rules. The system is imple-mented on top of CLIPS production rule system and builds upon R-DEVICE, an earlier deductive rule system over RDF metadata that also supports derived attribute and aggregate attribute rules. Rules can be expressed either in a native CLIPS-like language, or in an extension of the OO-RuleML syntax. The opera-tional semantics of defeasible logic are implemented through compilation into the generic rule language of R-DEVICE. The paper includes a complete exam-ple of a semantic web broker that reasons about apartment renting. RuleML04-bassiliades2.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE Hiroshima, Japan, 8 Nov. 2004 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Flink%2Easp%3Fid%3D6ah5yawwd8y2ck3a RDF rules reasoning defeasible logic rule markup languages semantic brokering PREVENT: An algorithm for mining inter-transactional patterns for the prediction of rare events 2 2nd European Starting AI Researcher Symposium (STAIRS' 04) IOS Press 2004 9 128-136 In this paper we propose a data mining technique for the efficient prediction of rare events, such as heat waves, network intrusions and engine failures, using inter transactional patterns. Data mining is a research area that attempts to assist the decision makers with a set of tools to treat a wide range of real world problems that the traditional statistical and mathematical approaches are not enough in terms of ef-ficiency and computational performance. Transaction databases, such as the ones in this paper that contain sets of events, require special approaches in order to extract valuable temporal knowledge. We utilize the framework of inter-transaction associa-tion rules, which associate events across a window of transactions. We propose an approach that extends sequential analysis to predict rare events in transaction data-bases. We formulate the problem of rare events prediction and we propose PREVENT, an algorithm that produces inter-transactional patterns for the fast and accurate prediction of a user-specified rare event. Finally, we provide experimental results and suggest some ideas for future research. STAIRS-04-Berberidis.pdf Valencia, Spain Data Mining Rare events prediction Association Rules Sequence Analysis DR-DEVICE: A Defeasible Logic RDF Rule Language 2 3rd International Semantic Web Conference (ISWC2004), Demonstrations Program 2004 5 Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is, among others, useful for ontology integration, where conflicting information arises naturally; and for the modeling of business rules and policies, where rules with exceptions are often used. In this demonstration we pre-sent a prototype system for defeasible reasoning on the Web. The system is called DR-DEVICE and is capable of reasoning about RDF metadata over multiple Web sources using defeasible logic rules. The system is implemented on top of CLIPS production rule system and builds upon R-DEVICE, an earlier deductive rule system over RDF metadata that also supports derived attribute and aggregate attribute rules. Rules can be expressed either in a native CLIPS-like language, or in an extension of the OO-RuleML syntax. The operational semantics of defeasible logic are implemented through compilation into the generic rule language of R-DEVICE. This demonstration includes a complete use case of a semantic web broker that reasons about apartment renting. iswc04-demo.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE Hiroshima, Japan, 7-11 Nov. 2004 http%3A%2F%2Fiswc2004%2Esemanticweb%2Eorg%2Fdemos%2Findex%2Ehtml Rules Semantic Web Reasoning RDF Defeasible Logic Rule Markup Languages A System for Automated Agent Negotiation with Defeasible Logic-Based Strategies – Preliminary Report 2 Third International Workshop on Rules and Rule Markup Languages for the Semantic Web (RuleML 2004) Springer-Verlag G. Antoniou, H. Boley LNCS 3323 2004 9 205-213 This paper reports on a system for automated agent negotiation. The negotiation strategies are expressed in defeasible logic, and are applied using the implemented reasoning system DR-DEVICE. The overall system architecture is described, and a particular 1-1 negotiation scenario is presented in detail. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE Hiroshima, Japan, November 8, 2004 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Flink%2Easp%3Fid%3Dc7k2t2nrk60ed2ry Agent Negotiation Defeasible Logic Rule System HAPrc: An Automatically Configurable Planning System 1 AI Communications 18 (1) 2005 44 41-60 This paper presents an adaptive planning system, called HAPRC, which automatically fine-tunes its planning parameters according to the morphology of the problem in hand, through a combination of Planning, Machine Learning and Knowledge-Based techniques. The adaptation is guided by a rule-based system that sets planner configuration parameters based on measurable characteristics of the problem instance. The knowledge of the rule system has been acquired through a rule induction algorithm. Specifically, the approach of propositional rule learning was applied to a dataset produced by results from experiments on a large number of problems from various domains, including those used in the three International Planning Competitions. The improvement of the adaptive system over the original planner is assessed through thorough experiments in problems of both known and unknown domains. vrakas-aicom18.pdf DR-BROKERING – A Defeasible Logic-Based System for Semantic Brokering 2 IEEE International Conference on E-Technology, E-Commerce and E-Service IEEE W.K.W. Cheung, J. Hsu 2005 8 414-417 Electronic Brokering, is a good candidate for taking up Semantic Web technology. In this paper we study the brokering and matchmaking problem that is, how a requester’s requirements and preferences can be matched against a set of offerings collected by a broker. The proposed solution uses the Semantic Web standard of RDF to represent the offerings, and a deductive logical language, based on non-monotonic reasoning, for expressing the requirements and preferences. We motivate and explain the approach we propose, and report on a prototypical implementation exhibiting the described functionality, in JADE agent environment. EEE05-a.pdf 29/3 - 1/4/2005, Hong Kong, China http%3A%2F%2Fdoi%2Eieeecomputersociety%2Eorg%2F10%2E1109%2FEEE%2E2005%2E60 Semantic Brokering Intelligent Agents Semantic Web Defeasible Logic RDF DR-NEGOTIATE – A System for Automated Agent Negotiation with Defeasible Logic-Based Strategies 2 IEEE International Conference on E-Technology, E-Commerce and E-Service IEEE W.K.W. Cheung, J. Hsu 2005 8 44-49 This paper reports on a system for automated agent negotiation. It uses the JADE agent framework, and its major distinctive feature is the use of declarative negotiation strategies. The negotiation strategies are expressed in a declarative rules language, defeasible logic and are applied using the implemented defeasible reasoning system DR-DEVICE. The choice of defeasible logic is justified. The overall system architecture is described, and a particular negotiation case is presented in detail. EEE05-b.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE 29/3 - 1/4/2005, Hong Kong, China http%3A%2F%2Fdoi%2Eieeecomputersociety%2Eorg%2F10%2E1109%2FEEE%2E2005%2E61 Agent Negotiation Semantic Web Defeasible Logic RDF Predicting Missing Parts in Time Series Using Uncertainty Theory 2 Biological and Medical Data Analysis: 5th International Symposium, ISBMDA 2004 Springer-Verlag J. M. Barreiro, F. Martin-Sanchez, V. Maojo, et al. 3337 2004 9 313-321 As extremely large time series data sets grow more prevalent in a wide variety of applications, including biomedical data analysis, diagnosis and monitoring systems and exploratory data analysis in scientific and business time series, the need of developing efficient analysis methods is high. However, essential preprocessing algorithms are required in order to obtain positive results. The goal of this paper is to propose a novel algorithm that is appropriate for filling missing parts of time series. This algorithm, named FiTS (Filling Time Series), was evaluated over 11 congestive heart failure patientsrsquo ECGs (Electrocardiogram). Those patients using electronic microdevices with which were recording their ECGs and sending them via telephone to a home care monitoring system, over a period of 8 to 16 months. Randomly missing parts in each ECG were introduced in the initial ECG. As a result, FiTS had 100% of successfully completion with high reconstructed signal accuracy. Barcelona, Spain, November 18-19, 2004 http%3A%2F%2Fspringerlink%2Emetapress%2Ecom%2Flink%2Easp%3Fid%3Dke147vew4k606bfu Time Series Analysis A Semantic Brokering System for the Tourism Domain 1 Journal of Information Technology and Tourism, special issue on "Semantic Web Technologies and Applications" Vol. 7, No. 3-4 2005 20 183-200 The tourism industry is a good candidate for taking up Semantic Web technology. In this paper we study the brokering and matchmaking problem in the tourism domain, that is, how a requester’s requirements and preferences can be matched against a set of offerings collected by a broker. The proposed solution uses the Semantic Web standard of RDF to represent the offerings, and a deductive logical language for expressing the requirements and preferences. We motivate and explain the approach we propose, and report on a prototypical implementation exhibiting the described functionality in a multi-agent environment. JITT.pdf http%3A%2F%2Fwww%2Eingentaconnect%2Ecom%2Fcontent%2Fcog%2Fitt%2F2005%2F00000007%2FF0020003%2Fart00006 Brokering Rules Ontologies RDF Schema Eds. 4 Intelligent Techniques for Planning IDEA Group Publishing 2005 300 Automated Planning is the area of Artificial Intelligence that deals with problems in which we are interested in finding a sequence of steps (actions) to apply to the world in order to achieve a set of predefined objectives (goals) starting from a given initial state. In the past, planning has been successfully applied in numerous areas including robotics, space exploration, transportation logistics, marketing and finance, assembling parts, crisis management, etc. The history of Automated Planning goes back to the early 1960s with the General Problem Solver (GPS) being the first automated planner reported in literature. Since then, it has been an active research field with a large number of institutes and researchers working on the area. Traditionally, planning has been seen as an extension of problem solving and it has been attacked using adaptations of the classical search algorithms. The methods utilized by systems in the �classical� planning era (until mid-1990s), include state-space or plan-space search, hierarchical decomposition, heuristic and various other techniques developed ad-hoc. The classical approaches in Automated Planning presented over the past years were assessed on toy-problems, such as the ones used in the International Planning Competitions, that simulate real world situations but with too many assumptions and simplifications. In order to deal with real world problems, a planner must be able to reason about time and resources, support more expressive knowledge representations, plan in dynamic environments, evolve using past experience, co-operate with other planners, etc. Although the above issues are crucial for the future of Automated Planning, they have been recently introduced to the planning community as active research directions. However, most of them are also the subject of researchers in other AI areas, such as Constraint Programming, Knowledge Systems, Machine Learning, Intelligent Agents and others, and therefore the ideal way is to utilize the effort already put into them. This edited volume, Intelligent Techniques for Planning, consists of 10 chapters bringing together a number of modern approaches in the area of Automated Planning. These approaches combine methods from classical planning, such as the construction of graphs and the use of domain-independent heuristics, with techniques from other areas of Artificial Intelligence. The book presents in detail a number of state-of-the-art planning systems that utilize Constraint Satisfaction Techniques in order to deal with time and resources, Machine Learning in order to utilize experience drawn from past runs, methods from Knowledge Representation and Reasoning for more expressive representation of knowledge, and ideas from other areas, such as Intelligent Agents. Apart from the thorough analysis and implementation details, each chapter of the book also provides extensive background information about its subject and presents and comments on similar approaches done in the past. http%3A%2F%2Fwww%2Eidea%2Dgroup%2Ecom%2Fbooks%2Fdetails%2Easp%3Fid%3D4496 Selective Fusion of Heterogeneous Classifiers 1 Intelligent Data Analysis IOS Press 9(6) 2005 28 511-525 There are two main paradigms in combining different classification algorithms: Classifier Selection and Classifier Fusion. The first one selects a single model for classifying a new instance, while the latter combines the decisions of all models. The work presented in this paper stands in between these two paradigms aiming tackle the disadvantages and benefit from the advantages of both. In particular, this paper proposes the use of statistical procedures for the selection of the best subgroup among different classification algorithms and the subsequent fusion of the decision of the models in this subgroup with simple methods like Weighted Voting. Extensive experimental results show that the proposed approach, Selective Fusion, improves over simple selection and fusion methods, leading to performance comparable with the state-of-the-art heterogeneous classifier combination method of Stacking, without the additional computational cost and learning problems of meta-training. tsoumakas-ida9.pdf Biological Data Mining 3 Encyclopedia of Database Technologies and Applications Idea Group Publishing Laura C. Rivero, Jorge H. Doorn and Viviana E. Ferraggine 2005 7 35-41 [... ]Recently, the collection of biological data has been increasing at explosive rates due to improvements of existing technologies and the introduction of new ones such as the microarrays. These technological advances have assisted the conduct of large scale experiments and research programs. An important example is the Human Genome Project, that was founded in 1990 by the U.S. Department of Energy and the U.S. National Institutes of Health (NIH) and was completed in 2003 (U.S. Department of Energy Office of Science, 2004). A representative example of the rapid biological data accumulation is the exponential growth of GenBank (Figure 1), the U.S. NIH genetic sequence database. (National Center for Biotechnology Information, 2004). The explosive growth in the amount of biological data demands the use of computers for the organization, the maintenance and the analysis of these data. Biological_Data_Mining.pdf http%3A%2F%2Fwww%2Eidea%2Dgroup%2Ecom%2Fencyclopedia%2Fdetails%2Easp%3FID%3D4462 The+official+web+page data mining bioinformatics molecular biology genomics proteomics gene expression analysis microarray classification clustering gene selection Preface of Special Issue on Knowledge Discovery from Distributed Information Sources 1 Information Sciences Elsevier 155(3-4) 2003 2 179-180 vlahavas-is155.pdf Mining for Contiguous Frequent Itemsets in Transaction Databases 2 IEEE 3rd International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS'2005) IEEE 2005 7 679-685 Mining a transaction database for association rules is a particularly popular data mining task, which involves the search for frequent co-occurrences among items. One of the problems often encountered is the large number of weak rules extracted. Item taxonomies, when available, can be used to reduce them to a more usable volume. In this paper we introduce a new data mining paradigm, which involves the discovery of contiguous frequent itemsets. We formulate the problem of mining contiguous frequent itemsets in a transaction database and we present a level-wise algorithm for finding these itemsets. Contiguous frequent itemsets may contain important knowledge about the dataset, that can not be exposed by the use of classic association rule mining approaches. This knowledge may well include serious hints for the generation of a taxonomy for all or part of the items. Sofia, Bulgaria data mining market basket analysis frequent itemset mining association rule Improving the Accuracy of Classifiers for the Prediction of Translation Initiation Sites in Genomic Sequences 2 10th Panhellenic Conference on Informatics (PCI'2005) Springer-Verlag P. Bozanis and E.N. Houstis LNCS 3746 2005 11 426-436 The prediction of the Translation Initiation Site (TIS) in a genomic sequence is an important issue in biological research. Although several methods have been proposed to deal with this problem, there is a great potential for the improvement of the accuracy of these methods. Due to various reasons, including noise in the data as well as biological reasons, TIS prediction is still an open problem and definitely not a trivial task. In this paper we follow a three-step approach in order to increase TIS prediction accuracy. In the first step, we use a feature generation algorithm we developed. In the second step, all the candidate features, including some new ones generated by our algorithm, are ranked according to their impact to the accuracy of the prediction. Finally, in the third step, a classification model is built using a number of the top ranked features. We experiment with various feature sets, feature selection methods and classification algorithms, compare with alternative methods, draw important conclusions and propose improved models with respect to prediction accuracy. TIS_PCI05.pdf http%3A%2F%2Fwww%2Espringer%2Ede%2Fcomp%2Flncs%2Findex%2Ehtml Springer%2DVerlag Volos, Greece, 11-13 November On the Utility of Incremental Feature Selection for the Classification of Textual Data Streams 2 10th Panhellenic Conference on Informatics (PCI 2005) Springer-Verlag P. Bozanis and E.N. Houstis LNCS 3746 2005 11 338-348 In this paper we argue that incrementally updating the fea- tures that a text classification algorithm considers is very important for real-world textual data streams, because in most applications the distri- bution of data and the description of the classification concept changes over time. We propose the coupling of an incremental feature ranking method and an incremental learning algorithm that can consider differ- ent subsets of the feature vector during prediction (what we call a feature based classifier), in order to deal with the above problem. Experimental results with a longitudinal database of real spam and legitimate emails shows that our approach can adapt to the changing nature of streaming data and works much better than classical incremental learning algo- rithms. tsoumakas-pci2005b.pdf Volos, Greece, 11-13 November http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2Fb100702 Text Mining Text Classification Feature Based Classifiers Dynamic Feature Space Dynamic Feature Selection Data Streams Concept Drift A Graphical Rule Authoring Tool for Defeasible Reasoning in the Semantic Web 2 10th Panhellenic Conference on Informatics (PCI 2005) Springer-Verlag P. Bozanis and E.N. Houstis LNCS 3746 2005 11 Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is useful for many applications in the Semantic Web, such as policies and business rules, agent brokering and negotiation, ontology and knowledge merging, etc. However, the syntax of defeasible logic may appear too complex for many users. In this paper we present a graphical authoring tool for defeasible logic rules that acts as a shell for the DR-DEVICE defeasible reasoning system over RDF metadata. The tool helps users to develop a rule base using the OO-RuleML syntax of DR-DEVICE rules, by constraining the allowed vocabulary through analysis of the input RDF namespaces, so that the user does not have to type-in class and property names. Rule visualization follows the tree model of RuleML. The DR-DEVICE reasoning system is implemented on top of the CLIPS production rule system and builds upon an earlier deductive rule system over RDF metadata that also supports derived attribute and aggregate attribute rules. pci2005-bassiliades_etal.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE%2C+VDR%2DDEVICE Volos, Greece, 11-13 November Graphical Rule Editor Defeasible Logic Rule Markup Languages Semantic Web Reasoning RDF Schema A Defeasible Logic Reasoner for the Semantic Web 1 International Journal on Semantic Web and Information Systems IDEA Group Publishing Amit Sheth , Miltiadis D. Lytras Vol. 2, No. 1 2006 35 1-41 Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is, among others, useful for ontology integration, where conflicting information arises naturally; and for the modeling of business rules and policies, where rules with ex-ceptions are often used. This paper describes these scenarios in more detail, and reports on the implementation of a system for defeasible reasoning on the Web. The system is called DR-DEVICE and is capable of reasoning about RDF metadata over multiple Web sources using defeasible logic rules. The system is implemented on top of CLIPS production rule system and builds upon R-DEVICE, an earlier deductive rule system over RDF metadata that also supports derived attribute and aggregate attribute rules. Rules can be expressed either in a native CLIPS-like language, or in an extension of the OO-RuleML syntax. The operational semantics of defeasible logic are implemented through compilation into the generic rule language of R-DEVICE. The paper also presents a full semantic web broker example for apartment renting. ijswis05-dr-device.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE http%3A%2F%2Fwww%2Eidea%2Dgroup%2Ecom%2Farticles%2Fdetails%2Easp%3FID%3D5575 defeasible logic RDF reasoning rule markup languages rules semantic brokering Protein Classification with Multiple Algorithms 2 10th Panhellenic Conference on Informatics (PCI 2005) Springer-Verlag P. Bozanis and E.N. Houstis LNCS 3746 2005 10 448-456 Nowadays, the number of protein sequences being stored in central protein databases from labs all over the world is constantly increasing. From these proteins only a fraction has been experimentally analyzed in order to detect their structure and hence their function in the corresponding organism. The reason is that experimental determination of structure is labor-intensive and quite time-consuming. Therefore there is the need for automated tools that can classify new proteins to structural families. This paper presents a comparative evaluation of several algorithms that learn such classification models from data concerning patterns of proteins with known structure. In addition, several approaches that combine multiple learning algorithms to increase the accuracy of predictions are evaluated. The results of the experiments provide insights that can help biologists and computer scientists design high-performance protein classification systems of high quality tsoumakas-pci2005a.pdf Volos, Greece, 11-13 November A Visualization Environment for Planning 1 International Journal on Artificial Intelligence Tools World Scientific Vol. 14(6) 2005 24 975-998 This article presents ViTAPlan-2, a visual tool for adaptive planning that is build on top of HAPRC, a rule-configurable planning system, which automatically adapts to each problem, in order to achieve best performance. Apart from HAPRC, ViTAPlan can be interfaced with any other planning system that supports the PDDL language. More than just being a user friendly environment for executing the underlying planner, the tool serves as a unified planning environment for encoding a new problem problem, solving it, visualizing the solution and monitoring its execution on a simulation of the problem’s word. The tool consists of various sub-systems, each one accompanied by a graphical interface, that collaborate with each other and assist the user, whether he is a knowledge engineer, a domain expert, an academic or even an end user in industry, to carry out complex planning tasks. IJAIT05-revised.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fvitaplan%2Findex%2Ehtml ViTA+Plan Automated Planning Graphical Interfaces Machine Learning Knowledge Engineering A Deductive Semantic Brokering System 2 9th International Conference on Knowledge-Based Intelligent Information and Engineering Systems (KES 2005) Springer-Verlag R. Khosla, R. J. Howlett, L. C. Jain LNCS 3682 2005 0 746-752 In this paper we study the brokering and matchmaking problem in the tourism domain, that is, how a requester’s requirements and preferences can be matched against a set of offerings collected by a broker. The proposed solution uses the Semantic Web standard of RDF to represent the offerings, and a deductive logical language for expressing the requirements and preferences. We motivate and explain the approach we propose, and report on a prototypical implementation exhibiting the described functionality in a multi-agent environment. Melbourne, Australia, September 14-16 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fopenurl%2Easp%3Fgenre%3Darticle%26id%3Ddoi%3A10%2E1007%2F11552451%5F102 Brokering Rules Ontologies RDF Schema A Visual Environment for Developing Defeasible Rule Bases for the Semantic Web 2 Proc. International Conference on Rules and Rule Markup Languages for the Semantic Web (RuleML-2005) Springer-Verlag A. Adi, S. Stoutenburg, S. Tabet LNCS 3791 2005 15 172-186 Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is useful for many applications in the Semantic Web, such as policies and business rules, agent brokering and negotiation, ontology and knowledge merging, etc., mainly due to interesting features, such as conflicting rules and priorities of rules. However, the RuleML syntax of defeasible logic may appear too complex for many users. Furthermore, the interplay between various technologies and languages, such as defeasible reasoning, RuleML, and RDF impose a demand for using multiple, diverse tools for building rule-based applications for the Semantic Web. In this paper we present VDR-Device, a visual integrated development environment for developing and using defeasible logic rule bases on top of RDF ontologies. VDR-Device integrates in a user-friendly graphical shell, a visual RuleML-compliant rule editor that constrains the allowed vocabulary through analysis of the input RDF ontologies and a defeasible reasoning system that processes RDF data and RDF Schema ontologies. ruleml-2005.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE%2C+VDR%2DDEVICE Galway, Ireland, 10-12 November http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fopenurl%2Easp%3Fgenre%3Darticle%26id%3Ddoi%3A10%2E1007%2F11580072%5F14 Integrated Development Environment Graphical Rule Editor Defeasible Logic Rule Markup Languages Semantic Web Reasoning RDF Schema Hybrid AcE: Combining Search Directions for Heuristic Planning 1 Computational Intelligence Blackwell 21(3) 2005 26 306-331 One of the most promising trends in Domain-Independent AI Planning, nowadays, is state-space heuristic planning. The planners of this category construct general but efficient heuristic functions, which are used as a guide to traverse the state space either in a forward or in a backward direction. Although specific problems may favor one or the other direction, there is no clear evidence why any of them should be generally preferred. This paper presents Hybrid-AcE, a domain-independent planning system that combines search in both directions utilizing a complex criterion that monitors the progress of the search, to switch between them. Hybrid AcE embodies two powerful domain-independent heuristic functions extending one of the AcE planning systems. Moreover, the system is equipped with a fact-ordering technique and two methods for problem simplification that limit the search space and guide the algorithm to the most promising states. The bi-directional system has been tested on a variety of problems adopted from the AIPS planning competitions with quite promising results. HybridAcE.pdf Towards an Object-Oriented Reasoning System for OWL 2 Int. Workshop on OWL Experiences and Directions B. Cuenca Grau, I. Horrocks, B. Parsia, P. Patel-Schneider 2005 11 In this paper we present O-DEVICE, a deductive object-oriented knowledge base system for reasoning over OWL documents. O-DEVICE imports OWL documents into the CLIPS production rule system by transforming OWL ontologies into an object-oriented schema of the CLIPS Object-Oriented Language (COOL) and instances of OWL classes into COOL objects. The pur-pose of this transformation is to be able to use a deductive object-oriented rule language for reasoning about OWL data. The O-DEVICE data model for OWL ontologies maps classes to classes, resources to objects, property types to class slot (or attribute) definitions and encapsulates resource properties inside re-source objects, as traditional OO attributes (or slots). In this way, when access-ing properties of a single resource, few joins are required. O-DEVICE is an ex-tension of a previous system, called R-DEVICE, which effectively maps RDF Schema and data into COOL objects and then reasons over RDF data using a deductive object-oriented rule language. owled05.pdf http%3A%2F%2Fiskp%2Ecsd%2Eauth%2Egr%2Fsystems%2Fo%2Ddevice%2Fo%2Ddevice%2Ehtml O%2DDEVICE 11-12 Nov. 2005, Galway, Ireland http%3A%2F%2Fwww%2Emindswap%2Eorg%2F2005%2FOWLWorkshop%2Faccepted%2Eshtml Ontologies OWL Object Model CLIPS Descriptive Semantics Production Rules Deductive Rules VDR-DEVICE: A Visual Editor for a Defeasible Logic Rule-ML-compatible Rule Language 2 4th International Semantic Web Conference (ISWC2005), Demo/Poster Session Edward Curry 2005 3 RuleML is a promising standardization effort for rule languages for the Semantic Web. However, the RuleML syntax may appear too complex for many users. Furthermore, the interplay between various Semantic Web technologies and languages impose a demand for using multiple, diverse tools for building rule-based applications for the Semantic Web. In this demonstration we present VDR-Device, a visual RuleML-compliant rule editor and an integrated development environment for developing and using defeasible logic rule bases on top of RDF ontologies. The visual rule editor constrains the allowed vocabulary through analysis of the input RDF ontologies. The development environment is supported by an RDF-aware defeasible reasoning system. Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and inconsistent information. Such reasoning is useful for many applications in the Semantic Web, such as policies and business rules, agent brokering and negotiation, ontology and knowledge merging, etc., mainly due to interesting features, such as conflicting rules and rule priorities. This demo presents a full example of using VDR-Device in a brokered trade application that takes place via an independent third party, the broker. The broker matches the buyer’s requirements and the sellers’ capabilities, and proposes a transaction when both parties can be satisfied by the trade. In our case, the concrete application is apartment renting and the landlord takes the role of the abstract seller. ISWC2005-poster.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE%2C+VDR%2DDEVICE Galway, Ireland, 6-10 November http%3A%2F%2Fiswc2005%2Esemanticweb%2Eorg%2FW%5FCallForPostersDemo%2Ehtml Integrated Development Environment Graphical Rule Editor Defeasible Logic Rule Markup Languages Semantic Web Reasoning RDF Schema Semantic Brokering Semantic Web: Vision and Technologies 2 2nd Int. Scientific Conf. on Computer Science IEEE Computer Society, Bulgarian Section Plamenka Borovska, Sofoklis Christofordis 2005 14 This paper introduces the vision behind the Semantic Web by using illustrative examples of how interaction with the future Web will be through the use of intelligent personal agents. Furthermore, the paper overviews the current Semantic Web technologies that will carry out this vision. Finally, the paper briefly presents the research on Semantic Web carried out at the Department of Informatics of the Aristotle University of Thessaloniki. conf-cs-halkidiki-bassiliades.pdf 30th Sep-2nd Oct 2005, Halkidiki, Greece Semantic Web Intelligent Agents Metadata Ontologies Logic Inference Applying neural networks with active neurons to sea-water quality measurements 2 2nd Int. Scientific Conf. on Computer Science IEEE Computer Society, Bulgarian Section Plamenka Borovska, Sofoklis Christofordis 2005 7 This study examines the presence of either linear or nonlinear relationships between a number of popular sea-water quality indicators such as water temperature, pH, amount of dissolved oxygen and turbidity. The data are obtained from a set of sensors in an underwater measurement station. The neural networks with active neurons are applied to the prediction of each one of the above four indicators and their performance is compared against a benchmark prediction method known as the random walk model. The random walk model is the simpler prediction method, which accepts as the best prediction for a variable its current value. The neural network with active neurons is a black box method, which contrary to neural networks with passive neurons does not require a long set of training data. The results show that for daily predictions the neural network with active neurons is able to beat the random walk model with regard to directional accuracy, namely the direction (upward or downwards) of the modelling object in the next day. hatzikos_et_al_1.pdf 30th Sep-2nd Oct 2005, Halkidiki, Greece Neural Networks System Modeling Active Neurons Simple Distributed Filtering on a CLP Platform 2 3rd Hellenic Conference on Artificial Intelligence (Companion Volume) G. Vouros and T. Panagiotopoulos 2004 10 318-327 The area of distributed constraint satisfaction has drawn significant attention in the past decade. The approaches proposed in the area can be classified in two large categories: distributed search techniques and distributed filtering techniques. The work described in this paper concerns the CLP implementation of the DSAC algorithm, a novel distributed filtering technique that is based on the singleton consistency algorithm. The advantages of the algorithm include a high pruning efficiency and a remarkable simplicity. The latter allows an unproblematic implementation of the algorithm in any constraint programming platform that supports network communication, without the need of tampering with the (low level) consistency algorithm employed. The present paper briefly describes DSAC along with its implementation in the CSPCONS distributed CLP platform and presents experimental results on a number of structured constraint problems. The motivation behind this work is twofold: to support our argument concerning the simple implementation of the algorithm and to further investigate the benefits of its application to constraint satisfaction problems. Animating Formal Models in a Communicating Sequential Process Platform 2 10th Panhellenic Conference on Informatics, Local Proceedings P. Bozanis and E. Houstis 2005 10 174-183 The X-machine formal method forms the basis for a specification/modeling language with a substantial potential value to software engineers. An X-machine is a more expressive and flexible state machine, capable of modeling both the dynamic and the static aspect of a system. Communicating X-machines provide a methodology for building communicating systems out of existing stand-alone X-machines. However, for practically using the model in an real-world system development process, a tool for demonstrating and informally verifying the properties of the modeled system is required. An ideal platform for efficiently implementing such a tool, should support, process oriented programming, efficient communication primitives and declarativeness. CSPCONS is a distributed CLP platform that supports program execution over multiple independent sequential CLP processes that synchronize though message and event passing. The present paper demonstrates the applicability of the \cons programming model to the implementation of a communicating X-machine animator tool that will act as the basis for an extended set of tools that will support the formal mathematical analysis of the specified X-machine models. Email Mining: Emerging Techniques for Email Management 3 Web Data Management Practices: Emerging Techniques and Technologies Idea Group Publishing Athena Vakali, George Pallis 2006 0 219-240 Email has met tremendous popularity over the past few years. People are sending and receiving many messages per day, communicating with partners and friends, or exchanging files and information. Unfortunately, the phenomenon of email overload has grown over the past years becoming a personal headache for users and a financial issue for companies. In this chapter, we will discuss how disciplines like Machine Learning and Data Mining can contribute to the solution of the problem by constructing intelligent techniques which automate email managing tasks and what advantages they hold over other conventional solutions. We will also discuss the particularity of email data and what special treatment it requires. Some interesting email mining applications like mail categorization, summarization, automatic answering and spam filtering will be also presented. katakis2006-idea.pdf To appear http%3A%2F%2Fwww%2Eidea%2Dgroup%2Dref%2Ecom%2Fbooks%2Fdetails%2Easp%3Fid%3D6152 Book+Page+at+the+IDEA+Group+Web+Site email e-mail mining text data streams email classification email clustering automatic answering spam filtering text mi ning Towards Automatic Synthesis of Educational Resources through Automated Planning 2 Proc. 4th Hellenic Conference on Artificial Intelligence (SETN-06) Springer Berlin Grigoris Antoniou, George Potamias, Costas Spyropoulos, Dimitris Plexousakis LNAI 3955 2006 11 421-431 This paper reports on the results of an ongoing project for the development of a platform for e-Learning, which automatically constructs curricula based on available educational resources and the learners needs and abilities. The system under development, called PASER (Planner for the Automatic Synthesis of Educational Resources), uses an automated planner, which given the initial state of the problem (learner’s profile, preferences, needs and abilities), the available actions (study an educational resource, take an exam, join an e-learning course, etc.) and the goals (obtain a certificate, learn a subject, acquire a skill, etc.) constructs a complete educational curriculum that achieves the goals. PASER is compliant with the evolving educational metadata standards that describe learning resources (LOM), content packaging (CP), educational objectives (RDCEO) and learner related information (LIP). paser.pdf Heraklion, Crete, 18-20 May Software Defect Prediction Using Regression via Classification 2 Proc. 4th ACS/IEEE International Conference on Computer Systems and Applications, AICCSA '06, (accepted for presentation) 2006 7 330- 336 In this paper we apply a machine learning approach to the problem of estimating the number of defects called Regression via Classification (RvC). RvC initially automatically discretizes the number of defects into a number of fault classes, then learns a model that predicts the fault class of a software system. Finally, RvC transforms the class output of the model back into a numeric prediction. This approach includes uncertainty in the models because apart from a certain number of faults, it also outputs an associated interval of values, within which this estimate lies, with a certain confidence. To evaluate this approach we perform a comparative experimental study of the effectiveness of several machine learning algorithms in a software dataset. The data was collected by Pekka Forselious and involves applications maintained by a bank of Finland. bibi-aiccsa2006.pdf Ensemble Pruning using Reinforcement Learning 2 Proc. 4th Hellenic Conference on Artificial Intelligence (SETN-06) Springer-Verlag G. Antoniou, G. Potamias, D. Plexousakis, C. Spyropoulos LNAI 3955 2006 10 301-310 Multiple Classifier systems have been developed in order to improve classification accuracy using methodologies for effective classifier combination. Classical approaches use heuristics, statistical tests, or a meta-learning level in order to find out the optimal combination function. We study this problem from a Reinforcement Learning perspective. In our modeling, an agent tries to learn the best policy for selecting classifiers by exploring a state space and considering a future cumulative reward from the environment. We evaluate our approach by comparing with state-of-the-art combination methods and obtain very promising results. partalas-setn2006.pdf Heraklion, Crete, 18-20 May http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2F11752912%5F31 Reinforcement Learning Multiple Classifier Systems Ensemble Prunning Classification Prediction of Translation Initiation Sites Using Classifier Selection 2 Proc. 4th Hellenic Conference on Artificial Intelligence (SETN-06) Springer-Verlag G. Antoniou, G. Potamias, D. Plexousakis, C. Spyropoulos LNAI 3955 2006 11 367-377 The prediction of the translation initiation site (TIS) in a genomic sequence is an important issue in biological research. Several methods have been proposed to deal with it. However, it is still an open problem. In this paper we follow an approach consisting of a number of steps in order to increase TIS prediction accuracy. First, all the sequences are scanned and the candidate TISs are detected. These sites are grouped according to the length of the sequence upstream and downstream them and a number of features is generated for each one. The features are evaluated among the instances of every group and a number of the top ranked ones are selected for building a classifier. A new instance is assigned to a group and is classified by the corresponding classifier. We experiment with various feature sets and classification algorithms, compare with alternative methods and draw important conclusions. TIS_SETN06.pdf Heraklion, Crete, 18-20 May Graphical Representation of Defeasible Logic Rules using Digraphs 2 Proc. 4th Hellenic Conference on Artificial Intelligence (SETN-06) Springer-Verlag G. Antoniou, G. Potamias, D. Plexousakis, C. Spyropoulos LNAI 3955 2006 5 529-533 Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and conflicting information. Nevertheless, it is based on solid mathematical formulations and is not fully comprehensible by end users, who often need graphical trace and explanation mechanisms for the derived conclu-sions. Directed graphs (or digraphs) can assist in this affair, but their applicabil-ity is balanced by the fact that it is difficult to associate data of a variety of types with the nodes and the connections in the graph. In this paper we try to utilize digraphs in the graphical representation of defeasible rules, by exploiting their expressiveness, but also trying to counter their major disadvantage, by de-fining multiple node and connection types. SETN06-skontopo-nbassili.pdf Heraklion, Crete, 18-20 May digraphs defeasible reasoning graphical representation O-DEVICE: An Object-Oriented Knowledge Base System for OWL Ontologies 2 Proc. 4th Hellenic Conference on Artificial Intelligence (SETN-06) Springer-Verlag G. Antoniou, G. Potamias, D. Plexousakis, C. Spyropoulos LNAI 3955 2006 11 256-266 Rule systems are a key component for the next step of evolution of the Semantic Web, allowing the development of declarative systems on top of ontologies. This paper reports on the implementation of such a system, called O-DEVICE, for reasoning over OWL documents. O-DEVICE exploits the rule language of the CLIPS production rule system and transforms OWL ontologies into an object-oriented schema of COOL. During the transformation procedure, OWL classes are mapped to COOL classes, OWL properties to class slots and OWL instances to COOL objects. The purpose of this transformation is two-fold: a) to exploit the advantages of the object-oriented representation and ac-cess all the properties of instances in one step, since properties are encapsulated inside resource objects; b) to be able to use a deductive object-oriented rule language for reasoning about OWL data, which operates over the object-oriented schema of CLIPS. The deductive rules are compiled into CLIPS pro-duction rules. The semantics of OWL are partly handled by the OWL transfor-mation procedure and partly by the rule compilation procedure. setn06-meditskos-nbassili-revised.pdf http%3A%2F%2Fiskp%2Ecsd%2Eauth%2Egr%2Fsystems%2Fo%2Ddevice%2Fo%2Ddevice%2Ehtml O%2DDEVICE Heraklion, Crete, 18-20 May http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2F11752912%5F27 OWL ontologies metadata object-oriented knowledge representation production rules deductive rules descriptive semantics prescriptive semantics ISKP Group Report 5 Semantic Web Factbook (Preliminary edition) AIS SIGSEMIS and OPEN RESEARCH SOCIETY publications M. Lytras ISSN: 1556-2301 2006 8 33-40 The Intelligent Systems and Knowledge Processing (ISKP) group belongs to the Department of Informatics, Aristotle University of Thessaloniki, Greece. The group’s main research areas include Logic Programming, Knowledge Representation and Reasoning, Automated Planning, Intelligent Applications and, of course, the Semantic Web. Research on those fields has led to a significant number of publications (over 140), including 6 authored and edited books. Nine people are currently associated with ISKP, including two faculty members, two associate researchers (post-doctoral), three PhD students and two external associates. The group is also collaborating with many distinctive researchers from Greek and international universities and research institutes. ISKP has participated with success in a variety of research and development projects, funded by the European Union and the Greek Government, many of which are related to the Semantic Web. Finally, ISKP is responsible for a number of under- and postgraduate taught courses at the Aristotle University of Thessaloniki, such as: Logic Programming, Artificial Intelligence, Knowledge Systems, Intelligent Autonomous Systems, Knowledge Management, Decision Support Systems, Intelligent Agents and the Semantic Web. SWFactBook-final.pdf http%3A%2F%2Fiskp%2Ecsd%2Eauth%2Egr%2F ISKP+Group http%3A%2F%2Fwww%2Esigsemis%2Eorg%2Ffactbook ISKP ISKP Projects LPIS Department of Informatics Aristotle University of Thessaloniki A Non-Monotonic Reasoning System for RDF Metadata 2 Proc. 11th Int. Workshop on Non-Monotonic Reasoning 2006 9 285-293 Non-monotonic reasoning constitutes an approach to rea-soning with incomplete or changing information and is sig-nificantly more powerful than standard reasoning, which simply deals with universal statements. Defeasible reason-ing, a member of the non-monotonic reasoning family, of-fers the extra capability of dealing with conflicting informa-tion and can represent facts, rules and priorities among rules. The main advantages of defeasible reasoning, how-ever, are not only limited to its enhanced representational capabilities, but also feature low computational complexity compared to mainstream non-monotonic reasoning. This paper presents a system for non-monotonic reasoning on the Semantic Web called VDR-Device, which is capable of rea-soning about RDF metadata over multiple Web sources us-ing defeasible logic rules. It is implemented on top of the CLIPS production rule system and features a RuleML com-patible syntax. The operational semantics of defeasible logic are implemented through compilation into a generic deduc-tive rule language. Since the RuleML syntax may appear complex for many users, we have also implemented a graphical authoring tool for defeasible logic rules that acts as a shell for the defeasible reasoning system. The tool con-strains the allowed vocabulary through analysis of the input RDF documents, so that the user does not have to type-in class and property names. NMR-final.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE%2C+VDR%2DDEVICE Lake District area, UK, 30 May - 1 June http%3A%2F%2Fcig%2Ein%2Etu%2Dclausthal%2Ede%2FNMR06%2F non-monotonic reasoning defeasible logic RDF RuleML Visualizing Defeasible Logic Rules for the Semantic Web 2 1st Asian Semantic Web Conference (ASWC'06) Springer-Verlag LNCS 4185 2006 14 278-292 Defeasible reasoning is a rule-based approach for efficient reasoning with incomplete and conflicting information. Such reasoning is useful in many Semantic Web applications, like policies, business rules, brokering, bargaining and agent negotiations. Nevertheless, defeasible logic is based on solid mathe-matical formulations and is, thus, not fully comprehensible by end users, who often need graphical trace and explanation mechanisms for the derived conclu-sions. Directed graphs can assist in confronting this drawback. They are a pow-erful and flexible tool of information visualization, offering a convenient and comprehensible way of representing relationships between entities. Their appli-cability, however, is balanced by the fact that it is difficult to associate data of a variety of types with the nodes and the arcs in the graph. In this paper we try to utilize digraphs in the graphical representation of defeasible rules, by exploiting the expressiveness and comprehensibility they offer, but also trying to leverage their major disadvantage, by defining two distinct node types, for rules and atomic formulas, and four distinct connection types for each rule type in defea-sible logic and for superiority relationships. The paper also briefly presents a tool that implements this representation methodology. ASWC-06.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml DR%2DDEVICE%2C+VDR%2DDEVICE Beijing, China, 3-7 September http%3A%2F%2Fwww%2Easwc2006%2Eorg%2Findex%2Ephp visualization defeasible logic RDF stratification DR-BROKERING: A Semantic Brokering System 1 Knowledge-Based Systems Elsevier Vol. 20, No. 1 2007 18 61-72 In this paper we study the brokering and matchmaking problem, that is, how a requester’s requirements and preferences can be matched against a set of offerings collected by a broker. The proposed solution uses the Semantic Web standard of RDF to represent the offerings, and a deductive logical language for expressing the requirements and preferences. We motivate and explain the approach we propose, and report on a prototypical implementation exhibiting the described functionality in a multi-agent environment. KBOS06final.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eknosys%2E2006%2E07%2E006 Brokering Rules Ontologies RDF Schema A Review of Multi-Label Classification Methods 2 2nd ADBIS Workshop on Data Mining and Knowledge Discovery 2006 11 99-109 Nowadays, multi-label classification methods are increasingly required by modern applications, such as protein function classification, music categorization and semantic scene classification. This paper introduces the task of multi-label classification, organizes the sparse related literature into a structured presentation and performs comparative experimental results of certain multi-label classification methods. It also contributes the presentation of an undocumented method and the definition of a concept for the quantification of the multi-label nature of a data set. Multi-Label Classification Classification Modern Applications of Machine Learning 2 Proceedings of the 1st Annual SEERC Doctoral Student Conference, Thessaloniki, Greece 2006 0 Machine learning is one of the older areas of artificial intelligence and concerns the study of computational methods for the discovery of new knowledge and for the management of existing knowledge. Machine learning methods have been applied to various application domains. However, in the few last years due to various technological advances and research efforts (e.g. completion of the Human Genome Project, evolution of the Web), new data have been available and consequently new domains where machine learning can be applied have been arisen. Some of these modern applications are learning from biological sequences, learning from email data, and learning in complex environments such as Web. In this paper we present the above three application domains as well as some recent efforts, where machine learning techniques are applied in order to analyze the data provided by these domains. tzanis_seerc06.pdf Machine Learning Applications Artificial Intelligence, 3rd edition (in Greek - Τεχνητή Νοημοσύνη, Γ' Έκδοση) 4 Book Έκδοση/Διάθεση: Εκδόσεις Πανεπιστημίου Μακεδονίας ISBN 978-960-8396-64-7 2011 900 <p>Τεχνητή Νοημοσύνη είναι ο τομέας της επιστήμης των υπολογιστών, που ασχολείται με τη σχεδίαση ευφυών υπολογιστικών συστημάτων, δηλαδή συστημάτων που επιδεικνύουν χαρακτηριστικά που σχετίζουμε με τη νοημοσύνη στην ανθρώπινη συμπεριφορά.</p> <p>Το βιβλίο προσεγγίζει την Τεχνητή Νοημοσύνη οριοθετώντας τα προβλήματα που αντιμετωπίζει, περιγράφοντας τρόπους αναπαράστασης γνώσης για αυτά και αλγορίθμους αναζήτησης λύσεών τους. Επιπλέον, παρουσιάζει κλασικές εφαρμογές που ιστορικά ανήκουν στην περιοχή, όπως το σχεδιασμό ενεργειών, τη μηχανική μάθηση, τα συστήματα γνώσης, καθώς και πιο σύγχρονες εφαρμογές όπως τους ευφυείς πράκτορες και το σημασιολογικό διαδίκτυο. Το βιβλίο συμπληρώνεται με αλγορίθμους και παραδείγματα εφαρμογών Τεχνητής Νοημοσύνης υλοποιημένα σε PROLOG και CLIPS, δύο διαδεδομένες γλώσσες προγραμματισμού του χώρου.</p> <p>Πρωταρχικός σκοπός του βιβλίου ήταν να αποτελέσει ένα διδακτικό βοήθημα για τη διδασκαλία ενός εισαγωγικού μαθήματος στην Τεχνητή Νοημοσύνη. Η πρώτη έκδοση έτυχε ευρείας αποδοχής από το σύνολο σχεδόν των Τμημάτων Πληροφορικής της τριτοβάθμιας εκπαίδευσης.</p> <p>Στη δεύτερη έκδοση η ύλη αναδιαρθρώθηκε για διευκόλυνση της διδασκαλίας και υπάρχοντα θέματα αναβαθμίστηκαν, ιδωμένα υπό νέα οπτική γωνία. Τέλος, προστέθηκαν νέα θέματα μεγάλου ερευνητικού και εμπορικού ενδιαφέροντος, κάνοντας το βιβλίο κατάλληλο για χρήση σε εξειδικευμένα προπτυχιακά ή μεταπτυχιακά μαθήματα.</p> <p>H τρίτη έκδοση συνοδεύεται από CD με διαφάνειες, προγράμματα και υλοποιήσεις αλγορίθμων, πολυμεσικό εκπαιδευτικό υλικό, γλωσσάριο ορολογίας, κτλ</p> http%3A%2F%2Faibook%2Ecsd%2Eauth%2Egr%2F %D4%E5%F7%ED%E7%F4%DE+%CD%EF%E7%EC%EF%F3%FD%ED%E7 http%3A%2F%2Faibook%2Ecsd%2Eauth%2Egr%2F Querying and Visualizing the Semantic Web 2 Proceedings of the 1st Annual SEERC Doctoral Student Conference, Thessaloniki, Greece 2006 0 The amount of unstructured information on the Web has reached a critical level, suggesting the need for the Semantic Web (SW). The SW expresses an initiative to improve the World Wide Web, by adding semantic content to the provided information, thus making the data accessible not only to humans but to machines as well. The fulfilment of this goal requires the contribution of a variety of research areas and tools: reasoning techniques that give the opportunity to agents/machines to act autonomously on the environment of the Web in order to fulfil the users’ needs and tools for assisting the end users in expressing these needs and comprehending the derived results. This paper presents two systems, namely O-DEVICE and VDR-DEVICE, developed by the Intelligent Systems and Knowledge Processing (ISKP) group at the Department of Informatics, Aristotle University of Thessaloniki, Greece. These two systems are designed to function in the SW environment. O-DEVICE is a system for querying and inferencing about ontologies expressed in the OWL Lite sublanguage and VDR-DEVICE is a visual integrated development environment for developing, using and visualizing defeasible logic rule bases on top of RDF ontologies. SEERC-ISKP.pdf Semantic Web Reasoning System Visualization RuleML Defeasible Logic Dynamic Feature Space and Incremental Feature Selection for the Classification of Textual Data Streams 2 ECML/PKDD-2006 International Workshop on Knowledge Discovery from Data Streams 2006 10 107-116 Real world text classification applications are of special interest for the machine learning and data mining community, mainly because they introduce and combine a number of special difficulties. They deal with high dimensional, streaming, unstructured, and, in many occasions, concept drifting data. Another important peculiarity of streaming text, not adequately discussed in the relative literature, is the fact that the feature space is initially unavailable. In this paper, we discuss this aspect of textual data streams. We underline the necessity for a dynamic feature space and the utility of incremental feature selection in streaming text classification tasks. In addition, we describe a computationally undemanding incremental learning framework that could serve as a baseline in the field. Finally, we introduce a new concept drifting dataset which could assist other researchers in the evaluation of new methodologies. katakis_ecml06_workshop_cr.pdf http%3A%2F%2Fmlkd%2Ecsd%2Eauth%2Egr%2Fdatasets%2Ehtml Concept+drifting+datasets+descibed+in+paper Berlin, Germany Text Classification Data Streams Concept Drift Dynamic Feature Space Incremental Feature Selection PersoNews: A Personalized News Reader Enhanced by Machine Learning and Semantic Filtering 2 5th International Conference on Ontologies, DataBases, and Applications of Semantics (ODBASE 2006) Springer-Verlag R. Meersman, Z. Tari 2006 8 975-982 In this paper, we present a web-based, machine-learning enhanced news reader (PersoNews). The main advantages of PersoNews are the aggregation of many different news sources, machine learning filtering offering personalization not only per user but also for every feed a user is subscribed to, and finally the ability for every user to watch a more abstracted topic of interest by employing a simple form of semantic filtering through a taxonomy of topics. (URL: http://news.csd.auth.gr). banos_odbase2006.pdf PersoNews Web Site: http://news.csd.auth.gr http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fm06t69466546023x%2F Springer+Link Montpellier, France News Filtering News Classification Text Mining Text Classification Personalization Ontology RSS Aggregator . On the Discovery of Mutually Exclusive Items in a Market Basket Database 2 2nd ADBIS Workshop on Data Mining and Knowledge Discovery 2006 12 1-12 Mining a transaction database for association rules is a particularly popular data mining task, which involves the search for frequent co-occurrences among items. One of the problems often encountered is the large number of weak rules extracted. Item taxonomies, when available, can be used to reduce them to a more usable volume. In this paper we introduce a new data mining paradigm, which involves the discovery of pairs of mutually exclusive items. We call this new type of knowledge mutual exclusion, as opposed to association, and we propose its use to tackle the aforementioned problem. We formulate the problem of mining for mutually exclusive items, provide important background information, propose a novel mutual exclusion metric and finally, present a mining algorithm that we test on transaction data. tzanis_ADMKD06.pdf Thessaloniki, Greece An interoperable and scalable Web-based system for classifier sharing and fusion 1 Expert Systems with Applications Elsevier 33(3) 2007 17 716-724 This paper describes CSF/DC, a Web-based system for classifier sharing and fusion. CSF/DC enables the sharing of classification models, by allowing the upload and download of such models expressed in the industry standard PMML language on the system’s online classifier repository. CSF/DC also leverages the individual knowledge shared by such (potentially heterogeneous) classification models and offers quality decision support to any user with an Internet connection through a guided procedure. However, some organizations or individuals might want to share the predictive capabilities of their classification models without compromising their internal structure. This is accommodated by CSF/DC through the use of Web services. Specifically, CSF/DC allows the participation of classifier Web services in the decision fusion process, by offering the necessary online mechanisms for the registration and invocation of such Web services developed and installed at remote sites. tsoumakas-eswa33(3).pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2006%2E06%2E021 Distributed Data Mining 3 Encyclopedia of Data Warehousing and Mining - 2nd Edition Idea Group Reference John Wang 2008 16 709-715 tsoumakas-dwm2.pdf Multi Label Classification: An Overview 1 International Journal of Data Warehousing and Mining Idea Group Publishing David Taniar 3(3) 2007 17 1-13 Nowadays, multi-label classification methods are increasingly required by modern applications, such as protein function classification, music categorization and semantic scene classification. This paper introduces the task of multi-label classification, organizes the sparse related literature into a structured presentation and performs comparative experimental results of certain multi-label classification methods. It also contributes the definition of concepts for the quantification of the multi-label nature of a data set. tsoumakas-ijdwm.pdf Multi-Label Classification Classfication A Novel Data Mining Approach for the Accurate Prediction of Translation Initiation Sites 2 7th International Symposium on Biological and Medical Data Analysis Springer-Verlag Nicos Maglaveras et al. 2006 12 92-103 In an mRNA sequence, the prediction of the exact codon where the process of translation starts (Translation Initiation Site – TIS) is a particularly important problem. So far it has been tackled by several researchers that apply various statistical and machine learning techniques, achieving high accuracy levels, often over 90%. In this paper we propose a mahine learning approach that can further improve the prediction accuracy. First, we provide a concise review of the literature in this field. Then we propose a novel feature set. We perform extensive experiments on a publicly available, real world dataset for various vertebrate organisms using a variety of novel features and classification setups. We evaluate our results and compare them with a reference study and show that our approach that involves new features and a combination of the Ribosome Scanning Model with a meta-classifier shows higher accuracy in most cases. Tzanis_ISBMDA06.pdf Thessaloniki, Greece Mining for Mutually Exclusive Items in Transaction Databases 1 International Journal of Data Warehousing and Mining Idea Group Publishing David Taniar 3(3) 2007 Association rule mining is a popular task that involves the discovery of co-occurences of items in transaction databases. Several extensions of the traditional association rule mining model have been proposed so far, however, the problem of mining for mutually exclusive items has not been directly tackled yet. Such information could be useful in various cases (e.g. when the expression of a gene excludes the expression of another) or it can be used as a serious hint in order to reveal inherent taxonomical information. In this paper, we address the problem of mining pairs of items, such that the presence of one excludes the other. First, we provide a concise review of the literature, then we define this problem, we propose a probability-based evaluation metric, and finally a mining algorithm that we test on transaction data. Tzanis_IJDWM07.pdf Detection and Prediction of Rare Events in Transaction Databases 1 International Journal on Artificial Intelligence Tools World Scientific Publishing Company 16(5) 2007 20 829 - 848 Rare events analysis is an area that includes methods for the detection and prediction of events, e.g. a network intrusion or an engine failure, that occur infrequently and have some impact to the system. There are various methods from the areas of statistics and data mining or that purpose. In this article we propose PREVENT, an algorithm which uses inter-transactional patterns for the prediction of rare events in transaction databases. PREVENT is a general purpose inter-transaction association rules mining algorithm that optimally fits the demands of rare event prediction. It requires only 1 scan on the original database and 2 over the transformed, which is considerably smaller and it is complete as it does not miss any patterns. We provide the mathematical formulation of the problem and experimental results that show PREVENT’s efficiency in terms of run time and effectiveness in terms of sensitivity and specificity. Berberidis_IJAIT2007CmrReady.pdf http%3A%2F%2Fwww%2Eworldscinet%2Ecom%2Fcgi%2Dbin%2Fdetails%2Ecgi%3Fid%3Dvoliss%3Aijait%5F1605%26type%3Dtoc rare events prediction data mining sequence mining Object-Oriented Modelling of RDF Schema Ontologies 2 11th Panhellenic Conference on Informatics (PCI 2007) 2007 11 479-489 Ontologies are the primary knowledge representation tool in the Semantic Web and are mainly used in defining common vocabularies, used in the exchange of information among Semantic Web applications. In the process of encoding ontologies, appropriate ontology languages are applied; such a language is RDF Schema, one of the dominant standards. A variety of commercial and educational tools that address the tasks of developing and manipulating RDF Schema ontologies has been developed. None of them, however, is specifically destined for the inexperienced Semantic Web user. In this paper we present RDFSbuilder, a Java-built visual authoring tool for developing RDF Schema ontologies. The system helps users to develop their model quickly and efficiently, without being concerned about syntax or semantic errors. Furthermore, it adopts a purely object-oriented representation of the RDF Schema model, emphasizing on functional flexibility and simplicity of use. As a result, the model produced is easy to understand and equally easy to handle. pci2007-kontopo_et_al.pdf Patras, Greece, 18-20 May RDF Schema ontology editor visual editor semantic web Proof Explanation in the DR-DEVICE System 2 1st International Conference on Web Reasoning and Rule Systems (RR 2007) Springer-Verlag LNCS 4524 2007 10 249-258 Trust is a vital feature for the Semantic Web: If users (humans and agents) are to use and integrate system answers, they must trust them. Thus, systems should be able to explain their actions, sources, and beliefs, and this issue is the topic of the proof layer in the design of the Semantic Web. This paper presents the design of a system for proof explanation on the Semantic Web, based on defeasible reasoning. The basis of this work is the DR-DEVICE system that is extended to handle proofs. A critical aspect is the representation of proofs in an XML language, which is achieved by a RuleML language extension. rr2007-nbassili-et-al.pdf DR%2DDEVICE http%3A%2F%2Fiskp%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml Innsbruck, Austria, 7-8 June 2007 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fc13w86447h200648%2F%3Fp%3Dc2e7e79bf0174971882e0733f14a8487%26pi%3D18 proof explanation rules rules rule defeasible logic semantic web RDF A Visualization Algorithm for Defeasible Logic Rule Bases over RDF Data 2 1st International Conference on Web Reasoning and Rule Systems (RR 2007) Springer LNCS 4524 2007 3 367-369 This work presents a visualization algorithm for defeasible logic rule bases as well as a software tool that applies this algorithm, according to which, a directed graph is produced that represents the rule base. The graph features distinct node types for rules and atomic formulas and distinct connection types for the various rule types of defeasible logic. rr2007-kontopo-et-al.pdf VDR%2DDEVICE http%3A%2F%2Fiskp%2Ecsd%2Eauth%2Egr%2Fsystems%2Fdr%2Ddevice%2Ehtml Innsbruck, Austria, 7-8 June http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fj5k00561283140mp%2F%3Fp%3D0a167c68a32a4f399ab7b872d0301dbc%26pi%3D32 visualization rule base rule base defeasible logic RDF RDF stratification PASER: A Curricula Synthesis System based on Automated Problem Solving 1 International Journal on Teaching and Case Studies, Special Issue on "Information Systems: The New Research Agenda, the Emerging Curriculum and the New Teaching Paradigm" Inderscience Vol. 1, Nos. 1/2 2007 13 159-170 This paper presents PASER, a system for automatically synthesizing curricula using AI Planning and Machine Learning techniques on an ontology of educational resources metadata. The ontology is a part–of hierarchy of learning themes which correspond to RDCEO competencies. The system uses an automated planner, which given the initial state of the problem (learner’s profile, preferences, needs and abilities), the available actions (study an educational resource, take an exam, join an e-learning course, etc.) and the goals (obtain a certificate, learn a subject, acquire a skill, etc.) constructs a complete educational curriculum that achieves the goals. PASER is accompanied by a Machine Learning module that classifies textually described users’ learning requests to competencies registered within the ontology. Furthermore, the ML module interactively assists content providers in constructing educational resources metadata (LOM records) that comply with the ontology concerning both learning objectives and prerequisites. IJTCS-vrakas_et_al.pdf http%3A%2F%2Fwww%2Einderscience%2Ecom%2Fsearch%2Findex%2Ephp%3Faction%3Drecord%26rec%5Fid%3D14217%26prevQuery%3D%26ps%3D10%26m%3Dor Curricula Synthesis Automated Planning Text Mining Monitoring Water Quality through a Telematic Sensor Network and a Fuzzy Expert System 1 Expert Systems Blackwell 24(3) 2007 29 143-161 In this paper we present an expert system that monitors sea water quality and pollution in Northern Greece, through a sensor network called "Andromeda". The expert system monitors sensor data collected by Local Monitoring Stations and reasons about the current level of water suitability for various aquatic uses, such as swimming and piscicultures. The aim of the expert system is to help the authorities in the "decisionmaking" process in the battle against the pollution of the aquatic environment, which is very vital for the public health and the economy of Northern Greece. The expert system determines, using fuzzy logic, when certain environmental parameters exceed certain "pollution" limits, which are specified either by the authorities or by environmental scientists, and flags out appropriate alerts. expert-systems-hatzikos_et_al.pdf http%3A%2F%2Fwww%2Eblackwell%2Dsynergy%2Ecom%2Fdoi%2Fabs%2F10%2E1111%2Fj%2E1468%2D0394%2E2007%2E00426%2Ex Sensor Network Pollution Monitoring Aquatic Uses Expert System Fuzzy Logic A Visual Programming Tool for Designing Planning Problems for Semantic Web Service Composition 3 Visual Languages for Interactive Computing: Definitions and Formalizations Idea Group Publishing F. Ferri 2007 25 302 - 326 This chapter is concerned with the issue of knowledge representation for AI Planning problems, especially those related to Semantic Web Service composition. It discusses current approaches in encoding planning problems using the PDDL formal language and it presents ViTAPlan, a user-friendly visual tool for planning. The tool is built on top of HAPRC, a rule-configurable planning system, which automatically adapts to each problem, in order to achieve best performance. Apart from HAPRC, ViTAPlan can be interfaced with any other planning system that supports the PDDL language. More than just being a user friendly environment for executing the underlying planner, the tool serves as a unified planning environment for encoding a new problem, solving it, visualizing the solution and monitoring its execution on a simulation of the problem’s world. The tool consists of various sub-systems, each one accompanied by a graphical interface, which collaborate with each other and assist the user, either a knowledge engineer, a domain expert, an academic or even an end-user in industry, to carry out complex planning tasks, such as composing complex Semantic Web Services from simple ones, in order to achieve complex tasks. The key feature of ViTAPlan is a visual programming module that enables the user to encode new planning problems just by using visual elements and simple mouse operations. The visual tool performs a validity check on the visual program created by the user and then compiles it to PDDL files that are ready to be used by any planning system. Finally, the planning system will solve the planning problem and then export the plan in an appropriate Web Service composition language to a Web Service execution monitoring system or just publish it in a UDDI registry. ferri.pdf Automated Planning Graphical Interfaces Knowledge Engineering Semantic Web Service Composition Visual Development of Defeasible Logic Rules for the Semantic Web 3 Visual Languages for Interactive Computing: Definitions and Formalizations Idea Group Publishing F. Ferri 2007 31 273-301 This chapter is concerned with the visualization of defeasible logic rules in the Semantic Web domain. The Semantic Web expresses an initiative to improve the World Wide Web, by adding semantic content to the provided information, thus making the data accessible not only to humans but to machines as well. Logic plays an important role in the development of the Semantic Web and defeasible reasoning seems to be a very suitable tool in this affair, since it allows reasoning with incomplete and conflicting information. However, defeasible reasoning is too complex for an end-user, who often needs graphical trace and explanation mechanisms for the derived conclusions. Directed graphs can assist towards this direction, by offering a powerful and flexible means of information visualization. The notion of direction, featured by directed graphs, appears to be extremely applicable for the representation of rule attacks and superiorities in defeasible reasoning. Their applicability, however, is balanced by the fact that it is difficult to associate data of a variety of types with the nodes and the connections between the nodes in the graph. In this chapter we try to utilize digraphs in the graphical representation of defeasible rules, by exploiting the expressiveness and comprehensibility they offer, but also trying to leverage their major disadvantage, by proposing a representation approach that features two distinct node types, for rules and atomic formulas, and four distinct connection types for each rule type in defeasible logic and for superiority relationships. Finally, the chapter briefly presents a tool that implements this representation methodology. chapter'07-kontopoulos-bassiliades.pdf Non-monotonic Reasoning Defeasible Logic Semantic Web Visual Representation Information Visualization DR-NEGOTIATE – A System for Automated Agent Negotiation with Defeasible Logic-Based Strategies 1 Data and Knowledge Engineering Elsevier 63(2) 2007 25 362-380 This paper reports on a system for automated agent negotiation, based on a formal and executable approach to capture the behavior of parties involved in a negotiation. It uses the JADE agent framework, and its major distinctive feature is the use of declarative negotiation strategies. The negotiation strategies are expressed in a declarative rules language, defeasible logic, and are applied using the implemented system DR-DEVICE. The key ideas and the overall system architecture are described, and a particular negotiation case is presented in detail. DKE-nego-2007.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Edatak%2E2007%2E03%2E004 Agent Negotiation Semantic Web Defeasible Logic RDF An Empirical Study of Sea Water Quality Prediction 1 Knowledge-Based Systems Elsevier 21(6) 2008 471-478 This paper studies the problem of predicting future values for a number of water quality variables, based on measurements from under-water sensors. It performs both exploratory and automatic analysis of the collected data with a variety of linear and nonlinear modeling methods. The paper investigates issues, such as the ability to predict future values for a varying number of days ahead and the effect of including values from a varying number of past days. Experimental results provide interesting insights on the predictability of the target variables and the performance of the different learning algorithms. hatzikos-kbs2008.pdf E. Hatzikos, G. Tsoumakas, G. Tzanis, N. Bassiliades, I. Vlahavas, “An Empirical Study of Sea Water Quality Prediction”, Knowledge-Based Systems 21, pp. 471-478, 2008. http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eknosys%2E2008%2E03%2E005 A Semantic Web Service Discovery and Composition Prototype Framework Using Production Rules 2 OWL-S: Experiences and Directions 2007 14 The full realization of the semantic Web services demands efficient algorithms able to perform the procedures of service discovery, composition and invocation. In this paper, we present ProSeDisCo, our approach for developing a semantic Web service discovery and composition framework on top of the CLIPS rule-based system. More specifically, we describe our methodology of utilizing production rules over Web services semantic descriptions expressed in the OWL-S ontology. The purpose of these rules is to discover and create a Web service choreography that matches users’ input and output requirements by utilizing a rule-based OWL reasoning engine in order to match semantically the requirements imposed by the users and the advertisements of the Web services. prosedisco.pdf Innsbruck, Austria http%3A%2F%2Fwww%2Eai%2Esri%2Ecom%2FOWL%2DS%2D2007%2Fagenda%2Ehtml Semantic Web Services OWL-S Production Rules OWL Reasoning Cooperative CG-Wrappers for Web Content Extraction 2 Proc. 15th International Conference on Conceptual Structures (ICCS '07) Springer-Verlag Berlin Heidelberg U. Priss, S. Polovina, and R. Hill LNAI 4604 2007 4 476-479 We use Conceptual Graphs (CGs) to model web content extraction rules (CG-Wrappers). The approach presented incorporates all major existing extraction techniques and allows the definition of synergies of cooperative wrappers for handling complex extraction task, without requiring programming. Sheffield, UK extraction rules conceptual graphs wrappers Ensemble Selection for Water Quality Prediction 2 Proceedings of the 10th International Conference on Engineering Applications of Neural Networks 2007 8 This paper studies the greedy ensemble selection algorithm for ensembles of regression models. We explore two interesting parameters of this algorithm: a) the direction of search (forward, backward), and b) the performance evaluation dataset (training set, validation set) on a large ensemble (200 models) of neural networks and support vector machines. Experimental comparison of the different parameters are performed on an application domain with important social and commercial value: water quality monitoring. In specific we experiment on real data collected from an underwater sensor system. partalas-eann2007.pdf Thessaloniki MANTIS: A Data Mining Methodology for Effective Translation Initiation Site Prediction 2 Proceedings of the 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society IEEE 2007 0 The prediction of the translation initiation site in a genomic sequence with the highest possible accuracy is an important problem that still has to be investigated by the research community. Current approaches perform quite well, however there is still room for a more general framework for the researchers who want to follow an effective and reliable methodology. We developed a prediction methodology that combines ad hoc as well as discovered knowledge in order to significantly increase the achieved accuracy reliably. Our methodology is modular and consists of three major decision components: a consensus component, a coding region classification component and a novel ATG location-based component that allows for the utilization of the advantages of the popular Ribosome Scanning Model while overcoming its limitations. All three of them are combined into a meta-classification system, using stacked generalization, in a highly effective prediction framework. We performed extensive comparative experiments on four different datasets, showing that the increase in terms of accuracy and adjusted accuracy is not only statistically significant, but also the highest reported. Tzanis_EMBC07.pdf Lyon, France An Ontology-based Planning System for e-Course Generation 1 Expert Systems with Applications Elsevier 35 (1-2) 2008 14 398-406 Researchers in the area of educational software have always shown a great deal of interest for the automatic synthesis of learning curricula. During the recent years, with the extensive use of metadata and the emergence of the Semantic Web, this vision is gradually turning into a reality. A number of systems for curricula synthesis have been proposed. These systems are based on strong relations defined in the metadata of learning objects, which allow them to be combined with other learning objects, in order to form a complete educational program. This article presents PASER, a system for automatically synthesizing curricula using AI Planning and Semantic Web technologies. The use of classical planning techniques allows the system to dynamically construct learning paths even from disjoint learning objects, meeting the learner’s profile, preferences, needs and abilities. eswa-pythagoras.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2007%2E07%2E034 ontology planning knowledge representation reasoning automatic curricula synthesis e-learning e-course A Web Service Reasoner for the Semantic Web 2 Proceedings of the 2nd Annual SEERC Doctoral Student Conference, Thessaloniki, Greece 2007 11 Web Services offer a standard interface for describing the available services on the Web. Common applications of Web Services are B2B communications and e-commerce, mainly because they are platform and network-independent, easily deployable and offer great reusability. Moreover, the Semantic Web initiative proposes technologies and languages for annotating information on the Web, so that it can be understood, interpreted and exploited by software agents. For the realization of such architecture, agents should be able to reason over the annotated information, in order to make decisions and to successfully cooperate with each other. To this end, logic and rules play an important role, enabling the description of assertions that can be used to derive new knowledge and the implementation of agent behaviour. In this paper we describe the Web Service implementation of a rule-based RDF reasoner, called DR-DEVICE. The deployment of the reasoner as a Web Service enables other applications to use the system over the Internet, by exploiting the well-defined interface that Web Service technology offers. Agents can use the service by interchanging messages, based on standards (SOAP) and already existing Internet protocols (HTTP), in order to enrich their functionality with reasoning capabilities. Furthermore, the system can serve as a software component of a more complex and distributed framework that would compose a variety of Web Services in order to achieve a new functionality. DR-DEVICE supports both deductive and defeasible rules and can be extended to the proof layer of the Semantic Web architecture, for validating the derivations stemming from the inferential logic activity. The service is available to end users through a Web interface. seerc-final.pdf RDF Rule-based Reasoning Semantic Web Web Services Ontology Development for Computer Supported Collaborative Learning Scripts 2 3rd Balkan Conference in Informatics (BCI'2007), to be presented 2007 9 This paper presents an ongoing effort to develop an ontology for Computer-Supported Collaborative Learning (CSCL) Scripts. Our work merges the field of collaborative learning with the field of semantic web and provides a framework for the formalization of collaboration scripts using the OWL language. Collaboration scripts are didactic scenarios that prescribe learners’ interactions in collaborative settings. A script comprises a number of phases and each phase specifies the task that learners have to perform, the composition of the group, the distribution of the task, the mode of interaction and the phase duration. Scripts can also be positioned along various design dimensions, such as their granularity, coercion degree and locus of control. The presented ontology is being developed with the purpose of formalizing scripts in order to promote their reusability and portability between various computer-supported learning platforms. We discuss design decisions and illustrate how this ontology could be beneficial when embedded in a knowledge-based system that supports collaborative learning. 27-29 Sep. 2007, Sofia, Bulgaria ontology computer supported collaborative learning collaboration script LPMO: An Ontology for Learning Problems 2 3rd Balkan Conference in Informatics (BCI'2007), to be presented 2007 9 This work presents the LPMO (Learning Problems Modeling Ontology), an ontology for the domain of learning problems that models the types of problems learners are faced with, their most important dimensions (such as complexity, structure and abstractness), the types of their representations and solutions and also the learner’s skills required to solve them. In general, the proposed ontology is an effort to formalize the domain of learning problems in order to enable its representation in a machine-processable way. Developing the ontology is a first step towards building adaptive and interoperable web-based environments for problem-based learning, which can support a variety of learners in a variety of problem-solving tasks. The paper presents the key aspects of the domain and the proposed ontology and it further explores how the ontology can be used in technology-enhanced learning environments that store, manage and share problem-based learning information concerning different kinds of learning problems and their instantiations. 27-29 Sep. 2007, Sofia, Bulgaria learning problem problem-based learning ontology well-structured problem ill-structured problem Random k-Labelsets: An Ensemble Method for Multilabel Classification 2 Proceedings of the 18th European Conference on Machine Learning (ECML 2007) Springer Verlag J.N. Kok, J. Koronacki, R.L. de Mantaras, S. Matwin, D. Mladenic, A. Skowron LNAI 4701 2007 12 406-417 This paper proposes an ensemble method for multilabel classification. The RAndom k-labELsets (RAKEL) algorithm constructs each member of the ensemble by considering a small random subset of labels and learning a single-label classifier for the prediction of each element in the powerset of this subset. In this way, the proposed algorithm aims to take into account label correlations using single-label classifiers that are applied on subtasks with manageable number of labels and adequate number of examples per label. Experimental results on common multilabel domains involving protein, document and scene classification show that better performance can be achieved compared to popular multilabel classification approaches. http%3A%2F%2Fvideolectures%2Enet%2Fecml07%5Ftsoumakas%5Fem%2F Watch+the+presentation+in+VideoLectures Warsaw, Poland Mining High Quality Clusters of SAGE Data 2 Proceedings of the 2nd VLDB Workshop on Data Mining in Bioinformatics 2007 10 Serial Analysis of Gene Expression (SAGE) is a method that allows the quantitative and simultaneous analysis of the whole gene function of a cell. One of the advantages of this method is that the experimenter does not have to select a priori the mRNA sequences that will be counted in a sample. This makes SAGE a powerful tool for analyzing gene expression and studying various diseases, such as cancer. An important concern in cancer studies is the discovery of the differences between healthy and cancerous samples and the accurate separation of these two groups of samples. However, the high dimensionality of the data, the multiple cell sources (i.e. bulk and cell line) and the multiple cancer subtypes make very difficult the effective clustering of SAGE libraries. Furthermore, the various sources of noise pose an extra challenge to data miners. For all these reasons we propose an approach that involves the discretization of the data, the selection of the most prominent gene tags and the use of a clustering algorithm in order to obtain more compact and reliable clusters that can assist cancer profiling. We experimented with two families of clustering algorithms, partitional and hierarchical, and we utilized various cluster validity criteria in order to evaluate the resulted clustering structures. The experimental results have shown that our approach provides more interesting clustering structures. Tzanis_VDMB07.pdf Vienna, Austria Clustering Gene Expression SAGE Cancer Accurate Classification of SAGE Data Based on Frequent Patterns of Gene Expression 2 Proceedings of the IEEE International Conference on Tools with Artificial Intelligence (ICTAI 2007) IEEE 2007 5 In this paper we present a method for classifying accurately SAGE (Serial Analysis of Gene Expression) data. The high dimensionality of the data, namely the large number of features, in combination with the small number of samples poses a great challenge and demands more accurate and robust algorithms for classification. The prediction accuracy of the up to now proposed approaches is moderate. In our approach we exploit the associations among the expressions of genes in order to construct more accurate classifiers. For validating the effectiveness of our approach we experimented with two real datasets using numerous feature selection and classification algorithms. The results have shown that our approach improves significantly the classification accuracy, which reaches 99%. Tzanis_ICTAI07.pdf Patras, Greece Classification Gene Expression SAGE Cancer Frequent Patterns Multi-Agent Reinforcement Learning using Strategies and Voting 2 Proceedings of the 19th IEEE International Conference on Tools with Artificial Intelligence (ICTAI), (to be presented) IEEE 2007 7 Multiagent learning attracts much attention in the past few years as it poses very challenging problems. Reinforcement Learning is an appealing solution to the problems that arise to Multi Agent Systems (MASs). This is due to the fact that Reinforcement Learning is a robust and well suited technique for learning in MASs. This paper proposes a multi-agent Reinforcement Learning approach, that uses coordinated actions, which we call strategies and a voting process that combines the decisions of the agents, in order to follow a strategy. We performed experiments to the predator-prey domain, comparing our approach with other multi-agent Reinforcement Learning techniques, getting promising results. multi-agent reinforcement learning voting Visual Stratification of Defeasible Logic Rule Bases 2 Proceedings of the 19th IEEE International Conference on Tools with Artificial Intelligence (ICTAI) IEEE 2007 8 238-245 Logic and proofs constitute key factors in increas-ing the user trust towards the Semantic Web. Defeasi-ble reasoning is a useful tool towards the development of the Logic layer of the Semantic Web architecture. However, having a solid mathematical notation, it may be confusing to end users, who often need graphical trace and explanation mechanisms for the derived con-clusions. In a previous work of ours, we outlined a methodology for representing defeasible logic rules, utilizing directed graphs that feature distinct node and connection types. However, visualizing a defeasible logic rule base also involves the placement of the mul-tiple graph elements in an intuitive way, a non-trivial task that aims at improving user comprehensibility. This paper presents a stratification algorithm for visu-alizing defeasible logic rule bases that query and rea-son about RDF data as well as a tool that applies this algorithm. ICTAI'07-kontopoulos-bassiliades.pdf Patras, Greece defeasible logic visual representation semantic web reasoning Object-Oriented Similarity Measures for Semantic Web Service Matchmaking 2 Proc. 5th IEEE European Conf. on Web Services (ECOWS 2007), Halle (Saale), Germany, November 26-28 IEEE 2007 10 The semantic annotation of Web services capabilities with ontological information aims at providing the neces-sary infrastructure for facilitating efficient and accurate service discovery. The main idea is to apply reasoning techniques over semantically enhanced Web service re-quests and advertisements in order to determine Web ser-vices that meet certain requirements. In this paper we present our work for introducing similarity measures in-spired from the domain of Object-Oriented paradigm for ontology concept matching. Our work focuses on the utili-zation of such measures over an Object-Oriented schema that is created through mapping rules of OWL constructs and semantics into the Object-Oriented model. The goal of the approach is to combine the Object-Oriented repre-sentation of the information and the reasoning over OWL semantics in order to enhance the retrieval of semanti-cally relevant, to some criteria, Web services. medi-ecows2007.pdf 57%2D66 A Rule-based Object-Oriented OWL Reasoner 1 IEEE Transactions on Knowledge and Data Engineering IEEE vol. 20, no. 3 2008 14 397-410 In this paper we describe O-DEVICE, a memory-based knowledge base system for reasoning and querying OWL ontologies by implementing RDF/OWL entailments in the form of production rules in order to apply the formal semantics of the language. Our approach is based on a transformation procedure of OWL ontologies into an Object-Oriented schema and the application of inference production rules over the generated objects in order to implement the various semantics of OWL. In order to enhance the performance of the system, we introduce a dynamic approach of generating production rules for ABOX reasoning and an incremental approach of loading ontologies. O-DEVICE is built over the CLIPS production rule system, using the object-oriented language COOL to model and handle ontology concepts and RDF resources. One of the contributions of our work is that we enable a well-known and efficient production rule system to handle OWL ontologies. We argue that although native OWL rule reasoners may process ontology information faster, they lack some of the key features that rule systems offer, such as the efficient manipulation of the information through complex rule programs. We present a comparison of our system with other OWL reasoners, showing that O-DEVICE can constitute a practical rule environment for ontology manipulation. http%3A%2F%2Fdoi%2Eieeecomputersociety%2Eorg%2F10%2E1109%2FTKDE%2E2007%2E190699 Visual Representation of Web Service Composition Problems through VLEPpO 2 International Workshop on Intelligent Web Based Tools, 19th IEEE International conference on Tools with Artificial Intelligence Juan D. Velasquez 2007 0 42 - 49 This paper discusses the problem of the automatic composition of Semantic Web Services. Web Services constitute a new computing paradigm, which provides a standardized framework that facilitates the interoperability among software systems and machines that are accessible through the Internet. Semantics can significantly improve software reuse and discovery and allow the automatic composition of Web Services in order to produce large scale applications. The use of VLEPpO for the automatic composition of Web Services is proposed. VLEPpO is a visual programming tool for designing planning problems using visual elements and simple mouse operations. In the tool the user simply defines the properties of the available Web Services and the global goals of the application. Then VLEPpO automatically forms the description as a planning problem, solves it by calling an appropriate planning system and exports the solution either to a Web Service execution monitoring system or to a UDDI registry. wscpp.pdf Patras, Greece Reinforcement Learnign and Automated Planning: A Survey 3 Artificial Intelligence for Advanced Problem Solving Techniques IGI D. Vrakas and I. Vlahavas 2008 0 148-165 This chapter presents a detailed survey on Artificial Intelligent approaches that combine Reinforcement Learning and Automated Planning. There is a close relationship between those two areas, as they both deal with the process of guiding an agent, situated in a dynamic environment, in order to achieve a set of predefined goals. Therefore, it is straightforward to integrate learning and planning in a single guiding mechanism and there have been many approaches in this direction during the past years. The approaches are organized and presented according to various characteristics, as the used planning mechanism or the reinforcement learning algorithm. rlplan.pdf Eds 4 Artificial Intelligence for Advanced Problem Solving Techniques IGI 2007 0 VLEPpO: A Visual Language for Problem Representation 2 26th Workshop of the UK Planning and Scheduling Special Interest Group Roman Bartak 2007 7 60 - 66 AI planning constitutes a field of interest as its techniques can be applied to many areas. Contemporary systems that are being developed deal with certain aspects of planning and focus mainly on dealing with advanced features such as resources, time and numerical expressions. This paper presents VLEPpO, a Visual Language for Enhanced Planning problem Orchestration. VLEPpO is a visual programming environment that allows the user to easily define planning domains and problems, acquire their PDDL representations, as well as receive solutions, utilizing web services infrastructure. vleppo.pdf Prague, Czech Republic Regression via Classification applied on Software Defect Prediction 1 Expert Systems with Applications Elsevier 34(3) 2008 10 2091-2101 In this paper we apply Regression via Classification (RvC) to the problem of estimating the number of software defects. This approach apart from a certain number of faults, it also outputs an associated interval of values, within which this estimate lies with a certain confidence. RvC also allows the production of comprehensible models of software defects exploiting symbolic learning algorithms. To evaluate this approach we perform an extensive comparative experimental study of the effectiveness of several machine learning algorithms in two software data sets. RvC manages to get better regression error than the standard regression approaches on both datasets. An Adaptive Personalized News Dissemination System 1 Journal of Intelligent Information Systems Springer 32 (2) 2009 21 191-201 With the explosive growth of the Word Wide Web, information overload became a crucial concern. In a data-rich information-poor environment like the Web, the discrimination of useful or desirable information out of tons of mostly worthless data became a tedious task. The role of Machine Learning in tackling this problem is thoroughly discussed in the literature, but few systems are available for public use. In this work, we bridge theory to practice, by implementing a web-based news reader enhanced with a specifically designed machine learning framework for dynamic content personalization. This way, we get the chance to examine applicability and implementation issues and discuss the effectiveness of machine learning methods for the classification of real-world text streams. The main features of our system named PersoNews are: a) the aggregation of many different news sources that offer an RSS version of their content, b) incremental filtering, offering dynamic personalization of the content not only per user but also per each feed a user is subscribed to, and c) the ability for every user to watch a more abstracted topic of interest by filtering through a taxonomy of topics. PersoNews is freely available for public use on the WWW (http://news.csd.auth.gr). Katakis_JIIS.pdf http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fc543376j5374772q%2F Web Text Classification RSS Personalization News Naive Bayes Machine Learning Dynamic Feature Space Data Streams Concept Drfit Deploying Defeasible Logic Rule Bases for the Semantic Web 1 Data and Knowledge Engineering Elsevier 66 (1) 2008 44 116-146 Logic is currently the target of the majority of the upcoming efforts towards the realization of the Semantic Web vision, namely making the content of the Web accessible not only to humans, as it is today, but to machines as well. Defeasible reasoning, a rule-based approach to reasoning with incomplete and conflicting information, is a powerful tool in many Semantic Web applications. Despite its strong mathematical background, logic, in general, and defeasible logic, in particular, may overload the user with tons of additional complex semantic relationships among data and metadata of the Semantic Web. To this end, a comprehensible, visual representation of these semantic relationships (rules) would help users understand them and make more use of them. This paper pre-sents VDR-DEVICE, a defeasible reasoning system, designed specifically for the Semantic Web environment. VDR-DEVICE is an integrated development envi-ronment for deploying and visualizing defeasible logic rule bases on top of RDF Schema ontologies. The system consists of a number of subcomponents, which, though developed autonomously, are combined efficiently, forming a flexible framework. The system employs a defeasible reasoning system that supports di-rect importing and processing of RDF data and RDF Schema ontologies as well as a number of user-friendly rule base and ontology visualization modules. DKE-Kontopoulos.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Edatak%2E2008%2E02%2E005 Semantic Web Reasoning Defeasible Logic Visual Representation Rule Editor Rule Markup Languages Use of Multimedia and the World Wide Web in Civil Engineering Learnin 1 Journal of Professional Issues in Engineering Education and Practice American Society for Civil Engineering 131 (2) 2005 9 129-137 The case study method, which has been proven to be a very useful learning tool, can be further enhanced with the use of multimedia and the World Wide Web. This paper demonstrates multimedia and Web-based enhancement with the design and construction of a port, a large-scale civil engineering project. The main purpose was to create an educational tool that brings into the classroom a "real-life" design and construction problem, including the construction field, operation of equipment, and details of construction methods. This enables civil engineering students to better understand the details of the planning, design, and construction of a complicated project. Furthermore, through the use of evaluation tests, feedback on the students' understanding of the case study can be provided to both the students and the educator. This application can be expanded beyond an academic environment for use as a learning tool in a business environment, which may be especially beneficial for new engineers. An Ensemble of Classifiers for coping with Recurring Contexts in Data Streams 2 18th European Conference on Artificial Intelligence IOS Press 2008 0 763-764 This paper proposes a general framework for classifying data streams by exploiting incremental clustering in order to dynamically build and update an ensemble of incremental classifiers. To achieve this, a transformation function that maps batches of examples into a new conceptual feature space is proposed. An incremental clustering algorithm is then applied in order to group different concepts and identify reoccurring themes. The ensemble is produced by creating and training an incremental classifier for every concept discovered in the data stream. An experimental study is performed using three new real-world datasets from the text domain, a basic implementation of the proposed framework and three baseline methods for dealing with drifting concepts. Results are promising and encourage further investigation. katakis_ecai08_short.pdf Patras, Greece text classification ensemble data stream concept drift classification Focused Ensemble Selection: A Diversity-Based Method for Greedy Ensemble Selection 2 18th European Conference on Artificial Intelligence IOS Press 2008 5 117-121 Ensemble selection deals with the reduction of an ensemble of predictive models in order to improve its efficiency and predictive performance. A number of ensemble selection methods that are based on greedy search of the space of all possible ensemble subsets have recently been proposed. This paper contributes a novel method, based on a new diversity measure that takes into account the strength of the decision of the current ensemble. Experimental comparison of the proposed method, dubbed Focused Ensemble Selection (FES), against state-of-the-art greedy ensemble selection methods shows that it leads to small ensembles with high predictive performance. partalas-ecai08.pdf Patras, Greece ensemble selection Reinforcement Learning with Classifier Selection for Focused Crawling 2 18th European Conference on Artificial Intelligence IOS Press 2008 0 759-760 Focused crawlers are programs that wander in the Web, using its graph structure, and gather pages that belong to a specific topic. The most critical task in Focused Crawling is the scoring of the URLs as it designates the path that the crawler will follow, and thus its effectiveness. In this paper we propose a novel scheme for assigning scores to the URLs, based on the Reinforcement Learning (RL) framework. The proposed approach learns an adaptive behavior of selecting the best classifier for ordering the URLs. This formulation reduces the size of the search space for the RL method and makes the problem tractable. We evaluate the proposed approach on-line on a number of topics, which offers us a realistic view of its performance, comparing it also with a RL method and a simple but effective classifier-based crawler. The results demonstrate the strength of the proposed approach. HOOPO: A Hybrid Object-Oriented Integration of Production Rules and OWL Ontologies 2 18th European Conference on Artificial Intelligence, pp, 729-730 IOS Press 2008 0 729-730 meditskos-ecai2008-hoopo.pdf Rule-based OWL Ontology Reasoning Using Dynamic ABOX Entailments 2 18th European Conference on Artificial Intelligence, pp. 731-732 IOS Press 2008 0 731-732 meditskos-ecai2008-dynamic.pdf Greedy Regression Ensemble Selection: Theory and an Application to Water Quality 1 Information Sciences Elsevier 178(20) 2008 0 3867-3879 This paper studies the greedy ensemble selection family of algorithms for ensembles of regression models. These algorithms search for the globally best subset of regressors by making local greedy decisions for changing the current subset. We abstract the key points of the greedy ensemble selection algorithms and present a general framework, which is applied to an application domain with important social and commercial value: water quality prediction. partalas-is178(20).pdf (accepted) http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eins%2E2008%2E05%2E025 Negotiation of meaning and digital textbooks in the CLIL classroom. 3 Diverse Contexts – Converging Goals, CLIL in Europe Peter Lang Publications David Marsh, Dieter Wolff 2007 11 319-330 A novel approach for incremental uncertainty rule generation from databases with missing values handling: Application to dynamic medical databases. 1 Med Inform Internet Med 30 (3) 2005 15 211-225 Current approaches for mining association rules usually assume that the mining is performed in a static database, where the problem of missing attribute values does not practically exist. However, these assumptions are not preserved in some medical databases, like in a home care system. In this paper, a novel uncertainty rule algorithm is illustrated, namely URG-2 (Uncertainty Rule Generator), which addresses the problem of mining dynamic databases containing missing values. This algorithm requires only one pass from the initial dataset in order to generate the item set, while new metrics corresponding to the notion of Support and Confidence are used. URG-2 was evaluated over two medical databases, introducing randomly multiple missing values for each record's attribute (rate: 5?-?20% by 5% increments) in the initial dataset. Compared with the classical approach (records with missing values are ignored), the proposed algorithm was more robust in mining rules from datasets containing missing values. In all cases, the difference in preserving the initial rules ranged between 30% and 60% in favour of URG-2. Moreover, due to its incremental nature, URG-2 saved over 90% of the time required for thorough re-mining. Thus, the proposed algorithm can offer a preferable solution for mining in dynamic relational databases. Predicting missing values in a home care database using an adaptive uncertainty rule method. 1 Methods Inf Med 44 (5) 2005 8 639-646 Contemporary literature illustrates an abundance of adaptive algorithms for mining association rules. However, most literature is unable to deal with the peculiarities, such as missing values and dynamic data creation, that are frequently encountered in fields like medicine. This paper proposes an uncertainty rule method that uses an adaptive threshold for filling missing values in newly added records. A new approach for mining uncertainty rules and filling missing values is proposed, which is in turn particularly suitable for dynamic databases, like the ones used in home care systems. Methods: In this study, a new data mining method named FiMV (Filling Missing Values) is illustrated based on the mined uncertainty rules. Uncertainty rules have quite a similar structure to association rules and are extracted by an algorithm proposed in previous work, namely AURG (Adaptive Uncertainty Rule Generation). The main target was to implement an appropriate method for recovering missing values in a dynamic database, where new records are continuously added, without needing to specify any kind of thresholds beforehand. Results: The method was applied to a home care monitoring system database. Randomly, multiple missing values for each record's attributes (rate 5-20% by 5% increments) were introduced in the initial dataset. FiMV demonstrated 100% completion rates with over 90% success in each case, while usual approaches, where all records with missing values are ignored or thresholds are required, experienced significantly reduced completion and success rates. Conclusions: It is concluded that the proposed method is appropriate for the data-cleaning step of the Knowledge Discovery process in databases. The latter, containing much significance for the output efficiency of any data mining technique, can improve the quality of the mined information. http%3A%2F%2Fwww%2Eschattauer%2Ede%2Findex%2Ephp%3Fid%3D1090%26L%3D1%26artikel%3D14276%26cHash%3Dab2e1777eb A Taxonomy and Short Review of Ensemble Selection 2 ECAI, Workshop on Supervised and Unsupervised Ensemble Methods and Their Applications 2008 6 Ensemble selection deals with the reduction of an ensemble of predictive models in order to improve its efficiency and predictive performance. The last 10 years a large number of very diverse ensemble selection methods have been proposed. In this paper we make a first approach to categorize them into a taxonomy. We also present a short review of some of these methods. We particularly focus on a category of methods that are based on greedy search of the space of all possible ensemble subsets. Such methods use different directions for searching this space and different measures for evaluating the available actions at each state. Some use the training set for subset evaluation, while others a separate validation set. This paper abstracts the key points of these methods and offers a general framework of the greedy ensemble selection algorithm, discussing its important parameters and the different options for instantiating these parameters. Patras, Greece ensemble selection Multilabel Classification of Music into Emotions 2 Proc. 9th International Conference on Music Information Retrieval (ISMIR 2008) 2008 6 325-330 In this paper, the automated detection of emotion in music is modeled as a multilabel classification task, where a piece of music may belong to more than one class. Four algorithms are evaluated and compared in this task. Furthermore, the predictive power of several audio features is evaluated using a new multilabel feature selection method. Experiments are conducted on a set of 593 songs with 6 clusters of music emotions based on the Tellegen-Watson-Clark model. Results provide interesting insights into the quality of the discussed algorithms and features. tsoumakas-ismir08.pdf Philadelphia, PA, USA Pruning an Ensemble of Classifiers via Reinforcement Learning 1 Neurocomputing Elsevier 72(7-9) 2009 0 1900-1909 This paper studies the problem of pruning an ensemble of classifiers from a Reinforcement Learning perspective. It contributes a new pruning approach that uses the Q-learning algorithm in order to approximate an optimal policy of choosing whether to include or exclude each classifier from the ensemble. Extensive experimental comparisons of the proposed approach against state-of-the-art pruning and combination methods show very promising results. Additionally, we present an extension that allows the improvement of the solutions returned by the proposed approach over time, which is very useful in certain performance-critical domains. partalas08Neuro.pdf Ensemble Pruning Reinforcement Learning A Hybrid Multiagent Reinforcement Learning Approach using Strategies and Fusion 1 International Journal of Artificial Intelligence Tools (IJAIT) World Scientific 17 (5) 2008 18 945-961 Reinforcement Learning comprises an attractive solution to the problem of coordinating a group of agents in a Multiagent System, due to its robustness for learning in uncertain and unknown environments. This paper proposes a multiagent Reinforcement Learning approach, that uses coordinated actions, which we call strategies and a fusing process to guide the agents. To evaluate the proposed approach, we conduct experiments in the Predator-Prey domain and compare it with other learning techniques. The results demonstrate the efficiency of the proposed approach. masterMARL-ijaitRevised.pdf multi-agent reinforcement learning Visual Modeling of Defeasible Logic Rules with DR-VisMo 1 International Journal of Artificial Intelligence Tools (IJAIT) World Scientific 17(5) 2008 22 903-924 The standardization of the Semantic Web has reached as far as ontologies and ontology languages. However, in order for the full potential of the Semantic Web to be achieved, the ability of reasoning over the available information is also essential. Rules can assist in this affair and various logics have been proposed for the Semantic Web domain. One of them is defeasible reasoning that deals with incomplete and conflicting information. However, despite its solid mathematical notation, it may be confusing to end users. To confront this downside, we proposed a representation schema for defeasible logic rule bases, which is based on directed graphs that feature distinct node and connection types. This paper presents DR-VisMo, a defeasible logic rule base editor and visualization system that implements this representation approach. The system also features a stratification algorithm for visualizing rule bases that deals with decisions, regarding the arrangement of the various elements in the graph. DR-VisMo is implemented as part of VDR-DEVICE, an environment for modeling and deploying defeasible logic rule bases on top of RDF ontologies. IJAIT-final-kontopoulos.pdf Semantic Web Defeasible Reasoning Directed Graphs Visualization A Synergy of Planning and Ontology Concept Ranking for Semantic Web Service Composition 2 11th Ibero-American Conference on Artificial Intelligence Springer-Verlag H. Geffner et al. LNAI 5290 2008 0 42-51 This paper presents a prototype system that exploits planning and an ontology concept ranking algorithm for composing semantic Web services (PORSCE). The system exploits the inferencing capabilities of a Description Logics Reasoner in order to compute the subsumption hierarchy of the ontologies whose concepts are used in the OWL-S Profile descriptions as input and output concepts. The concept ranking algorithm is applied over this hierarchy in order to determine similar concepts based on different degrees of semantic matching relaxation, such as subclass or sibling hierarchical relationships. The domain independent planning system’s role is to semantically search the space of possible compositions of Web services, generating plans according to the desirable level of relaxation. porsce.pdf Lisbon, Portugal Combining a DL Reasoner and a Rule Engine for Improving Entailment-based OWL Reasoning 2 Proc. 7th International Semantic Web Conference (ISWC-2008), 26-30 Oct 2008, Karlsruhe, Germany, Springer, LNCS, Vol. 5318, pp. 277-292 2008 16 We introduce the notion of the mixed DL and entailment-based (DLE) OWL reasoning, defining a framework inspired from the hybrid and homogeneous paradigms for integration of rules and ontologies. The idea is to combine the TBox inferencing capabilities of the DL algorithms and the scalability of the rule paradigm over large ABoxes. Towards this end, we define a framework that uses a DL reasoner to reason over the TBox of the ontology (hybrid-like) and a rule engine to apply a domain-specific version of ABox-related entailments (homogeneous-like) that are generated by TBox queries to the DL reasoner. The DLE framework enhances the entailment-based OWL reasoning paradigm in two directions. Firstly, it disengages the manipulation of the TBox semantics from any incomplete entailment-based approach, using the efficient DL algorithms. Secondly, it achieves faster application of the ABox-related entailments and efficient memory usage, comparing it to the conventional entailment-based approaches, due to the low complexity and the domainspecific nature of the entailments. med-iswc08.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2FDLEJena%2F DLEJena+Web+Site http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2F5k226206h2747708%2F%3Fp%3Daa3b384f084b4486ba3a7db06e0e40c6%26pi%3D17 An Empirical Study of Lazy Multilabel Classification Algorithms 2 Proc. 5th Hellenic Conference on Artificial Intelligence (SETN 2008) Springer LNAI Vol.5138 2008 6 401-406 Multilabel classification is a rapidly developing field of machine learning. Despite its short life, various methods for solving the task of multilabel classification have been proposed. In this paper we focus on a subset of these methods that adopt a lazy learning approach and are based on the traditional k-nearest neighbor (kNN) algorithm. Two are our main contributions. Firstly, we implement BRkNN, an adaptation of the kNN algorithm for multilabel classification that is conceptually equivalent to using the popular Binary Relevance problem transformation method in conjunction with the kNN algorithm, but much faster. We also identify two useful extensions of BRkNN that improve its overall predictive performance. Secondly, we compare this method against two other lazy multilabel classification methods, in order to determine the overall best performer. Experiments on different real-world multilabel datasets, using a variety of evaluation metrics, expose the advantages and limitations of each method with respect to specific dataset characteristics. spyromitros-setn08.pdf Syros, Greece http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2F978%2D3%2D540%2D87881%2D0%5F40 Effective and Efficient Multilabel Classification in Domains with Large Number of Labels 2 Proc. ECML/PKDD 2008 Workshop on Mining Multidimensional Data (MMD'08) 2008 15 This paper contributes a novel algorithm for effective and computationally efficient multilabel classification in domains with large label sets L. The HOMER algorithm constructs a Hierarchy Of Multilabel classifiERs, each one dealing with a much smaller set of labels compared to L and a more balanced example distribution. This leads to improved predictive performance along with linear training and logarithmic testing complexities with respect to |L|. Label distribution from parent to children nodes is achieved via a new balanced clustering algorithm, called balanced k means. tsoumakas-mmd08.pdf Antwerp, Belgium Polyadenylation Site Prediction Using Interesting Emerging Pattern 2 8th IEEE International Conference on Bioinformatics and Bioengineering IEEE 2008 6 This paper presents a study on polyadenylation site prediction in mRNA sequences. We describe a method, called PolyA-EP, that we developed for predicting polyadenylation sites and we present a systematic study of the problem of recognizing mRNA 3΄ ends which contain a polyadenylation site using the proposed method. PolyA-EP exploits the advantages of emerging patterns, namely high understandability and discriminating power and can be used for both descriptive and predictive analysis. In particular, PolyA-EP is a parameterizable tool that can be used in order to extract interesting emerging patterns for describing or predicting polyadenylation sites. Moreover, the extracted emerging patterns can span across many elements around the polyadenylation site. We discuss the results of the experiments we conducted with Arabidopsis thaliana sequences drawing important conclusions and finally we propose a framework that improves the accuracy of polyadenylation site prediction. Tzanis_BIBE08.pdf Athens, Greece prediction polyadenylation mRNA messanger RNA emerging patterns data mining Arabidopsis thaliana Multilabel Text Classification for Automated Tag Suggestion 2 Proceedings of the ECML/PKDD 2008 Discovery Challenge 2008 9 The increased popularity of tagging during the last few years can be mainly attributed to its embracing by most of the recently thriving user-centric content publishing and management Web 2.0 applications. However, tagging systems have some limitations that have led researchers to develop methods that assist users in the tagging process, by automatiically suggesting an appropriate set of tags. We have tried to model the automated tag suggestion problem as a multilabel text classification task in order to participate in the ECML/PKDD 2008 Discovery Challenge. katakis_ecmlpkdd08_challenge.pdf Antwerp Machine Learning and Data Mining in Bioinformatics 3 Handbook of Research on Innovations in Database Technologies and Applications: Current and Future Trends IGI Global L.C. Rivero, J.H. Door, V.E. Ferraggine 2009 0 Tzanis_Handbook09.pdf Visualization of Proofs in Defeasible Logic 2 Proceedings of the 2008 International Symposium on Rule Interchange and Applications (RuleML-2008) Springer Nick Bassiliades, Guido Governatori, Adrian Paschke LNCS, Vol. 5321 2008 14 197-210 The development of the Semantic Web proceeds in steps, building each layer on top of the other. Currently, the focus of research efforts is concentrated on logic and proofs, both of which are essential, since they will allow systems to infer new knowledge by applying principles on the existing data and explain their actions. Research is shifting towards the study of non-monotonic systems that are capable of handling conflicts among rules and reasoning with partial information. As for the proof layer of the Semantic Web, it can play a vital role in increasing the reliability of Semantic Web systems, since it will be possible to provide explanations and/or justifications of the derived answers. This paper reports on the implementation of a system for visualizing proof explanations on the SemanticWeb. The proposed system applies defeasible logic, a member of the non-monotonic logics family, as the underlying inference system. The proof representation schema is based on a graph-based methodology for visualizing defeasible logic rule bases. kontopoulos-RuleML'08.pdf Orlando, Florida, USA Extending a Defeasible Reasoner with Modal and Deontic Logic Operators 2 Proceedings of Workshop on Logics for Intelligent Agents and Multi-Agent Systems (WLIAMAS 2008) at IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology IEEE Computer Society Press 2008 4 626-629 Defeasible logic is a non-monotonic formalism that deals with incomplete and conflicting information. Modal logic deals with necessity and possibility, exhib-iting defeasibility; thus, it is possible to combine defea-sible logic with modal operators. This paper reports on the extension of the DR-DEVICE defeasible reasoner with modal and deontic logic operators. The aim is a practical defeasible reasoner that will take advantage of the expressiveness of modal logics and the flexibility to define diverse agent types and behaviors. kontopoulos-WLIAMAS'08.pdf Sydney, Australia defeasible reasoning modal logic deontic logic reasoning engine modal DL Rule-based OWL Reasoning Systems: Implementations, Strengths and Weaknesses 3 Handbook of Research on Emerging Rule-Based Languages and Technologies: Open Solutions and Approaches IGI Global Adrian Giurca, Dragan Gasevic, Kuldar Taveter ISBN 978-1-60566-402-6 2009 0 124-148 This chapter is focused on the basic principles behind the utilization of rules in order to perform reasoning about the Web Ontology Language (OWL), a Description Logic-based language that is the W3C recommendation for creating and sharing ontologies in the Semantic Web. More precisely, we elaborate on the entailment-based OWL reasoning (EBOR) paradigm, which is based on the utilization of RDF/RDFS and OWL entailment rules that run on a rule engine, applying the formal semantics of the ontology language. To this end, seven EBOR systems are described and compared, analyzing the different approaches. Despite the closed rule environment, which comes in contrast with the open nature of the Semantic Web, and the fact that OWL semantics are partially mapped into rules, the rule-based OWL reasoning paradigm can give great potentials in the Semantic Web, enabling the utilization of rule engines on top of ontology information. medi-HRER2008.pdf http%3A%2F%2Fwww%2Eigi%2Dglobal%2Ecom%2FBookstore%2FTitleDetails%2Easpx%3FTitleId%3D465 ontologies and rules OWL reasoning On the Combination of Textual and Semantic Descriptions for Automated Semantic Web Service Classification 2 Proceedings of the 5th IFIP Conference on Artificial Intelligence Applications and Innovations (AIAI 2009) Springer 2009 10 Semantic Web services have emerged as the solution to the need for automating several aspects related to service-oriented architectures, such as service discovery and composition, and they are realized by combining Semantic Web technologies and Web service standards. In the present paper, we tackle the problem of automated classification of Web services according to their application domain taking into account both the textual description and the semantic annotations of OWL-S advertisements. We present results that we obtained by applying machine learning algorithms on textual and semantic descriptions separately and we propose methods for increasing the overall classification accuracy through an extended feature vector and an ensemble of classifiers. katakis-aiai09.pdf http%3A%2F%2Fmlkd%2Ecsd%2Eauth%2Egr%2Fws%2Ehtml Automated+Web+Service+Classification Thessaloniki web service semantic web classification machine learning data mining owls text mining text classification OntoLife: An Ontology for Semantically Managing Personal Information 2 Proceedings of the 5th IFIP Conference on Artificial Intelligence Applications and Innovations (AIAI 2009) Springer 2009 7 127-134 Knowledge management, along with the more recent trend of personal knowledge management, have attracted the attention of researchers from various angles, one of which is the Semantic Web. Since semantics promise to add value to the interaction of users with computers, many applications try to incorporate them. Ontologies, the primary knowledge representation tool for the Semantic Web, can play a significant role in semantically managing personal knowledge. The scope of this paper focuses on addressing the issue of effective personal knowledge management, by proposing an ontology for modelling the domain of biographical events. The proposed ontology also undergoes a thorough evaluation, based on specific criteria presented in the literature. AIAI'09-kargioti-kontopoulos-bassiliades.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fontologies%2Fontolist%2Ehtml%23ontolife OntoLife+Ontology Thessaloniki knowledge management personal knowledge management semantic wikis personal information ontology Visualizing RDF Documents 2 Proceedings of the 5th IFIP Conference on Artificial Intelligence Applications and Innovations (AIAI 2009) Springer 2009 6 151-156 The Semantic Web (SW) is an extension to the current Web, enhancing the available information with semantics. RDF, one of the most prominent standards for representing meaning in the SW, offers a data model for referring to objects and their interrelations. Managing RDF documents, however, is a task that demands experience and expert understanding. Tools have been developed that al-leviate this drawback and offer an interactive graphical visualization environment. This paper studies the visualization of RDF documents, a domain that exhibits many applications. The most prominent approaches are presented and a novel graph-based visualization software application is also demonstrated. AIAI'09-athanassiades-kontopoulos-bassiliades.pdf Thessaloniki RDF visualization display-at-once navigational-centric centric-graph-at-once An Ensemble Pruning Primer 3 Supervised and Unsupervised Methods and Their Applications to Ensemble Methods (SUEMA 2009) Springer Verlag Oleg Okun and Giorgio Valentini Volume 245/2009 2009 13 1-13 Ensemble pruning deals with the reduction of an ensemble of predictive models in order to improve its e±ciency and predictive performance. The last 12 years a large number of ensemble pruning methods have been proposed. This work proposes a taxonomy for their organization and reviews important representative methods of each category. It abstracts their key components and discusses their main advantages and disadvantages.We hope that this work will serve as a good starting point and reference for researchers working on the development of new ensemble pruning methods. tsoumakas09.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2F978%2D3%2D642%2D03999%2D7 Transferring Experience in Reinforcement Learning through Task Decomposition 2 International Conference on Autonomous Agents and Multiagent Systems 2009 0 aamas09Master-CR.pdf Tracking Recurring Contexts using Ensemble Classifiers: An Application to Email Filtering 1 Knowledge and Information Systems Springer 22(3) 2010 23 371-391 Concept drift constitutes a challenging problem for the machine learning and data mining community that frequently appears in real world stream classification problems. It is usually defined as the unforeseeable concept change of the target variable in a prediction task. In this paper, we focus on the problem of recurring contexts, a special sub-type of concept drift, that has not yet met the proper attention from the research community. In the case of recurring contexts, concepts may re-appear in future and thus older classification models might be beneficial for future classifications. We propose a general framework for classifying data streams by exploiting stream clustering in order to dynamically build and update an ensemble of incremental classifiers. To achieve this, a transformation function that maps batches of examples into a new conceptual representation model is proposed. The clustering algorithm is then applied in order to group batches of examples into concepts and identify recurring contexts. The ensemble is produced by creating and maintaining an incremental classifier for every concept discovered in the data stream. An experimental study is performed using a) two new real-world concept drifting datasets from the email domain, b) an instantiation of the proposed framework and c) five methods for dealing with drifting concepts. Results indicate the effectiveness of the proposed representation and the suitability of the concept-specific classifiers for problems with recurring contexts. katakis_kais_09.pdf http%3A%2F%2Fmlkd%2Ecsd%2Eauth%2Egr%2Fconcept%5Fdrift%2Ehtml Datasets http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2Fs10115%2D009%2D0206%2D2 data streams classification concept drift text mining text classification recurring contexts recurring themes text streams email mining email classification An Empirical Study Of Multi-Label Learning Methods For Video Annotation 2 7th International Workshop on Content-Based Multimedia Indexing IEEE 2009 6 This paper presents an experimental comparison of different approaches to learning from multi-labeled video data. We compare state-of-the-art multi-label learning methods on the Mediamill Challenge dataset. We employ MPEG-7 and SIFT-based global image descriptors independently and in conjunction using variations of the stacking approach for their fusion. We evaluate the results comparing the different classifiers using both MPEG-7 and SIFT-based descriptors and their fusion. A variety of multi-label evaluation measures is used to explore advantages and disadvantages of the examined classifiers. Results give rise to interesting conclusions. tsoumakas-cbmi09.pdf Chania, Crete Mining Multi-label Data 3 Data Mining and Knowledge Discovery Handbook, Part 6 Springer O. Maimon, L. Rokach 2nd edition 2010 19 667-685 tsoumakas09-dmkdh.pdf http%3A%2F%2Fmlkd%2Ecsd%2Eauth%2Egr%2Fmultilabel%2Ehtml Mining+from+multi%2Dlabel+data+web+page+%40+MLKD http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2F978%2D0%2D387%2D09822%2D7%2F Structural and Role-Oriented Web Service Discovery with Taxonomies in OWL-S 1 IEEE Transactions on Knowledge and Data Engineering (TKDE) IEEE 22(2) 2010 12 278-290 In this paper, we describe and evaluate a Web service discovery framework using OWL-S advertisements, combined with the distinction between service and Web service of the WSMO Discovery Framework. More specifically, we follow the Web service discovery model, which is based on abstract and lightweight semantic Web service descriptions, using the Service Profile ontology of OWL-S. Our goal is to determine fast an initial set of candidate Web services for a specific request. This set can then be used in more fine-grained discovery approaches, based on richer Web service descriptions. Our Web service matchmaking algorithm extends object-based matching techniques used in Structural Case-based Reasoning, allowing (a) the retrieval of Web services not only based on subsumption relationships, but exploiting also the structural information of OWL ontologies, and (b) the exploitation of Web services classification in Profile taxonomies, performing domain-dependent discovery. Furthermore, we describe how the typical paradigm of Profile input/output annotation with ontology concepts can be extended, allowing ontology roles to be considered as well. We have implemented our framework in the OWLS-SLR system, which we extensively evaluate and compare to the OWLS-MX matchmaker. meditskos-TKDE.pdf http%3A%2F%2Fdoi%2Eieeecomputersociety%2Eorg%2F10%2E1109%2FTKDE%2E2009%2E89 A Process-Oriented Ontology-Based Knowledge Management System for Facilitating Operational Procedures in Public Administration 1 Expert Systems with Applications Elsevier Vol. 36, Iss. 3, Part 1 2009 24 4467-4478 Public organizations produce daily a great volume of administrative documents, in order to fulfill their mission. This requires the use of a certain, unique for each procedure, legal framework. This article adopts a process oriented approach, through a web-based knowledge management system that provides this legal framework in an up-to-date and accurate manner. The system also supports the interpretation of the legal framework, supplying civil servants, citizens and businesses with precedents and opinions. The system employs an ontology in OWL for representing the public administration structure and any kind of document that flows among administrative units, during the execution of the procedures, which are mapped into OWL-S service models. eswa-savvas-final.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fontologies%2Fontolist%2Ehtml%23gpap Public+Administration+Ontology http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2008%2E05%2E022 public administration e-Government process-oriented knowledge management ontology-based knowledge management OWL OWL-S Applying adaptive prediction to sea-water quality measurements 1 Expert Systems with Applications Elsevier Vol. 36, Iss. 3, Part 2 2009 20 6773-6779 This study explores the possibility of using adaptive filters to predict sea-water quality indicators such as water temperature, pH and dissolved oxygen based on measurements produced by an under-water measurement set-up. Two alternative adaptive approaches are tested, namely a projection algorithm and a least squares algorithm. These algorithms were chosen for comparison because they are widely used prediction algorithms. The results indicate that if the measurements remain reasonably stationary, it is possible to make one-day ahead predictions, which perform better than the prediction that the value of a certain quality variable tomorrow is going to be equal to the value today. SWA-hatzikos-final.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2008%2E08%2E051 Projection algorithm Least square algorithm one-day ahead prediction A collision detection and resolution multi agent approach using utility functions 2 4th Balkan Conference in Informatics (BCI-09) IEEE Computer Society P. Kefalas, D. Stamatis, C. Douligeris 2009 5 3-7 Free Flight is the concept introduced by NASA and FAA in order to change the aviation of the 21st century, allowing for pilots to choose dynamically (“on the fly”) their nominal paths. Despite its many advantages as opposed to today’s situation, the free flight concept raises many new issues that need to be addressed before being applicable, with the main interest focusing on conflict detection and resolution (CD&R), in order to ensure the safety of the aircrafts. In this paper we present a decentralized CD&R approach using agents, as well as a general multi-agent framework where not only the proposed but new agent-based approaches may be implemented and tested. bci09.pdf Thessaloniki, Greece, September 17th-19th 2009 http%3A%2F%2Fwww%2Ecity%2Eacademic%2Egr%2Fspecial%2Facademix%2Fevents%2Fbci%2Fbci09%2Findex%2Ehtml Collision detection and resolution Multi-Agent systems Utility functions A Trusted Defeasible Reasoning Service for Brokering Agents in the Semantic Web 2 3rd International Symposium on Intelligent Distributed Computing (IDC 2009), 13-14 October Springer Berlin / Heidelberg Vol. 237 2009 6 243-248 Based on the plethora of proposals and standards for logic- and rule-based reasoning for the Semantic Web (SW), a key factor for the success of SW agents is interoperability of reasoning tasks. This paper reports on the first steps towards a framework for interoperable reasoning among agents in the SW that deploys third-party trusted reasoning services. This way, agents can exchange their arguments, without the need to conform to a common rule or logic paradigm – via an external reasoning service, the receiving agent can grasp the semantics of the received rule set. The paper presents how a multi-agent system was extended with a third-party trusted defeasible reasoning service, which offers agents the ability of reasoning with incomplete and inconsistent information. In addition, a brokering trade scenario is presented that illustrates the usability of the approach. idc2009-kravari_et_al.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald EMERALD Ayia Napa, Cyprus intelligent agents multi-agent system defeasible reasoning brokering agents Towards a Knowledge-based Framework for Agents Interacting in the Semantic Web 2 2009 IEEE/WIC/ACM International Conference on Intelligent Agent Technology (IAT'09), 15-18 September Vol. 2 2009 4 482-485 The Semantic Web aims at making Web content un-derstandable both for people and machines. Although intelligent agents can assist towards this vision, they do not have to conform to a common rule or logic paradigm. This paper reports on the first steps towards a framework for interoperating knowledge-based intel-ligent agents. A multi-agent system was extended with defeasible reasoning and a reusable agent model is proposed for customizable agents, equipped with a knowledge base and a Jess rule engine. Two use case scenarios display the integration of these technologies. iat2009-kravari_et_al.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald EMERALD Milan, Italy intelligent agents multi-agent system defeasible reasoning brokering agents agent negotiation Clustering Based Multi-Label Classification for Image Annotation and Retrieval 2 2009 IEEE International Conference on Systems, Man, and Cybernetics IEEE 2009 6 This paper presents a novel multi-label classification framework for domains with large numbers of labels. Automatic image annotation is such a domain, as the available semantic concepts are typically hundreds. The proposed framework comprises an initial clustering phase that breaks the original training set into several disjoint clusters of data. It then trains a multi-label classifier from the data of each cluster. Given a new test instance, the framework first finds the nearest cluster and then applies the corresponding model. Empirical results using two clustering algorithms, four multi-label classification algorithms and three image annotation data sets suggest that the proposed approach can improve the performance and reduce the training time of standard multi-label classification algorithms, particularly in the case of large number of labels. tsoumakas-smc09.pdf On the combination of two decompositive multi-label classification methods 2 Workshop on Preference Learning, ECML PKDD 09 Eyke Hullermeir, Johannes Furnkranz 2009 114-133 In this paper, we compare and combine two approaches for multi-label classification that both decompose the initial problem into sets of smaller problems. The Calibrated Label Ranking approach is based on interpreting the multi-label problem as a preference learning problem and decomposes it into a quadratic number of binary classifiers. The HOMER approach reduces the original problem into a hierarchy of considerably simpler multi-label problems. Experimental results indicate that the use of HOMER is beneficial for the pairwise preference-based approach in terms of computational cost and quality of prediction. tsoumakas-pl09.pdf Bled, Slovenia Correlation-Based Pruning of Stacked Binary Relevance Models for Multi-Label Learning 2 Proceedings of the 1st International Workshop on Learning from Multi-Label Data (MLD'09) G. Tsoumakas, Min-Ling Zhang, Zhi-Hua Zhou 2009 16 101-116 Binary relevance (BR) learns a single binary model for each different label of multi-label data. It has linear complexity with respect to the number of labels, but does not take into account label correlations and may fail to accurately predict label combinations and rank labels according to relevance with a new instance. Stacking the models of BR in order to learn a model that associates their output to the true value of each label is a way to alleviate this problem. In this paper we propose the pruning of the models participating in the stacking process, by explicitly measuring the degree of label correlation using the phi coefficient. Exploratory analysis of phi shows that the correlations detected are meaningful and useful. Empirical evaluation of the pruning approach shows that it leads to substantial reduction of the computational cost of stacking and occasional improvements in predictive performance. tsoumakas-mld09.pdf Bled, Slovenia Semantic Web Service Composition Using Planning and Ontology Concept Relevance 2 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology 2009 4 418-421 This paper presents PORSCE II, a system that combines planning and ontology concept relevance for automatically composing semantic web services. The presented approach includes transformation of the web service composition problem into a planning problem, enhancement with semantic awareness and relaxation and solution through external planners. The produced plans are visualized and their accuracy is assessed. wii-porsce.pdf Milan, Italy PORSCE II: Using Planning for Semantic Web Service Composition 2 3rd International Competition on Knowledge Engineering for Planning and Scheduling, ICAPS-09 Roman Bartak, Simone Frattini, Lee McCluskey 2009 8 38-45 This paper presents PORSCE II, an integrated system that performs automatic semantic web service composition through planning. In order to achieve that, an essential step is the translation of the web service composition problem into a planning problem. The planning problem is then solved using external domain-independent planning systems, and the solutions are visualized and evaluated. The system exploits semantic information to enhance the translation and planning processes. porsce-ickeps.pdf Thessaloniki, Greece DLEJena: A Practical Forward-Chaining OWL 2 RL Reasoner Combining Jena and Pellet 1 Journal of Web Semantics Elsevier 8(1) 2010 12 89-94 This paper describes DLEJena, a practical reasoner for the OWL 2 RL profile that combines the forward-chaining rule engine of Jena and the Pellet DL reasoner. This combination is based on rule templates, instantiating at run-time a set of ABox OWL 2 RL/RDF Jena rules dedicated to a particular TBox that is handled by Pellet. The goal of DLEJena is to handle efficiently, through instantiated rules, the OWL 2 RL ontologies under direct semantics, where classes and properties cannot be at the same time individuals. The TBox semantics are treated by Pellet, reusing in that way efficient and sophisticated TBox DL reasoning algorithms. The experimental evaluation shows that DLEJena achieves more scalable ABox reasoning than the direct implementation of the OWL 2 RL/RDF rule set in the Jena's production rule engine, which is the main target of the system. DLEJena can be also used as a generic framework for applying an arbitrary number of entailments beyond the OWL 2 RL profile. meditskos-jws09.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Ewebsem%2E2009%2E11%2E001 A visual programming system for automated problem solving 1 Expert Systems with Applications Elsevier 37(6) 2010 15 4611-4625 Although new AI planning algorithms and techniques are being developed and improved rapidly, there is a lack of efficient and easy to use systems able to incorporate and utilize them. Furthermore, while visual representation facilitates design, maintenance and comprehension of planning domains and problems, very few systems incorporate it. This paper presents VLEPPO, an integrated system aiming at visually modeling planning domains and problems through a convenient graphical interface, while maintaining compatibility with the Planning Domain Definition Language (PDDL), with import and export features. Solutions to planning problems can be obtained by invoking different planners employing the web services technology. The demonstration of the system is performed through a case study involving web service composition viewed as a planning problem. eswaPORSCEII.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2009%2E12%2E047 An Intelligent System for Monitoring and Predicting Water Quality 2 Proceedings of the European conference TOWARDS eENVIRONMENT 2009 9 534-542 In this paper we present an intelligent system for monitoring and predicting water quality, whose main aim is to help the authorities in the "decision-making" process in the battle against the pollution of the aquatic environment, which is very vital for the public health and the economy of Northern Greece. Two sensor-telematic networks for collecting water quality measurements in real time (Andromeda, for sea waters, and Interrisk, for surface/fresh waters) were developed and deployed. Sensor readings (water temperature, pH, dissolved oxygen, conductance, turbidity, sea currents, and salinity) are transmitted to a main station for processing and storage. The intelligent system monitors sensor data, reasons, using fuzzy logic, about the current level of water suitability for various aquatic uses, such as swimming and piscicultures, and flags out appropriate alerts. Furthermore, the system employs Machine Learning and Adaptive Filtering techniques and algorithms which successfully predict measurements a day ahead, as well as techniques to incorporate the window of past values in order to be able to make a more precise prediction. The results showed that these algorithms can help make accurate predictions one day ahead and are better than the naive prediction that the value will be similar to today. eEnvironment09.pdf Prague, Czech Republic, March 2009 http%3A%2Fwww%2Ee%2Denvi2009%2Eorg%2Fproceedings%2F Sensor Network Pollution Monitoring Aquatic Uses Water quality Pollution Prediction Pollution Pre diction Prediction Knowledge-Based System EMERALD: A Multi-Agent System for Knowledge-based Reasoning Interoperability in the Semantic Web 2 6th Hellenic Conference on Artificial Intelligence (SETN 2010) Springer Berlin / Heidelberg LNCS, Vol. 6040/2010 2010 10 173-182 The Semantic Web aims at augmenting the WWW with meaning, assisting people and machines in comprehending Web content and better satisfying their requests. Intelligent agents are considered to be greatly favored by Semantic Web technologies, because of the interoperability the latter will achieve. One of the main problems in agent interoperation is the great variety in reasoning formalisms, as agents do not necessarily share a common rule or logic formalism. This paper reports on the implementation of EMERALD, a knowledge-based framework for interoperating intelligent agents in the Semantic Web. More specifically, a multi-agent system was developed on top of JADE, featuring trusted, third party reasoning services, a reusable agent prototype for knowledge-customizable agent behavior, as well as a reputation mechanism for ensuring trust in the framework. Finally, a use case scenario is presented that illustrates the viability of the proposed framework. setn10-kravari.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald%2Femerald%2Ehtml EMERALD Athens, Greece, 4-7 May http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2F978%2D3%2D642%2D12842%2D4%5F21 semantic web intelligent agents multi-agent system reasoning Semantic Awareness in Automated Web Service Composition through Planning 2 6th Hellenic Conference on Artificial Intelligence (SETN 2010) Springer LNAI, Vol. 6040 2010 10 123-132 PORSCE II is a framework that performs automatic web service composition by transforming the composition problem into AI planning terms and utilizing external planners to obtain solutions. A distinctive feature of the system is that throughout the entire process, it achieves semantic awareness by exploiting semantic information extracted from the OWL-S descriptions of the available atomic web services and the corresponding ontologies. This information is then used in order to enhance the planning domain and problem. Semantic awareness facilitates approximations when searching for suitable atomic services, as well as modification of the produced composite service. The alternatives for modification include the replacement of a certain atomic service that takes part in the composite service by an equivalent or a semantically relevant service, the replacement of an atomic service through planning, or the replanning from a certain point in the composite service. The system also provides semantic representation of the produced composite service. setn10-hatzi.pdf Athens, Greece, 4-7 May http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1007%2F978%2D3%2D642%2D12842%2D4%5F16 Automatic Web Service Composition Semantic Web Services AI Planning Semantic Awareness Semantic Matching Relaxation Towards Compositional Safety Analysis via Semantic Representation of Component Failure Behaviour 2 Proc. 8th Joint Conference on Knowledge - Based Software Engineering 2008 (JCKBSE 08) IOS Press M. Virvou and T. Nakamura 2008 0 405-414 In dependable systems engineering safety assessment of complex designs that involve software and hardware components is one of the most difficult tasks required. Due to the different modelling languages and models that are used for complementary tasks, the model and specification artefacts are not easily shared by the experts involved in the design process. Moreover, the structural and semantic differences of the used language representations open a possibility for inconsistencies between the corresponding models. This work explores the role of an ontology representation of component failure behaviour as a basis for automated model transformations, as well as a library of reusable knowledge artefacts to be used in different modelling languages and models. The presented approach was motivated by recent findings and requirements derived from European industrial-driven research and development projects. August 2008, Piraeus, Greece Ontology Model Driven Software Development Dependability Safety Analysis Rule Representation, Interchange and Reasoning on the Web 4 Proceedings of the 2008 International Symposium on Rule Interchange and Applications (RuleML-2008) Springer LNCS, Vol. 5321, ISBN: 978-3-540-88807-9 2008 0 243 This book constitutes the refereed proceedings of the International RuleML Symposium on Rule Interchange and Applications, RuleML 2008, held in Orlando, FL, USA, in October 2008 - collocated with the 11th International Business Rules Forum. The 10 revised full papers and 10 revised short papers presented together with 2 demo papers and the abstracts of 3 keynote lectures were carefully reviewed and selected from 35 submissions. The papers are organized in topical sections on rule engineering, rule-based methodologies and applications in policies, electronic contracts and security, rule representation languages and reasoning engines, rule-based methodologies and applications in distributed and heterogeneous environments, natural-language and graphical rule representation and processing, as well as exemplary contributions to the RuleML-2008 challenge. Orlando, Florida, USA, 30-31 Oct. http%3A%2F%2Fwww%2Espringer%2Ecom%2Fcomputer%2Finformation%2Bsystems%2Band%2Bapplications%2Fbook%2F978%2D3%2D540%2D88807%2D9 RuleML Web rules semantic Web rule interchange business rules integration of rules and ontologies rule-based environments hybrid rule systems AIAI 2009 Workshops Proceedings, Proceedings of the Workshops of the 5th IFIP Conference on Artificial Intelligence Applications and Innovations (AIAI-2009) 4 CEUR Workshop Proceedings, ISSN 1613-0073 Vol. 475 2009 0 349 The ever expanding abundance of information and computing power enables researchers and users to tackle highly interesting issues, such as applications providing personalized access and interactivity to multimodal information based on user preferences and semantic concepts or human-machine interface systems utilizing information on the affective state of the user. The general focus of the AIAI conference is to provide insights on how AI can be implemented in real world applications. During the main AIAI-2009 conference, four (4) Workshops, on various specific AI application areas, were held, with parallel sessions: - Workshop on Biomedical Informatics and Intelligent Approaches in the Support of Genomic Medicine (BMIINT) - Workshop on Artificial Intelligence Approaches for Biometric Template Creation and Multibiometrics Fusion (ArtIBio) - 2nd Workshop on Artificial Intelligence Techniques in Software Engineering (AISEW 2009) - Workshop on Artificial Intelligence Applications in Environmental Protection (AIAEP) This volume contains papers selected for presentation at the Workshops of the 5th IFIP Conference on Artificial Intelligence Applications and Innovations (AIAI 2009) being held from 23rd till 25th of April, in Thessaloniki, Greece. The Workshops were held on 24th (BMIINT, ArtIBio) and 25th of April (AISEW 2009, AIAEP). The IFIP AIAI 2009 conference is co-organized by the Aristotle University of Thessaloniki, by the University of Macedonia Thessaloniki and by the Democritus University of Thrace. AIAI 2009 is the official conference of the WG12.5 "Artificial Intelligence Applications" working group of IFIP TC12 the International Federation for Information Processing Technical Committee on Artificial Intelligence (AI). The response to call for workhsop proposals was satisfactory. Five workshop proposals were received; finally, four of them managed to receive a critical mass of submitted papers and make it to the AIAI-2009 conference. The wide range of topics and high level of contributions guarantees a very successful set of workshops. Thessaloniki, Greece, April 23-25 http%3A%2F%2Fceur%2Dws%2Eorg%2FVol%2D475 Artificial Intelligence Biomedical Informatics Genomic Medicine Biometric Template Creation Multibiometrics Fusion Software Engineering Environmental Protection An Ensemble Uncertainty Aware Measure for Directed Hill Climbing Ensemble Pruning 1 Machine Learning Springer 2010 23 This paper proposes a new measure for ensemble pruning via directed hill climbing, dubbed Uncertainty Weighted Accuracy (UWA), which takes into account the uncertainty of the decision of the current ensemble. Empirical results on 30 data sets show that using the proposed measure to prune a heterogeneous ensemble leads to significantly better accuracy results compared to state-of-the-art measures and other baseline methods, while keeping only a small fraction of the original models. Besides the evaluation measure, the paper also studies two other parameters of directed hill climbing ensemble pruning methods, the search direction and the evaluation dataset, with interesting conclusions on appropriate values. partalas-mlj10.pdf Random k-Labelsets for Multi-Label Classification 1 IEEE Transactions on Knowledge and Data Engineering IEEE 23(7) 2011 12 1079-1089 A simple yet effective multi-label learning method, called label powerset (LP), considers each distinct combination of labels that exist in the training set as a different class value of a single-label classification task. The computational efficiency and predictive performance of LP is challenged by application domains with large number of labels and training examples. In these cases the number of classes may become very large and at the same time many classes are associated with very few training examples. To deal with these problems, this paper proposes breaking the initial set of labels into a number of small random subsets, called {\em labelsets} and employing LP to train a corresponding classifier. The labelsets can be either disjoint or overlapping depending on which of two strategies is used to construct them. The proposed method is called RA$k$EL (RAndom $k$ labELsets), where $k$ is a parameter that specifies the size of the subsets. Empirical evidence indicate that RA$k$EL manages to improve substantially over LP, especially in domains with large number of labels and exhibits competitive performance against other high-performing multi-label learning methods. tsoumakas-tkde10.pdf http%3A%2F%2Fwww%2Ecomputer%2Eorg%2Fportal%2Fweb%2Fcsdl%2Fdoi%2F10%2E1109%2FTKDE%2E2010%2E164 Mining for Mutually Exclusive Gene Expressions 2 6th Hellenic Conference on Artificial Intelligence Springer Verlag S. Konstantopoulos, S. Perantonis, et al. 2010 0 Association rules mining is a popular task that involves the discovery of co-occurences of items in transaction databases. Several extensions of the traditional association rules mining model have been proposed so far, however, the problem of mining for mutually exclusive items has not been investigated. Such information could be useful in various cases in many application domains like bioinformatics (e.g. when the expression of a gene excludes the expression of another) In this paper, we address the problem of mining pairs and triples of genes, such that the presence of one excludes the presence of the other. First, we provide a concise review of the literature, then we define this problem, we propose a probability-based evaluation metric, and finally a mining algorithm that we apply on gene expression data gaining new biological insights. Tzanis_SETN10.pdf Athens, Greece Trusted Reasoning Services for Semantic Web Agents 1 Informatica: International Journal of Computing and Informatics Slovenian Society Informatika 34(4) 2010 12 429-440 The Semantic Web aims at enriching information with well-defined semantics, making it possible both for people and machines to understand Web content. Intelligent agents are the most prominent approach towards realizing this vision. Nevertheless, agents do not necessarily share a common rule or logic formalism, neither would it be realistic to attempt imposing specific logic formalisms in a rapidly changing world like the Web. Thus, based on the plethora of proposals and standards for logic- and rule-based reasoning for the Semantic Web, a key factor for the success of Semantic Web agents lies in the interoperability of reasoning tasks. This paper reports on the implementation of trusted, third party reasoning services wrapped as agents in a multi-agent system framework. This way, agents can exchange their arguments, without the need to conform to a common rule or logic paradigm – via an external reasoning service, the receiving agent can grasp the semantics of the received rule set. Finally, a use case scenario is presented that illustrates the viability of the proposed approach. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald%2Femerald%2Ehtml EMERALD http%3A%2F%2Fwww%2Einformatica%2Esi%2FPDF%2F34%2D4%2F04%5FKravari%2520%2D%2520Trusted%2520Reasoning%2520Services%2520for%2520Semantic%2520Web%2Epdf Semantic Web Intelligent Agents Multi-Agent Systems Reasoning Deploying a Semantically-Enabled Content Management System in a State University 2 Proceedings of EGOVIS 2010 Springer, Heidelberg K. N. Andersen et al. LNCS 6267 2010 8 257-264 Public institutes often face the challenge of managing vast volumes of administrative documents, a need that is often met via Content Management Systems (CMSs). CMSs offer various advantages, like separation of data struc-ture from presentation and variety in user roles, but also present certain disad-vantages, like inefficient keyword-based search facilities. The new generation of content management solutions imports the notion of semantics and is based on Semantic Web technologies, such as metadata and ontologies. The benefits include semantic interoperability, competitive advantages and dramatic cost re-duction. In this paper a leading Enterprise CMS is extended with semantic ca-pabilities for automatically importing and exporting ontologies. This functional-ity enables reuse of repository content, semantically-enabled search and interoperability with third-party applications. The extended system is deployed in semantically managing the large volumes of documents for a state university. Befa-EGOVIS'10.pdf semantic web metadata content management ontology Artificial Intelligence Applications and Innovations: Proceedings of the 5th IFIP Conference on Artificial Intelligence Applications and Innovations (AIAI'2009) 4 ISBN: 978-1441902207 Springer 2009 0 http%3A%2F%2Fwww%2Espringer%2Ecom%2Fcomputer%2Fai%2Fbook%2F978%2D1%2D4419%2D0220%2D7 System Architecture for a Smart University Building 2 ICANN '10 Intelligent Environmental Monitoring, Modelling and Management Systems for better QoL Workshop 2010 0 This paper presents a system architecture that provides smart building monitoring and management. The proposed solution integrates heterogeneous geographically disparate sensor networks and devices, and enables optimal operations of the building while reducing its energy footprint. The platform is based on Semantic Web Services composition using AI Planning, that integrates and manages WiFi, RFiD and ZigBee networks providing connectivity to the devices. The goal is to develop a model that follows the latest guidelines in the area of Information Communication Technologies (ICT) for sustainable growth, energy efficiency and better quality of life. thanosICANN10.pdf Thessaloniki Obtaining Bipartitions from Score Vectors for Multi-Label Classification 2 22nd International Conference on Tools with Artificial Intelligence, 27-29 October 2010. IEEE 2010 8 Multi-label classification is a popular learning task. However, some of the algorithms that learn from multi-label data, can only output a score for each label, so they cannot be readily used in applications that require bipartitions. In addition, several of the recent state-of-the-art multi-label classification algorithms, actually output a score vector primarily and employ one (sometimes simple) thresholding method in order to be able to output bipartitions. Furthermore, some approaches can naturally output both a score vector and a bipartition, but whether a better bipartition can be obtained through thresholding has not been investigated. This paper contributes a theoretical and empirical comparative study of existing thresholding methods, highlighting their importance for obtaining bipartitions of high quality. ioannou-ictai10.pdf Arras, France Instance-Based Ensemble Pruning via Multi-Label Classification 2 22nd International Conference on Tools with Artificial Intelligence, 27-29 October 2010. IEEE 2010 8 Ensemble pruning is concerned with the reduction of the size of an ensemble prior to its combination. Its purpose is to reduce the space and time complexity of the ensemble and/or to increase the ensemble’s accuracy. This paper focuses on instance-based approaches to ensemble pruning, where a different subset of the ensemble may be used for each different unclassified instance. We propose modeling this task as a multi-label learning problem, in order to take advantage of the recent advances in this area for the construction of effective ensemble pruning approaches. Results comparing the proposed framework against a variety of other instance-based ensemble pruning approaches in a variety of datasets using a heterogeneous ensemble of 200 classifiers, show that it leads to improved accuracy. markatopoulou-ictai10.pdf Arras, France Mulan: A Java Library for Multi-Label Learning 1 Journal of Machine Learning Research 12 2011 4 2411-2414 Mulan is a Java library for learning from multi-label data. It offers a variety of classiffication, ranking, thresholding and dimensionality reduction algorithms, including an algorithm for learning from hierarchically structured labels. In addition, it contains an evaluation framework that calculates a rich variety of performance measures. tsoumakas-jmlr11.pdf http%3A%2F%2Fjmlr%2Ecsail%2Emit%2Eedu%2Fpapers%2Fv12%2Ftsoumakas11a%2Ehtml The PORSCE II Framework: Using AI Planning for Automated Semantic Web Service Composition 1 The Knowledge Engineering Review Cambridge University Press Vol. 28, No. 2 2013 17 137-156 This paper presents PORSCE II, an integrated system that performs automatic semantic web service composition exploiting AI techniques, specifically planning. Essential steps in achieving web service composition include the translation of the web service composition problem into a solver-ready planning domain and problem, followed by the acquisition of solutions, and the translation of the solutions back to web service terms. The solutions to the problem, that is, the descriptions of the desired composite service, are obtained by means of external domain-independent planning systems, they are visualized and finally evaluated. Throughout the entire process, the system exploits semantic information extracted from the semantic descriptions of the available web services and the corresponding ontologies, in order to perform composition under semantic awareness and relaxation. ker.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1017%2FS0269888912000392 Translating Web Services Composition Plans to OWL-S Descriptions 2 Proc. 3rd International Conference on Agents and Artificial Intelligence (ICAART 2011) 2011 10 167-176 Web Services technology has led to simpler and more rapid development of Web Applications with improved functionality by which several platforms through the globe can communicate to exchange data and cooperate for problem solving. Methods for automated web services composition are studied so as to enhance this type of software development. Many studies focus on converting the composition problem to a planning problem and solving it using known planning algorithms. This paper suggests a method for translating the produced PDDL plans of the above algorithms to OWL-S descriptions of the final composite web services. The result is a totally new web service that can later be discovered and invoked or even take part in a new composition. icaart.pdf 28-30 Jan 2011, Rome, Italy web services composition AI planning semantic web services OWL-S PDDL MOpiS: a Multiple Opinion Summarizer 2 Proc. 5th Hellenic Conference on Artificial Intelligence, LNAI 5138 Springer-Verlag Berlin Heidelberg J. Darzentas et al. 2008 0 110-122 Product reviews written by on-line shoppers is a valuable source of information for potential new customers who desire to make an informed purchase decision. Manually processing quite a few dozens, or even hundreds, of reviews for a single product is tedious and time consuming. Although there exist mature and generic text summarization techniques, they are focused primarily on article type content and do not perform well on short and usually repetitive snippets of text found at on-line shops. In this paper, we propose MOpiS, a multiple opinion summarization algorithm that generates improved summaries of product reviews by taking into consideration metadata information that usually accompanies the on-line review text. We demonstrate the effectiveness of our approach with experimental results. kokkoras-MOpiS-SETN08-Springer.pdf F. Kokkoras, E. Lampridou, K. Ntonas, I. Vlahavas, "MOpiS: a Multiple Opinion Summarizer", in Proc. 5th Hellenic Conference on Artificial Intelligence, J. Darzentas et al. (Eds.) SETN 2008, Springer-Verlag Berlin Heidelberg, LNAI 5138, pp.110–122, Syros, October 2-4, 2008. http%3A%2F%2Fdeixto%2Ecom%2F DEiXTo Syros, Greece web mining opinion mining summarization ahp Summarization of Multiple, Metadata Rich, Product Reviews 2 Workshop on Mining Social Data (MSoDa), 18th European Conference on Artificial Intelligence (ECAI '08) 2008 0 Modern successful on-line shops and product comparison sites allow consumers to express their opinion on products and services they purchased. Although such information can be useful to other potential customers, reading and mentally processing quite a few dozens or even hundreds of reviews for a single product is tedious and time consuming. In this paper, we propose ReSum a novel summarization ap-proach for multiple, metadata augmented, product reviews. We argue that the contribution of additional information (metadata) such as the user's expertise, the usefulness of the review to other users, etc., is significant and can result in improved summaries. The summarization algorithm we propose outperforms two commercial, general purpose summarizers that ignore such metadata. kokkoras-MSoDa-ECAI08-camera.pdf F. Kokkoras, E. Lampridou, K. Ntonas, I. Vlahavas, "Summarization of Multiple, Metadata Rich, Product Reviews", in Proc. Workshop on Mining Social Data (MSoDa), 18th European Conference on Artificial Intelligence (ECAI '08), July 20, 2008. http%3A%2F%2Fdeixto%2Ecom%2F DEiXTo Patras, Greece web mining opinion mining summarization ahp PolyA-iEP: A Data Mining Method for the Effective Prediction of Polyadenylation Sites 1 Expert Systems with Applications Elsevier 38(10) 2011 0 1239812408 This paper presents a study on polyadenylation site prediction, which is a very important problem in bioinformatics and medicine, promising to give a lot of answers especially in cancer research. We describe a method, called PolyA-iEP, that we developed for predicting polyadenylation sites and we present a systematic study of the problem of recognizing mRNA 3΄ ends which contain a polyadenylation site using the proposed method. PolyA-iEP is a modular system consisting of two main components that both contribute substantially to the descriptive and predictive potential of the system. In specific, PolyA-iEP exploits the advantages of emerging patterns, namely high understandability and discriminating power and the strength of a distance-based scoring method that we propose. The extracted emerging patterns may span across many elements around the polyadenylation site and can provide novel and interesting biological insights. The outputs of these two components are finally combined by a classifier in a highly effective framework, which in our setup reaches 93.7% of sensitivity and 88.2% of specificity. PolyA-iEP can be parameterized and used for both descriptive and predictive analysis. We have experimented with Arabidopsis thaliana sequences for evaluating our method and we have drawn important conclusions. Tzanis_ESWA11.pdf data mining machine learning classification emerging pattern bioinformatics polyadenylation A Triple-Random Ensemble Classification Method for Mining Multi-label Data 2 Proc. 2010 IEEE International Conference on Data Mining Workshops 2010 8 49-56 This paper presents a triple-random ensemble learning method for handling multi-label classification problems. The proposed method integrates and develops the concepts of random subspace, bagging and random k-labelsets ensemble learning methods to form an approach to classify multi-label data. It applies the random subspace method to feature space, label space as well as instance space. The devised subsets selection procedure is executed iteratively. Each multi-label classifier is trained using the randomly selected subsets. At the end of the iteration, optimal parameters are selected and the ensemble MLC classifiers are constructed. The proposed method is implemented and its performance compared against that of popular multi-label classification methods. The experimental results reveal that the proposed method outperforms the examined counterparts in most occasions when tested on six small to larger multi-label datasets from different domains. This demonstrates that the developed method possesses general applicability for various multi-label classification problems. nasierding-icdm10.pdf Dealing with Concept Drift and Class Imbalance in Multi-label Stream Classification 2 Proc. 22nd International Conference on Artificial Intelligence (IJCAI 2011) AAAI press 2011 6 Streams of objects that are associated with one or more labels at the same time appear in many applications. However, stream classification of multi-label data is largely unexplored. Existing approaches try to tackle the problem by transferring traditional single-label stream classification practices to the multi-label domain. Nevertheless, they fail to consider some of the unique properties of the problem such as within and between class imbalance and multiple concept drift. To deal with these challenges, this paper proposes a novel multi-label stream classification approach that employs two windows for each label, one for positive and one for negative examples. Instance-sharing is exploited for space efficiency, while a time-efficient instantiation based on the k-Nearest Neighbor algorithm is also proposed. Finally, a batch-incremental thresholding technique is proposed to further deal with the class imbalance problem. Results of an empirical comparison against two other methods on three real world datasets are in favor of the proposed approach. spyromitrosijcai11.pdf Barcelona, Spain Multi-label Classification Concept Drift Class Imbalance Multiple Windows Multi-label Learning Approaches for Music Instrument Recognition 2 Proc. 9th International Symposium on Methodologies for Intelligent Systems (ISMIS 2011) 2011 10 This paper presents the two winning approaches that we developed for the instrument recognition track of the ISMIS 2011 contest on Music Information. The solution that ranked first was based on the Binary Relevance approach and built a separate model for each instrument on a selected subset of the available training data. Moreover, a new ranking approach was utilized to produce an ordering of the instruments according to their degree of relevance to a given track. The solution that ranked second was based on the idea of constraining the number of pairs that were being predicted. It applied a transformation to the original dataset and utilized a variety of post-processing filters based on domain knowledge and exploratory analysis of the evaluation set. Both solutions were developed using the Mulan open-source software for multi-label learning. spyromitrosIsmis.pdf http%3A%2F%2Ftunedit%2Eorg%2Fchallenge%2Fmusic%2Dretrieval%3Fm%3Dsummary ISMIS+2011+Contest%3A+Music+Information+Retrieval Warsaw, Poland Multi-label Learning Instrument Recognition Mulan: A Java Library for Multi-Label Learning 2 ICML 2010 Workshop on Machine Learning Open Source Software (MLOSS 2010), Demo/Poster Session 2010 0 spyromitrosMLOSS2010poster.pdf http%3A%2F%2Fmulan%2Esourceforge%2Enet%2F Mulan%27s+website Haifa, Israel A Combinatory Framework of Web 2.0 Mashup Tools, OWL-S and UDDI 1 Expert Systems with Applications Elsevier 38(6) 2011 0 6657-6668 The increasing number of Web 2.0 applications, such as wikis or social networking sites, indicates a movement to large-scale collaborative and social Web activities. Users can share information, add value to Web applications by using them or aggregate data from different sources creating Web applications (mashups) using specialized tools (mashup tools). However, Web 2.0 is not a new technology, but it rather embraces a new philosophy, treating the Internet as a platform. Several issues related to the Semantic Web vision, such as interoperability or machine understandable data semantics, are not tackled by Web 2.0. In this paper, we present our effort to combine semantic Web services (SWS) discovery frameworks, UDDI repositories and existing mashup tools in order to enhance the procedure of developing mashups with semantic mashup discovery capabilities. Towards this end, we introduce a social-oriented extension of OWL-S advertisements, their mapping algorithm on UDDI repositories and a semantic mashup discovery algorithm. Finally, we elaborate on the way our framework has been realized using the Yahoo Pipes mashup tool. http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2010%2E11%2E072 Mashup tools Semantic mashup discovery OWL-S OWL-S UDDI Visualizing Semantic Web Proofs of Defeasible Logic in the DR-DEVICE System 1 Knowledge-based Systems Elsevier 24(3) 2011 0 406-419 The Semantic Web aims at improving the current Web, by augmenting its content with semantics and encouraging the cooperation among human users and machines. Since the basic Semantic Web infrastructure is reaching sufficient maturity, research efforts are shifting towards logic, proof and trust and rule-based systems inevitably concentrate most of the attention. Nevertheless, in order for human users to trust system answers, they have to be presented with adequate explanations that justify the derived results. And, even more importantly, these explanations have to be presented in a user-comprehensible format. Consequently, the focus in this work is on humans and the research area called proof visualization that features three main approaches: tree-based, graphical and logical/textual. Since each of the approaches presents advantages and disadvantages, this article proposes a fourth, hybrid visualization approach that combines the pros of all three approaches and attempts to leverage the respective cons. The article also presents a software tool that implements the proposed hybrid approach. The tool is called VProofH and visualizes defeasible logic proofs, offering multiple representations that adapt to user needs. Extensive scalability and user evaluation tests prove the software tool’s usability. http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eknosys%2E2010%2E12%2E001 Semantic Web Defeasible logic Explanations Proof visualization Rule base SPARSE: A Symptom-based Antipattern Retrieval Knowledge-based System Using Semantic Web Technologies 1 Expert Systems with Applications Elsevier 38(6) 2011 0 7633-7646 Antipatterns provide information on commonly occurring solutions to problems that generate negative consequences. The number of software project management antipatterns that appears in the literature and the Web increases to the extent that makes using antipatterns problematic. Furthermore, antipatterns are usually inter-related and rarely appear in isolation. As a result, detecting which antipatterns exist in a software project is a challenging task which requires expert knowledge. This paper proposes SPARSE, an OWL ontology based knowledge-based system that aims to assist software project managers in the antipattern detection process. The antipattern ontology documents antipatterns and how they are related with other antipatterns through their causes, symptoms and consequences. The semantic relationships that derive from the antipattern definitions are determined using the Pellet DL reasoner and they are transformed into the COOL language of the CLIPS production rule engine. The purpose of this transformation is to create a compact representation of the antipattern knowledge, enabling a set of object-oriented CLIPS production rules to run and retrieve antipatterns relevant to some initial symptoms. SPARSE is exemplified through 31 OWL ontology antipattern instances of software development antipatterns that appear on the Web. http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2010%2E12%2E097 Antipatterns Symptom-based retrieval OWL ontology Production rules Objects CLIPS-OWL: A framework for providing object-oriented extensional ontology queries in a production rule engine 1 Data and Knowledge Engineering Elsevier 70(7) 2011 43 661-681 In this paper, we define a framework, namely CLIPS–OWL, for enabling the CLIPS production rule engine to represent the extensional results of DL reasoning on OWL ontologies in the form of Object-Oriented (OO) models. The purpose of this transformation is to allowCLIPS to use these OO models as static querymodels that are able to answer extensional ontology queries directly by the RETE reasoning engine during the development of custom CLIPS production rule programs, without interfacing at runtime the external DL reasoner. In that way, any CLIPS-based application may enhance its functionality by incorporating ontological knowledge without modifying the architecture of the CLIPS rule engine. CLIPS–OWL has been implemented using the Pellet DL reasoner and the CLIPS Object-Oriented Language (COOL). meditskos-datak-final.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Edatak%2E2011%2E04%2E001 production rules ontologies object-oriented model CLIPS OWL An Integrated Approach to Automated Semantic Web Service Composition through Planning 1 IEEE Transactions on Services Computing IEEE 5(3) 2012 0 The paper presents an integrated approach for automated semantic web service composition using AI planning techniques. An important advantage of this approach is that the composition process, as well as the discovery of the atomic services that take part in the composition, are significantly facilitated by the incorporation of semantic information. OWL-S web service descriptions are transformed into a planning problem described in a standardized fashion using PDDL, while semantic information is used for the enhancement of the composition process as well as for approximating the optimal composite service when exact solutions are not found. Solving, visualization, manipulation and evaluation of the produced composite services are accomplished, while, unlike other systems, independence from specific planners is maintained. Implementation was performed through the development and integration of two software systems, namely PORSCE II and VLEPPO. PORSCE II is responsible for the transformation process, semantic enhancement and management of the results. VLEPPO is a general-purpose planning system used to automatically acquire solutions for the problem by invoking external planners. A case study is also presented to demonstrate the functionality, performance and potential of the approach. tsc.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1109%2FTSC%2E2011%2E20 A Framework for Access Control with Inference Constraints 2 IEEE Computer Software and Applications Conference (COMPSAC 2011) 2011 14 The paper presents an integrated approach for automated semantic web service composition using AI planning techniques. An important advantage of this approach is that the composition process, as well as the discovery of the atomic services that take part in the composition, are significantly facilitated by the incorporation of semantic information. OWL-S web service descriptions are transformed into a planning problem described in a standardized fashion using PDDL, while semantic information is used for the enhancement of the composition process as well as for approximating the optimal composite service when exact solutions are not found. Solving, visualization, manipulation and evaluation of the produced composite services are accomplished, while, unlike other systems, independence from specific planners is maintained. Implementation was performed through the development and integration of two software systems, namely PORSCE II and VLEPPO. PORSCE II is responsible for the transformation process, semantic enhancement and management of the results. VLEPPO is a general-purpose planning system used to automatically acquire solutions for the problem by invoking external planners. A case study is also presented to demonstrate the functionality, performance and potential of the approach. compsac.pdf Munich, Germany Ontology-based Model Driven Engineering for Safety Verification 2 Proc. 36th EUROMICRO Conference on Software Engineering and Advanced Applications (SEAA 2010) IEEE Computer Society 2010 0 47-54 Safety assessment of dependable systems is a complex verification task that is desirable to be explicitly incorporated into the development cycle during the very early stages of a project. The main reason is that the cost to correct a safety error at the late stages of system development is excessively high. Towards this aim, we introduce an ontology-based model-driven engineering process for automating transformations of models that are utilized as reusable artifacts. The logical and syntactical structures of the design and safety models have to conform to a number of metamodel constraints. These constraints are semantically represented by mapping them onto an OWL domain ontology, allowing the incorporation of a Description Logic OWL reasoner and inference rules, in order to detect lacks of model elements and semantically inconsistent parts. Model validation throughout the ontology-based transformation assures that the generated formal safety model fulfils a series of requirements that render it analyzable. Our approach has been implemented as a response to an industrial problem, where the architecture design is expressed in Architecture Analysis and Design Language (AADL) and safety models are specified in the AltaRica formal language. Lille, France, 1-3 Sep 2010 http%3A%2F%2Fdoi%2Eieeecomputersociety%2Eorg%2F10%2E1109%2FSEAA%2E2010%2E60 model driven engineering safety verification and validation ontology reasoning model transformation Detecting antipatterns using a Web-based collaborative antipattern ontology knowledge base 2 ONTOSE 2011: 5th International Workshop on Ontology, Models, Conceptualization and Epistemology in Social, Artificial and Natural Systems, pre conference workshop of CAiSE 2011 Springer LNBIP 83 2011 0 478-488 The enrichment of the antipattern ontology that acts as the lexicon of terms to communicate antipatterns between people and software tools, is a labor intensive task. Existing work has implemented SPARSE, an ontology based intelligent system that uses a symptom based approach in order to semantically detect and retrieve inter-related antipatterns that exist in a software project. In this paper, we propose a Web-based environment that uses the Protege platform, in order to allow collaborative ontology editing as well as annotation and voting of both ontology components and ontology changes. This technology allows multiple users to edit and enrich the antipattern ontology simultaneously. Preliminary results on SPARSE show the effectiveness of the antipattern detection process during the research and development of a software project. June 20th 2011, London, United Kingdom http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%2F978%2D3%2D642%2D22056%2D2%5F50%3Fnull Collaborative Ontology Development Collaborative software Collaborative software engineering Antipatterns Special issue on Artificial Intelligence Techniques for Pervasive Computing: Preface 1 International Journal on Artificial Intelligence Tools World Scientific 19(2) 2010 0 133-135 This special issue focuses on how Artificial Intelligence techniques of various AI sub-areas, such as Knowledge Representation and Reasoning, Machine Learning, Machine Vision, Speech Recognition, Intelligent Human-Computer Interaction, Intelligent Agents, can contribute to the vision of pervasive computing to build electronic environments that are sensitive and responsive to the presence of people, by tackling highly interesting issues, such as applications providing personalized access and interactivity to multimodal information based on user preferences and semantic concepts or human-machine interface systems utilizing information on the affective state of the user. http%3A%2F%2Fwww%2Eworldscinet%2Ecom%2Fijait%2F19%2F1902%2FS02182130101902%2Ehtml Artificial Intelligence Pervasive Computing Ambient Intelligence Ubiquitous computing Guest editors' introduction: Rule representation, interchange, and reasoning in distributed, heterogeneous environments 1 IEEE Transactions on Knowledge and Data Engineering IEEE Computer Society 22(11) 2010 0 1489-1491 The eight papers in this special section focus on the state-of-the-art approaches, solutions, and applications in the area of rule representation, reasoning, and interchange in the context of distributed, (partially) open, heterogeneous environments, such as the semantic Web, intelligent multiagent systems, event-driven architectures. and service-oriented computing. http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1109%2FTKDE%2E2010%2E179 Knowledge based systems Semantic Web Rules Multiagent systems T-REX: A hybrid agent trust model based on witness reputation and personal experience 2 Proc. 11th International Conference on Electronic Commerce and Web Technologies (EC-Web 2010) Spinger 61 - Part 3 2010 12 107-118 Semantic Web will transform the way people satisfy their requests letting them delegate complex actions to intelligent agents, which will act on behalf of their users into real-life applications, under uncertain and risky situations. Thus, trust has already been recognized as a key issue in Multi-Agent Systems. Current computational trust models are usually built either on an agent’s direct experience or reports provided by others. In order to combine the advantages and overcome the drawbacks of these approaches, namely interaction trust and witness reputation, this paper presents a hybrid trust model that combines them in a dynamic and flexible manner. The main advantage of our approach is that it provides a reliable and flexible model with low bandwidth and storage cost. Moreover, we present the integration of this model in JADE, a multi-agent framework and provide an evaluation and an e-Commerce scenario that illustrate the usability of the proposed model. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald%2Femerald%2Ehtml EMERALD Bilbao, Spain http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fj37t22563341m342%2F Trust Reputation Multi-Agent Systems A Contract Agreement Policy-Based Workflow Methodology for Agents Interacting in the Semantic Web 2 Semantic Web Rules, Proc. 4th International Web Rule Symposium (RuleML 2010) Springer 6403 2010 15 225-239 The Semantic Web aims at automating Web content understanding and user request satisfaction. Intelligent agents assist towards this by performing complex actions on behalf of their users into real-life applications, such as e-Contracts, which make transactions simple by modeling the processes involved. This paper, presents a policy-based workflow methodology for efficient contract agreement among agents interacting in the Semantic Web. In addition, we present the integration of this methodology into a multi-agent knowledgebased framework, providing flexibility, reusability and interoperability of behavior between agents. The main advantage of our approach is that it provides a safe, generic, and reusable framework for modeling and monitoring e-Contract agreements, which could be used for different types of on-line transactions among agents. Furthermore, our framework is based on Semantic Web and FIPA standards, to maximize interoperability and reusability. Finally, an e-Commerce contract negotiation scenario is presented that illustrates the usability of the approach. Best Paper Award Washington, DC, USA http%3A%2F%2Fportal%2Eacm%2Eorg%2Fcitation%2Ecfm%3Fid%3D1929601 Semantic Web Intelligent Agents e-Contracts defeasible reasoning Cross-Community Interoperation Between the EMERALD and Rule Responder Multi-Agent Systems 2 Proc. 5th International Symposium on Rules: Research Based and Industry Focused (RuleML-2011@IJCAI) Springer Berlin / Heidelberg LNCS 6826 2011 8 44-51 The vision of the Semantic Web allows users to delegate complex actions to intelligent agents, which will act on behalf of their users in a variety of real-life applications. This paper focuses on two Semantic Web enabled multi-agent systems, EMERALD and Rule Responder, which can be employed to assist communities of users based on Semantic Web and multi-agent standards such as RDF, OWL, RuleML, and FIPA. The present work demonstrates how these multi-agent systems can interoperate to automate collaboration across communities using a declarative, knowledge-based approach. In addition, a multi-step interaction scenario among agents is presented, demonstrating the usefulness of interoperating between the above systems, exemplifying a general approach to cross-community collaboration. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2FemeraldRR%2Findex%2Ehtml EMERALD+%2D+Rule+Responder+gateway Barcelona, Spain http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%2F978%2D3%2D642%2D22546%2D8%5F5%3Fnull RuleML Semantic Web Web Rules intelligent multi-agent systems EMERALD Rule Responder Extending a Multi-agent Reasoning Interoperability Framework with Services for the Semantic Web Logic and Proof Layers 2 5th International Symposium on Rules: Research Based and Industry Focused (RuleML-2011@IJCAI) Springer Berlin / Heidelberg Volume 6826/2011 2011 0 29-43 The ultimate vision of the Semantic Web (SW) is to offer an interoperable and information-rich web environment that will allow users to safely delegate complex actions to intelligent agents. Much work has been done for agents’ interoperability; a plethora of proposals and standards for ontology-based metadata and rule-based reasoning are already widely used. Nevertheless, the SW proof layer has been neglected so far, although it is vital for SW agents and human users to understand how a result came about, in order to increase the trust in the interchanged information. This paper focuses on the implementation of third party SW reasoning and proofing services wrapped as agents in a multi-agent framework. This way, agents can exchange and justify their arguments without the need to conform to a common rule paradigm. Via external reasoning and proofing services, the receiving agent can grasp the semantics of the received rule set and check the validity of the inferred results. Selected among the best 2 papers of RuleML-2011@IJCAI to be presented at IJCAI-2011 http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald%2Femerald%2Ehtml EMERALD Barcelona, Spain http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fj1w61701wv8m7q08%2F intelligent multi-agent systems EMERALD DR-Prolog defeasible reasoning semantic web Advanced Agent Discovery Services 2 Proc. 2nd International Conference on Web Intelligence, Mining and Semantics (WIMS '12) ACM, New York, NY, USA Article 38 2012 12 338-349 The Semantic Web aims at augmenting the WWW with meaning, assisting people and machines in comprehending Web content and better satisfying their requests. Intelligent agents are considered to be greatly favored by Semantic Web technologies, because of the interoperability the latter will achieve. As a result, a plethora of multi-agent systems is already in use. One of the main problems in these systems is the lack of sufficient discovery and interaction monitoring tools. This paper reports on the design and implementation of advanced agent discovery services. These services are distinguished in two main categories; namely advanced semantic directory and discovery services (yellow pages) and trust-based discovery services which support among others interaction monitoring tools. Their implementation, including the associated graphical interface, was integrated into the EMERALD framework, a knowledge-based framework for interoperating agents. Finally, a software trade use case scenario is presented, which better displays the potential of the proposed approach. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald%2Femerald%2Ehtml EMERALD Craiova, Romania, June 13-15, 2012 http%3A%2F%2Fdl%2Eacm%2Eorg%2Fcitation%2Ecfm%3Fid%3D2254177 Semantic Web Intelligent Agents Discovery Services On the Stratification of Multi-Label Data 2 Proceedings of ECML PKDD 2011 2011 15 Stratified sampling is a sampling method that takes into account the existence of disjoint groups within a population and produces samples where the proportion of these groups is maintained. In single-label classification tasks, groups are differentiated based on the value of the target variable. In multi-label learning tasks, however, where there are multiple target variables, it is not clear how stratified sampling could/should be performed. This paper investigates stratification in the multi-label data context. It considers two stratification methods for multi-label data and empirically compares them along with random sampling on a number of datasets and based on a number of evaluation criteria. The results reveal some interesting conclusions with respect to the utility of each method for particular types of multi-label datasets. sechidis-ecmlpkdd-2011.pdf Athens, Greece Transferring Models in Hybrid Reinforcement Learning Agents 2 Proc. Engineering Applications of Neural Networks Springer Berlin Heidelberg 2011 12 162-171 The main objective of transfer learning is to reuse knowledge acquired in a previous learned task, in order to enhance the learning procedure in a new and more complex task. Transfer learning comprises a suitable solution for speeding up the learning procedure in Reinforcement Learning tasks. In this work, we propose a novel method for transferring models to a hybrid reinforcement learning agent. The models of the transition and reward functions of a source task, will be transferred to a relevant but different target task. The learning algorithm of the target task's agent takes a hybrid approach, implementing both model-free and model-based learning, in order to fully exploit the presence of a model. The empirical evaluation, of the proposed approach, demonstrated significant results and performance improvements in the 3D Mountain Car task, by successfully using the models generated from the standard 2D Mountain Car. EANN2011.pdf A System for Energy Savings in an Ambient Intelligence Environment 2 Proc. ICT-GLOW 2011 2011 0 This work presents an Ambient Intelligence system, that targets energy consumption awareness and savings. The system was deployed at the School of Science and Technology of the International Hellenic University. It was implemented following a three-layer approach. The first layer hosts devices (in our implementation smart plugs, sensor boards and smart clampers) suited for the purpose. The second layer, namely the aWESoME Web Service Middleware, resolves interoperability issues on the first layer, by universally exposing all actuator functions and sensor data through Web Services. Finally, a prototype application for monitoring and control, named iDEALISM, has been developed to reside on the topmost layer. iDEALISM has actually been tested and evaluated by students, for its usability and effectiveness in raising awareness. thanosICTGLOW11.pdf Toulouse, France A Hybrid Approach for Cold-start Recommendations of Videolectures 2 Proceedings ECML/PKDD 2011 Discovery Challenge Workshop 2011 12 This paper presents the solution which ranked 2nd in the "cold-start" recommendations task of the ECML/PKDD 2011 discovery challenge. The task was the recommendation of new videolectures to new users of the Videolectures.netWeb site. The proposed solution is a hybrid recommendation approach which combines content-based and collaborative information. Structured and unstructured textual attributes which describe each lecture are synthesized to create a vector representation with tf/idf weights. Collaborative information is incorporated for query expansion with a novel method which identies neighboring lectures in a co-viewing graph and uses them to supplement missing attributes. The cosine similarity measure is used to nd similar lectures and nal recommendations are made by also accounting the coexistence duration of lectures. The results of the competition show that the proposed approach is able to give accurate "cold-start" recommendations. spyromitrosECML2011.pdf The paper describes the solution that ranked 2nd in the ECML/PKDD 2011 Discovery Challenge http%3A%2F%2Ftunedit%2Eorg%2Fchallenge%2FVLNetChallenge%3Fm%3Dsummary Discovery+Challenge Athens, Greece hybrid recommender cold-start problem query expansion MLKD's Participation at the CLEF 2011 Photo Annotation and Concept-Based Retrieval Tasks 2 ImageClef Lab of CLEF 2011 Conference on Multilingual and Multimodal Information Access Evaluation 2011 15 We participated both in the photo annotation and concept-based retrieval tasks of CLEF 2011. For the annotation task we developed visual, textual and multi-modal approaches using multi-label learning algorithms from the Mulan open source library. For the visual model we employed the ColorDescriptor software to extract visual features from the images using 7 descriptors and 2 detectors. For each combination of descriptor and detector a multi-label model is built using the Binary Relevance approach coupled with Random Forests as the base classifier. For the textual models we used the boolean bag-of-words representation, and applied stemming, stop words removal, and feature selection using the chi-squared-max method. The multi-label learning algorithm that yielded the best results in this case was Ensemble of Classifier Chains using Random Forests as base classifier. Our multi-modal approach was based on a hierarchical late-fusion scheme. For the concept based retrieval task we developed two different approaches. The first one is based on the concept relevance scores produced by the system we developed for the annotation task. It is a manual approach, because for each topic we manually selected the relevant topics and manually set the strength of their contribution to the final ranking produced by a general formula that combines topic relevance scores. The second approach is based solely on the sample images provided for each query and is therefore fully automated. In this approach only the textual information was used in a query-by-example framework. spyromitrosCLEF2011.pdf (not peer-reviewed) http%3A%2F%2Fwww%2Eimageclef%2Eorg%2F2011%2FPhoto ImageClef+2011 Amsterdam, Netherlands image annotation image retrieval multi-label learning Transferring Evolved Reservoir Features in Reinforcement Learning Tasks 2 Recent Advances in Reinforcement Learning, Lecture Notes in Computer Science Springer-Verlag 7188 2012 12 213-224 The major goal of transfer learning is to transfer knowledge acquired on a source task in order to facilitate learning on another, dif- ferent, but usually related, target task. In this paper, we are using neu- roevolution to evolve echo state networks on the source task and transfer the best performing reservoirs to be used as initial population on the tar- get task. The idea is that any non-linear, temporal features, represented by the neurons of the reservoir and evolved on the source task, along with reservoir properties, will be a good starting point for a stochastic search on the target task. In a step towards full autonomy and by taking advan- tage of the random and fully connected nature of echo state networks, we examine a transfer method that renders any inter-task mappings of states and actions unnecessary.We tested our approach and that of inter- task mappings in two RL testbeds: the mountain car and the server job scheduling domains. Under various setups the results we obtained in both cases are promising. EVOLVED.pdf Transfer Learning in Multi-agent Reinforcement Learning Domains 2 Accepted for presentation at the 9th European Workshop on Reinforcement Learning and to be published in the Workshop Proceedings 2011 12 Transfer learning refers to the process of reusing knowledge from past tasks in order to speed up the learning procedure in new tasks. In reinforcement learning, where agents often require a consider- able amount of training, transfer learning comprises a suitable solution for speeding up learning. Transfer learning methods have primarily been applied in single-agent reinforcement learning algorithms, while no prior work has addressed this issue in the case of multi-agent learning. This work proposes a novel method for transfer learning in multi-agent rein- forcement learning domains. We test the proposed approach in a multi- agent domain under various setups. The results demonstrate that the method helps to reduce the learning time and increase the asymptotic performance. MULTIAGENT.pdf Transfer Learning via Multiple Inter-Task Mappings 2 Recent Advances in Reinforcement Learning Springer Berlin Heidelberg 2012 12 225-236 In this paper we investigate using multiple mappings for transfer learning in reinforcement learning tasks. We propose two dif- ferent transfer learning algorithms that are able to manipulate multiple inter-task mappings for both model-learning and model-free reinforce- ment learning algorithms. Both algorithms incorporate mechanisms to select the appropriate mappings, helping to avoid the phenomenon of negative transfer. The proposed algorithms are evaluated in the Moun- tain Car and Keepaway domains. Experimental results show that the use of multiple inter-task mappings can significantly boost the performance of transfer learning methodologies, relative to using a single mapping or learning without transfer. MAPPINGS.pdf A Survey of Service Composition in Ambient Intelligence Environments 1 Artificial Intelligence Review Springer 2011 0 This article presents a comparative review of systems performing service composition in Ambient Intelligence Environments. Such environments should comply to ubiquitous or pervasive computing guidelines by sensing the user needs or wishes and offering intuitive human-computer interaction and a comfortable non-intrusive experience. To achieve this goal service orientation is widely used and tightly linked with AmI systems. Some of these employ the Web Service technology, which involves well-defined web technologies and standards that facilitate interoperable machine to machine interaction. Other systems regard services of different technologies (e.g. UPnP, OSGi etc) or generally as abstractions of various actions. Service operations are sometimes implemented as software based functions or actions over hardware equipment (e.g. UPnP players). However, a single service satisfies an atomic only user need, so services need to be composed (i.e. combined), in order to provide the usually requested complex tasks. Since manual service composition is obviously a hassle for the user, ambient systems struggle to automate this process by applying various methods. The approaches that have been adopted during the last years vary widely in many aspects, like domain of application, modeling of services, composition method, knowledge representation and interfaces. This work presents a comparative view of these approaches revealing similarities and differences, while providing additional information. thanosAIReview.pdf Multi-label classification of music by emotion 1 EURASIP Journal on Audio, Speech, and Music Processing 2011:4 2011 20 This work studies the task of automatic emotion detection in music. Music may evoke more than one different emotion at the same time. Single-label classification and regression cannot model this multiplicity. Therefore, this work focuses on multi-label classification approaches, where a piece of music may simultaneously belong to more than one class. Seven algorithms are experimentally compared for this task. Furthermore, the predictive power of several audio features is evaluated using a new multi-label feature selection method. Experiments are conducted on a set of 593 songs with six clusters of emotions based on the Tellegen-Watson-Clark model of affect. Results show that multi-label modeling is successful and provide interesting insights into the predictive quality of the algorithms and features. http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1186%2F1687%2D4722%2D2011%2D426793 StackTIS: A Stacked Generalization Approach for Effective Prediction of Translation Initiation Sites 1 Computers in Biology and Medicine Elsevier Vol. 42, No. 1 2012 0 61-69 The prediction of the translation initiation site in an mRNA or cDNA sequence is an essential step in gene prediction and an open research problem in bioinformatics. Although recent approaches perform well, more effective and reliable methodologies are solicited. We developed an adaptable data mining method, called StackTIS, which is modular and consists of three prediction components that are combined into a meta-classification system, using stacked generalization, in a highly effective framework. We performed extensive experiments on sequences of two diverse eukaryotic organisms (Homo sapiens and Oryza sativa), indicating that StackTIS achieves statistically significant improvement in performance. http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Ecompbiomed%2E2011%2E10%2E009 Translation initiation Bioinformatics Data mining Machine learning Classification Stacking A Study on Greedy Algorithms for Ensemble Pruning 5 … 2012 0 Ensemble selection deals with the reduction of an ensemble of predictive models in order to improve its efficiency and predictive performance. A number of ensemble selection methods that are based on greedy search of the space of all possible ensemble subsets have recently been proposed. They use different directions for searching this space and different mea- sures for evaluating the available actions at each state. Some use the training set for subset evaluation, while others a separate validation set. This paper abstracts the key points of these methods and offers a general framework of the greedy ensemble selection algorithm, discussing its important parameters and the different options for instantiating these parameters. partalas-tr-2012.pdf BOnSAI: a Smart Building Ontology for Ambient Intelligence 2 Proc. 2nd International Conference on Web Intelligence, Mining and Semantics (WIMS-2012) ACM 2012 11 263-274 This work introduces an ontology for incorporating Ambient Intelligence in Smart Buildings. The ontology extends and benefits from existing ontologies in the field, but also adds classes needed to sufficiently model every aspect of a service-oriented smart building system. Namely, it includes concepts modeling all functionality (i.e. services, operations, inputs, outputs, logic, parameters and environmental conditions), QoS (resources, QoS parameters), hardware (smart devices, sensors and actuators, appliances, servers) users and context (user profiles, moods, location, rooms etc.). The ontology is instantiated and put to use at the Smart Building setting of the International Hellenic University, enabling knowledge representation in machineinterpretable form and hence is expected to enhance service-based intelligent applications. thanosWIMS2012.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsmartihu%2F Smart+IHU Craiova, Romania, June 13-15, 2012 http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1145%2F2254129%2E2254166 Ambient Intelligence Semantic Web Ontologies Semantically-enhanced Authoring of Defeasible Logic Rule Bases in the Semantic Web 2 Proc. 2nd International Conference on Web Intelligence, Mining and Semantics (WIMS'12) ACM Article 56 2012 4 489-492 The Semantic Web represents an initiative to improve the current Web, by augmenting content with semantics and encouraging cooperation among human and software agents. The development of the logic and proof layers of the Semantic Web is currently concentrating the related research effort and is vital, since these layers allow systems to infer new knowledge from existing information, assisting them in explaining their actions and, ultimately, increasing user trust towards the Semantic Web. However, there is a lack of applications that could contribute towards developing logic-based applications. Consequently, users resort to inadequate tools that offer syntactic support, without being able to support the user semantically as well. This work presents S2DRREd, a software tool that introduces a supplementary level of semantic assistance during rule base development. The tool allows creating meta-models of the main notions of the loaded rule sets and assists the user in authoring rule bases, independently of the explicitly chosen rule language syntax. The domain of application is defeasible logic, a type of logic that allows reasoning with incomplete and conflicting information and, as such, it can play an increasingly important role in a drastically dynamic environment like the Web. WIMS'12-Kontopoulos-Zetta-Bassiliades.pdf http%3A%2F%2Fsourceforge%2Enet%2Fprojects%2Fs2red%2F S2Red Craiova, Romania, June 13-15, 2012 http%3A%2F%2Fdl%2Eacm%2Eorg%2Fcitation%2Ecfm%3Fid%3D2254129%2E2254199 Semantic Web RuleML Rule Bases Defeasible Logic Non-monotonic Reasoning A Modal Defeasible Reasoner of Deontic Logic for the Semantic Web 1 International Journal on Semantic Web and Information Systems (IJSWIS) IGI Global 7(1) 2011 26 18-43 Defeasible logic is a non-monotonic formalism that deals with incomplete and conflicting information, whereas modal logic deals with the concepts of necessity and possibility. These types of logics play a significant role in the emerging Semantic Web, which enriches the available Web information with meaning, leading to better cooperation between end-users and applications. Defeasible and modal logics, in general, and, particularly, deontic logic provide means for modeling agent communities, where each agent is characterized by its cognitive profile and normative system, as well as policies, which define privacy requirements, access permissions, and individual rights. Toward this direction, this article discusses the extension of DR-DEVICE, a Semantic Web-aware defeasible reasoner, with a mechanism for expressing modal logic operators, while testing the implementation via deontic logic operators, concerned with obligations, permissions, and related concepts. The motivation behind this work is to develop a practical defeasible reasoner for the Semantic Web that takes advantage of the expressive power offered by modal logics, accompanied by the flexibility to define diverse agent behaviours. A further incentive is to study the various motivational notions of deontic logic and discuss the cognitive state of agents, as well as the interactions among them. IJSWIS-2011-Kontopoulos-etal.pdf http%3A%2F%2Fwww%2Eigi%2Dglobal%2Ecom%2Farticle%2Fmodal%2Ddefeasible%2Dreasoner%2Ddeontic%2Dlogic%2F55390 Semantic Web defeasible logic modal logic deontic logic reasoning reasoning engine DR-DEVICE An empirical study on the combination of surf features with VLAD vectors for image search 2 13th International Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS) IEEE 2012 4 The study of efficient image representations has attracted significant interest due to the computational needs of large-scale applications. In this paper we study the performance of the recently proposed VLAD method for aggregating local image descriptors when combined with SURF features, in the domain of image search. The experiments show that when SURF features are used as local image descriptors, VLAD attains better performance compared to using SIFT features. We also study how the average number of local image descriptors extracted per image affects the performance and show that by controlling this number we are able to adjust the trade off between feature extraction time and search accuracy. Finally, we examine the retrieval performance of the proposed scheme with varying levels of distractor images. spyromitrosWIAMIS2012.pdf http%3A%2F%2Fieeexplore%2Eieee%2Eorg%2Fxpl%2Ffreeabs%5Fall%2Ejsp%3Farnumber%3D6226771 VLAD SURF Image representation Feature extraction Principal Component Analysis RuleML Representation and Simulation of Fuzzy Cognitive Maps 1 Expert Systems with Applications Elsevier 40 (5) 2013 23 1413-1426 Fuzzy Cognitive Map (FCM) technique is a combination of Fuzzy Logic and Artificial Neural Networks that is extensively used by experts and scientists of a diversity of disciplines, for strategic planning, decision making and predictions. A standardized representation of FCMs accompanied by a system that would assist decision makers to simulate their own developed Fuzzy Cognitive Maps would be highly appreciated by them, and would help the dissemination of FCMs. In this paper, a) a RuleML representation of FCM is proposed and b) a system is designed and implemented in Prolog programming language to assist experts to simulate their own FCMs. This system returns results in valid RuleML syntax, making them readily available to other cooperative systems. The representation capabilities and the design choices of the implemented system are discussed and a variety of examples are given to demonstrate the use of the system. FCM-RuleML.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2012%2E08%2E035 Fuzzy Cognitive Maps RuleML Simulation Decision Making Predictions Prolog Bridging the HASM: An OWL ontology for modelling the information pathways in haptic interfaces software 1 Expert Systems with Applications Elsevier Vol. 40, No. 4 2012 34 1358-1371 Haptics technology has received enormous attention to enhance human computer interaction. The last decade has witnessed a rapid progress in haptic application software development due to the fact that the underlying technology has become mature and has opened up novel research areas. In an attempt to organize the path between cause and effect we envision a need for a standard for haptic application softwaremodeling. In order for the software to better enhance the tactile information sensation, flow and perception and also make interaction between humans and haptics more efficient and natural, we need a formal representation of the haptics domain. This article proposes the use of HASM, a haptic applications softwaremodelingontology to formally model the haptics domain in order to be used during the specifications and design phases of developing software applications for hapticinterfaces. The presented ontology captures the existing knowledge in haptics domain, using OWL, and defines the pathways that the hapticinformation follows between the human and the machine haptic system, using SWRL rules. The hapticontology that has been developed will be used as a basis to design effective user interfaces and assist the development of softwaremodeling for haptic devices. A case study is demonstrating how this hapticontology can be used to design a software model that analyzes the perception of a haptic property of an object by interacting with a haptic device. myrgioti_etal.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fontologies%2Fontolist%2Ehtml%23hasm HASM+ontology http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2012%2E08%2E053 Haptic devices Tactile displays Tactile information Haptic-Tactile interfaces Human-computer interaction Ontology Special issue on Intelligent Distributed Systems: Preface 1 International Journal on Artificial Intelligence Tools World Scientific 20(6) 2011 0 981-984 This special issue focuses on how Artificial Intelligence techniques of various AI sub-areas, such as Intelligent Agents and Multi-agent Systems, Knowledge Representation and Reasoning, Semantic Web and Ontologies, Machine Learning can contribute to intelligent distributed computing in order to build distributed systems that are able to communicate and coordinate their actions to exhibit intelligent and adaptive behavior. The special issue brings to the reader new results, applications and tools of intelligent distributed systems, with a special focus on synergies between agents, services and processes on one side and semantic technologies including ontologies and rules on the other side. In particular, some of the papers address interesting applications of intelligent distributed approaches in the areas of ambient intelligence, human collaboration and e-business. http%3A%2F%2Fwww%2Eworldscientific%2Ecom%2Ftoc%2Fijait%2F20%2F06 distributed computing intelligent systems PLIS+: A Rule-Based Personalized Location Information System 2 Proc. of the RuleML2012@ECAI Challenge, at the 6th International Symposium on Rules CEUR Workshop Proceedings Hassan Ait-Kaci, Yuh-Jong Hu, Grzegorz J. Nalepa, Monica Palmirani, Dumitru Roman Vol-874 2012 0 In this paper, the idea of providing personalized, location-based information services via rule-based policies is demonstrated. After a short introduction, an innovative Personalized Location Information System (PLIS+) is designed and implemented. PLIS+ delivers personalized and contextualized information to users according to rule-based policies. More specifically, many categories of points of interest (for example shops, restaurants) have rule-based policies to expose and deploy their marketing strategy on special offers, discounts, etc. PLIS+ evaluates these rules on-the-fly and delivers personalized information according to the user’s context and the corresponding rules fired within this context. After discussing the design and the implementation of PLIS+, illustrative examples of PLIS+ functionality are presented. As a result, PLIS+ proves that combining contextual data and rules can lead to powerful personalized information services. Best Challenge Award http%3A%2F%2Fplaton%2Eecon%2Eauth%2Egr%2Fexamples%2Fplis%2Flogin%2Ejsp PLIS Montpellier, France, August 27-31, 2012 http%3A%2F%2Fceur%2Dws%2Eorg%2FVol%2D874%2F RuleML Rules Location Based Services Context Points of Interest Jess Personalizing Location Information through Rule-Based Policies 2 Rules on the Web: Research and Applications, Proc. 6th International Symposium on Rules: Research Based and Industry Focused (RuleML 2012) Springer A. Bikakis and A. Giurca LNCS 7438 2012 0 215-223 In this paper, the idea of providing personalized, location-based information services via rule-based policies is demonstrated. After a short introduction about related technologies and approaches, an innovative Personalized Location Information System (PLIS) is designed and implemented. PLIS delivers personalized and contextualized information to users according to rule-based policies. More specifically, many categories of points of interest (e.g. shops, restaurants) have rule-based policies to expose and deploy their marketing strategy on special offers, discounts, etc. PLIS evaluates these rules on-the-fly and delivers personalized information according to the user’s context and the corresponding rules fired within this context. After discussing the design and the implementation of PLIS, illustrative examples of PLIS functionality are presented. As a result, PLIS proves that combining contextual data and rules can lead to powerful personalized information services. http%3A%2F%2Fplaton%2Eecon%2Eauth%2Egr%2Fexamples%2Fplis%2Flogin%2Ejsp PLIS Montpellier, France, August 27-31, 2012 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fju42276636227n82%2F RuleML Rules Location Based Services Context Points of Interest Jess IRISPortal: A Semantic Portal for Industrial Risk Cases Management 2 Proc. 2nd International Conference on Web Intelligence, Mining and Semantics (WIMS’ 12) ACM Dumitru Dan Burdescu, Rajendra Akerkar, Costin Badica 2012 0 243-250 In this paper, we describe the architecture and functionality of IRISPortal, a semantic portal that allows the management of industrial risk cases by exploiting state-of-the-art semantic technologies, such as the OWL 2 language and the OWLIM semantic repository. The portal allows the web-based management of risk cases that are modeled in terms of a risk ontology, assisting the domain experts to perform administrative tasks, such as adding, deleting and updating risk cases. Furthermore, the portal provides the functionality to the end-users for searching and browsing the modeled risk cases and their corresponding characteristics, based on the semantic relationships that derive from the ontology model after the reasoning procedure. http%3A%2F%2Firisportal%2Ecsd%2Eauth%2Egr%2F IRISPortal Craiova, Romania, June 13-15, 2012 http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1145%2F2254129%2E2254164 semantic web OWL rule-based reasoning web application risk management Agents and Knowledge Interoperability in the Semantic Web Era 2 Proc. 2nd International Conference on Web Intelligence, Mining and Semantics (WIMS’ 12) ACM Dumitru Dan Burdescu, Rajendra Akerkar, Costin Badica 2012 0 46-58 This tutorial will discuss about issues, technologies and tools that concern the way that the Semantic Web affects knowledge and information interchange among intelligent agents in multi-agent systems, as well as reasoning interoperability. First, the tutorial will discuss how semantic web rules and ontologies interact with each other in order to be used as the agent's internal knowledge base for environment awareness and decision making. Then, interoperability between reasoning systems for agents will be discussed. The issues involved in all the previous discussion will be exemplified using actual implemented tools for semantic web reasoning in multi-agents systems. Invited tutorial http%3A%2F%2Fsoftware%2Eucv%2Ero%2FWims12%2Fslides%2FNick%2FWIMS12%2Fbassiliades%2Dtutorial%2Dwims12%2Epptx Tutorial+slides Craiova, Romania, June 13-15, 2012 http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1145%2F2254129%2E2254140 Multiagent systems Semantic Web Rules Ontologies Reasoning Engines Knowledge Interoperability SWRL2COOL: Object-Oriented Transformation of SWRL in the CLIPS Production Rule Engine 2 Proc. 7th Hellenic Conference on Artificial Intelligence (SETN 2012) Springer Ilias Maglogiannis , Vassilis Plagianakos and Ioannis Vlahavas LNAI 7297 2012 0 49-56 The Semantic Web Rule Language (SWRL) is a W3C member submission rule language for ontologies. It is based on a combination of the OWL DL and OWL Lite sublanguages of the OWL Web Ontology Language with the Unary/Binary Datalog RuleML sublanguages of the Rule Markup Language. In this paper we propose a transformation of SWRL rules into the object-oriented rule language of CLIPS (COOL). The purpose of this transformation is to enhance an already existing CLIPS-based OWL ontology reasoner, namely O-DEVICE, with the ability to import and execute SWRL rules during the process of building custom ontology-based production rule programs. Lamia, Greece, May 28-31, 2012 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fj361x54378677865%2F SWRL Production Rules CLIPS OWL Paving the way for a Transformational Public Administration global security, safety and sustainability and e-Democracy 2 Global Security, Safety And Sustainability and E-Democracy, Proc. 17th ICGS3 / 4th e-Democracy Joint Conferences 2011 Springer, Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering C. K. Georgiadis, H. Jahankhani, E. Pimenidis, R. Bashroush, and A. Al-Nemrat 99 2012 0 194-203 Transformational government as a newborn scientific field seeks for implementation through integration of its components. As a contribution to this end this work impresses a Public Administration’s operation ontology modeling and an algorithm for tracing malfunctions and changing the case. PA is considered as a production unit and any administrative act as the output of its processes. This output creates effects and consequences which are to be met stakeholders’ goals in order to balance socioeconomic problems. Thessaloniki, Greece, 24-26 Aug, 2011 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fwj41561411277v38%2F Transformational government PA ontology modeling Service transformation algorithm stakeholder goals Monitoring Conformance to the Internal Regulation of an MSc Course using Ontologies and Rules 2 Proc. 2nd International Conference on Electronic Government and the Information Systems Perspective (EGOVIS'11) Springer Kim Normann Andersen, Enrico Francesconi, Ake Gronlund and Tom M. van Engers LNCS 6866 2011 0 212-226 The representation of information in the Web today is mainly through natural language and its meaning is only interpretable to users who have studied understand the specific natural language. Thus, in the case of the Internal Regulation (IR) of an MSc course of a Greek State University in order to extract an (indispensable) conclusion, one must understand the Greek language, must comprehend the content of the regulation and finally must combine information maybe from many disparate parts of the corpus. For example, if a candidate post-graduate student wanted to know if and how he can attend these courses he should consider all the articles of the IR to find the answer to this question. On the other hand, a computer program could not draw such a conclusion using natural language text. To solve problems of this nature one can use the technologies of the Semantic Web. This paper presents the development of a system that gives solution to these issues, based on Semantic Wed mechanisms, such as ontologies in OWL and rule in SWRL. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fontologies%2Fontolist%2Ehtml%23iro Internal+Regulation+Ontology Toulouse, France, Aug 29-Sep 2, 2011 http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Ff70352w51u254813%2F Course Regulations Semantic Web Ontologies OWL Rules SWRL An Ontological Business Process Modeling Approach for Public Administration: The Case of Human Resource Management 3 Handbook of Research on E-Business Standards and Protocols: Documents, Data and Advanced Web Technologies IGI Global Vol. II, Ch. 33 2012 0 725-753 In this chapter, an electronic model of Public Administration’s operation using an ontology as a means to a formalized representation of knowledge is presented. According to the proposed model, every public administration procedure is viewed as a service offered to some external entity and is represented as a (Semantic) Web service, semantically annotating its functional parameters, profile, and workflow. The modeling of public administration services/procedures involved the commonly used IOPE (Inputs – Outputs – Preconditions – Effects) model of OWL-S for Semantic Web Service description. This chapter also presents a specific use case about the Human Resource Management department of the Region of Central Macedonia. In order to do so, certain extensions/adaptations of the general methodology were needed. In this chapter the authors fully present and justify these adaptations that were deployed in order to turn the general methodology into a really flexible and re-usable tool to model any public administration procedure. Furthermore, the authors describe the full knowledge engineering cycle for developing the ontology of this department’s business processes. http%3A%2F%2Fwww%2Eigi%2Dglobal%2Ecom%2Fchapter%2Fontological%2Dbusiness%2Dprocess%2Dmodeling%2Dapproach%2F63495 HARM: A Hybrid Rule-based Agent Reputation Model based on Temporal Defeasible Logic 2 6th International Symposium on Rules: Research Based and Industry Focused Springer Berlin / Heidelberg Antonis Bikakis, Adrian Giurca 7438 2012 0 193 - 207 Multi-agent systems are considered a modern medium of communication and interaction with limited or no human intervention. As intelligent agents are gradually enriched with Semantic Web technology, their use is constantly increasing. To this end, the degree of trust that can be invested in a certain agent is recognized as a vital issue. Current trust models are mainly based on agents’ direct experience (interaction trust) or reports provided by others (witness reputation). Though, lately, some combinations of them (hybrid models) were also proposed. To overcome their main drawbacks, in this paper we propose HARM, a hybrid, rule-based reputation model based on temporal defeasible logic. It combines the advantages of the hybrid approach and the benefits of a rule-based reputation modeling approach, providing a stable and realistic estimation mechanism with low bandwidth and computational complexity. Moreover, an evaluation of the reputation model is presented, demonstrating the added value of the approach. Montpellier, France http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Fbxw20125273146l4%2F Temporal Defeasible Logic Semantic Web Intelligent Multi-agent Systems Defeasible Reasoning Agent Reputation Cross-Community Interoperation Between Knowledge-Based Multi-Agent Systems: A Study on EMERALD and Rule Responder 1 Journal of Expert Systems With Applications Elsevier 39 (10) 2012 0 9571-9587 The ultimate vision of the Semantic Web (SW) is to provide users with the capability of delegating complex tasks to intelligent agents. The latter, acting in an interoperable and information-rich Web environment, will efficiently satisfy their users’ requests in a variety of real-life applications. Much work has been done on SW information agents for Web-based query answering; a variety of multi-agent platforms and Web language standards has been proposed. However, the platform- and language-bridging interoperation across multi-agent systems has been neglected so far, although it will be vital for large-scale agent deployment and wide-spread adoption of agent technology by human users. This article defines the space of possible interoperation methods for heterogeneous multi-agent systems based on the communication type, namely symmetric or asymmetric, and the MASs status, namely open or closed systems. It presents how heterogeneous multi-agent systems can use one of these methods to interoperate and, eventually, automate collaboration across communities. The method is exemplified with two SW-enabled multi-agent systems, EMERALD and RuleResponder, which assist communities of users based on declarative SW and multi-agent standards such as RDF, OWL, RuleML, and FIPA. This interoperation employs a declarative, knowledge-based approach, which enables information agents to make smart and consistent decisions, relying on high-quality facts and rules. Multi-step interaction use cases between agents from both communities are presented, demonstrating the added value of interoperation. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2FemeraldRR%2Findex%2Ehtml EMERALD+%2D+Rule+Responder+gateway http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%2Farticle%2Fpii%2FS095741741200423X Semantic Web RuleResponder Intelligent multi-agent systems EMERALD Cross-Community Collaboration Rule-Based Reasoning, Programming, and Applications 4 Proceedings of 5th International Symposium, RuleML 2011 – Europe Springer LNCS, Vol. 6826 2011 383 This book constitutes the refereed proceedings of the 5th International Symposium on Rules: Research Based and Industry Focused (RuleML-2011@IJCAI), collocated in Barcelona, Spain, with the 22nd International Joint Conference on Artificial Intelligence. The technical program showed a carefully selected presentation of current rule research and development in 18 full papers, 8 short papers, 3 invited track papers, and 2 keynote talks (abstracts included) detailed in this book. Accepted papers covered several aspects of rules, such as rule-based distributed/multiagent systems, rules, agents and norms, rule-based event processing and reaction rules, fuzzy rules and uncertainty, rules and the Semantic Web, rule learning and extraction, rules and reasoning, and finally, rule-based applications. The papers were selected from 58 submissions received from 22 countries. Four of the submissions were selected from some worthy papers on topics related to RuleML originally submitted at IJCAI that marginally missed the acceptance cut-off. In order to increase the quality of the papers, this year a two-round reviewing scheme was adopted to assess the revised papers from IJCAI and revised papers originally submitted to RuleML but not immediately accepted in the first reviewing round. http%3A%2F%2Fwww%2Edefeasible%2Eorg%2Fruleml2011ijcai%2F RuleML%2D2011%40IJCAI http%3A%2F%2Fwww%2Espringerlink%2Ecom%2Fcontent%2Ft337435l6505%2F%23section%3D924901%26page%3D1 Rules and Automated Reasoning Logic Programming and Non-monotonic Reasoning Rules, Agents and Norms Rule-Based Distributed/Multi-Agent Systems Rule-Based Policies, Reputation and Trust Rule-based Event Processing and Reaction Rules Fuzzy Rules and Uncertainty Rule Transformation and Extraction Vocabularies, Ontologies, and Business rules Rule interchange and reasoning interoperation Reasoning and Proofing Services for Semantic Web Agents 2 In proceedings of the 22nd International Join Conference on Artificial Intelligence (IJCAI-2011) AAAI Press 3 2011 0 2662-2667 The Semantic Web aims to offer an interoperable environment that will allow users to safely delegate complex actions to intelligent agents. Much work has been done for agents' interoperability; especially in the areas of ontology-based metadata and rule-based reasoning. Nevertheless, the SW proof layer has been neglected so far, although it is vital for agents and humans to understand how a result came about, in order to increase the trust in the interchanged information. This paper focuses on the implementation of third party SW reasoning and proofing services wrapped as agents in a multi-agent framework. This way, agents can exchange and justify their arguments without the need to conform to a common rule paradigm. Via external reasoning and proofing services, the receiving agent can grasp the semantics of the received rule set and check the validity of the inferred results. http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald%2Fresource%2Ehtml EMERALD Barcelona, Spain http%3A%2F%2Fijcai%2Eorg%2Fpapers11%2FPapers%2FIJCAI11%2D443%2Epdf Semantic Web Intelligent Agents Reasoning Services Proofing Services Mining Frequent Patterns and Association Rules from Biological Data 3 Biological Knowledge Discovery Handbook: Preprocessing, Mining and Postprocessing of Biological Data Wiley-Blackwell (John Wiley & Sons) Mourad Elloumi and Albert Y. Zomaya 2014 0 This chapter presents a method called PolyA-iEP that has been developed for the prediction of polyadenylation sites. More precisely PolyA-iEP is a method that recognizes mRNA 3’ends which contain polyadenylation sites. It is a modular system which consists of two main components. The first exploits the advantages of emerging patters and the second is a distance-based scoring method. The outputs of the two components are finally combined by a classifier. The final results reach very high scores of sensitivity and specificity. (in press) Data Mining Machine Learning Classification Emerging Patterns Bioinformatics Polyadenylation Transferring Task Models in Reinforcement Learning Agents 1 Neurocomputing Elsevier 107 2013 21 23-32 The main objective of Transfer Learning is to reuse knowledge acquired in a previous learned task, in order to enhance the learning procedure in a new and more complex task. Transfer learning comprises a suitable solution for speeding up the learning procedure in Reinforcement Learning tasks. In this work, we propose a novel method for transferring models to Reinforcement Learning agents. The models of the transition and reward functions of a source task, will be transferred to a relevant but different target task. The learning algorithm of the target task's agent takes a hybrid approach, implementing both model-free and model-based learning, in order to fully exploit the presence of a source task model. Moreover, a novel method is proposed for transferring models of potential-based, reward shaping functions. The empirical evaluation, of the proposed approaches, demonstrated significant results and performance improvements in the 3D Mountain Car and Server Job Scheduling tasks, by successfully using the models generated from their corresponding source tasks. TaskModels.pdf (in press) Virtual Laboratories on Wireless Communications: A contemporary, extensible approach 2 IEEE EDUCON IEEE 2012 0 EDUCON_Publish_VLABS.pdf Morocco Development of a Logic Programming Machine 5 PhD Thesis (in Greek) University of Thessaloniki 1987 0 The Programming Language Pascal 4 ISBN 960-7013-17-4 Gartaganis Publications 1991 0 Methods and Applications of Artificial Intelligence 4 LNAI Springer - Verlag 2308 2002 0 aWESoME: a Web Service Middleware for Ambient Intelligence 1 Expert Systems With Applications Elsevier 40 (11) 2013 13 4380-4392 This work presents a Web Service Middleware infrastructure for Ambient Intelligence environments, named aWESoME. aWESoME is a vital part of the Smart IHU project, a large-scale Smart University deployment. The purpose of the proposed middleware within the project is twofold: for one, to ensure universal, homogeneous access to the system’s functions and secondly, to fulfill functional and non-functional requirements of the system. Namely, the infrastructure itself should consume significantly low power (as it is meant for energy savings in addition to automations), without compromising reliability and fast response time. The infrastructure should enable fast and direct discovery, invocation and execution of services. Finally, on hardware level, the wireless sensor and actuator network should be optimally configured for speed and reliability as well. The proposed solution employs widely used web open standards for description and discovery to expose hardware and software functions and ensure interoperability, even outside the borders of this university deployment. It proposes a straightforward method to integrate low-cost and resource-constrained heterogeneous devices found in the market and a large-scale placement of servers and wireless sensor networks. Different server hardware installations have been evaluated to find the optimum trade-off between response time and power consumption. Finally, a range of client applications that exploit the middleware on different platforms are demonstrated, to prove its usability and effectiveness in enabling, in this scenario, energy monitoring and savings. thanosESWA2013.pdf http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%2Farticle%2Fpii%2FS0957417413000936 Web Services Real-time and Embedded Systems Ubiquitous Computing Wireless Sensor Networks Iridescent: a Tool for Rapid Semantic Annotation of Web Service Descriptions 2 Proc. 3nd International Conference on Web Intelligence, Mining and Semantics (WIMS’ 13) ACM 2013 9 12:1-12:9 Although the Semantic Web and Web Service technologies have al-ready formed a synergy towards Semantic Web Services, their use remains limited. Potential adopters are usually discouraged by the number of different methodologies and the lack of tools, which both force them to acquire expert knowledge and commit to exhausting manual labor. This work proposes a novel functional and user-friendly graphical tool, named Iridescent, intended for both expert and non-expert users, to create and edit Semantic Web Service descriptions, following the SAWSDL recommendation. The tool’s aim is twofold: to enable users manually create descriptions in a visual manner, providing a complete alternative to coding, and to semi-automate the process by matching elements and concepts and suggesting annotations. A state-of-the-art survey has been carried out to reveal critical points and requirements. The tool’s functionality is presented along with usage scenarios that demonstrate how the tool and SAWSDL enable Intelligence in an Ambient Intelligence environment. Finally, Iridescent was methodically tested for its usability and evalu-ated by a range of both expert and non-expert users. thanosWIMS13.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Fthanosgstavr%2Fapplications%2Firidescent%2Ehtml http%3A%2F%2Fdoi%2Eacm%2Eorg%2F10%2E1145%2F2479787%2E2479797 Ontology-based Sentiment Analysis of Twitter Posts 1 Expert Systems with Applications Elsevier 40 (10) 2013 0 4065-4074 The emergence of Web 2.0 has drastically altered the way users perceive the Internet, by improving information sharing, collaboration and interoperability. Micro-blogging is one of the most popular Web 2.0 applications and related services, like Twitter, have evolved into a practical means for sharing opinions on almost all aspects of everyday life. Consequently, micro-blogging web sites have since become rich data sources for opinion mining and sentiment analysis. Towards this direction, text-based sentiment classifiers often prove inefficient, since tweets typically do not consist of representative and syntactically consistent words, due to the imposed character limit. This paper proposes the deployment of original ontology-based techniques towards a more efficient sentiment analysis of Twitter posts. The novelty of the proposed approach is that posts are not simply characterized by a sentiment score, as is the case with machine learning-based classifiers, but instead receive a sentiment grade for each distinct notion in the post. Overall, our proposed architecture results in a more detailed analysis of post opinions regarding a specific topic. http%3A%2F%2Fwww%2Esciencedirect%2Ecom%2Fscience%2Farticle%2Fpii%2FS0957417413000043 Micro-blogging Twitter Tweet Sentiment Analysis Ontology Web Service Composition Plans in OWL-S 3 Agents and Artificial Intelligence, Communications in Computer and Information Science (CCIS) Springer Berlin Heidelberg J. Filipe and A. Fred Vol. 271 2013 0 240-254 One of the main visions of Semantic Web has been the ability of software agents to compose atomic web services in order to facilitate the automation of complex tasks. One of the approaches used in the past in order to automatically construct composite web services has been AI planning. The most important advantage of this approach is its dynamic character that reduces the interference of the user. Although there have been various attempts to utilize planning algorithms and systems in the composition process, there has been little work in the field of converting web service composition plans in OWL-S. This paper studies the use of two well established standards in expressing plans and composite web services, namely the Planning Domain Definition Language (PDDL) and the Ontology Web Language for Services (OWL-S) and suggests a method for translating the produced PDDL plans of any planning system to OWL-S descriptions of the final composite web services. The result is a totally new web service that can later be discovered and invoked or even take part in a new composition. Post-conference book chapter version of: E. Ziaka, D. Vrakas, N. Bassiliades, “Translating Web Services Composition Plans to OWL-S Descriptions”, Proc. 3rd International Conference on Agents and Artificial Intelligence (ICAART 2011), pp. 167-176, 28-30 Jan 2011, Rome, Italy, 2011. Web services composition AI planning Semantic web services OWL-S PDDL Computational Models for Normative Multi-Agent Systems 3 Normative Multi-Agent Systems. Dagstuhl Follow-Ups Schloss Dagstuhl G. Andrighetto, G. Governatori, P. Noriega, L. W.N. van der Torre Vol. 4 2013 0 71-92 This chapter takes a closer look at computational logic approaches for the design, verification and the implementation of normative multi-agent systems. After a short overview of existing formalisms, architectures and implementation languages, an overview of current research challenges is provided. http%3A%2F%2Fdrops%2Edagstuhl%2Ede%2Fportals%2Fdfu%2Findex%2Ephp%3Fsemnr%3D13003 Norm verification Computational Architectures for Normative MAS Programming Normative Systems The Development of a New Framework for Managing Risks in the European Industry: The IRIS RISK PARADIGM 3 Industrial Safety and Life Cycle Engineering: Technologies / Standards / Applications VCE Helmut Wenzel ISBN 978-3-200-03179-1 2013 34 23-56 Motivation: Risk assessment has been performed in a fragmented way creating problems with interfaces and quantification. A consistent methodology for risk quantification is required. Main Results: The new IRIS Risk Paradigm provides a conceptual framework for consistent harmonized risk assessment and quantification. http%3A%2F%2Fwww%2Evce%2Eat%2Firis%2F The+IRIS+Project+%28Integrated+European+Industrial+Risk+Reduction+System%29 http%3A%2F%2Fwww%2Evce%2Eat%2Fsites%2Fdefault%2Ffiles%2Fuploads%2Fdownloads%2Findustrial%2Dsafety%2Dand%2Dlife%2Dcycle%2Dengineering%5Fabstract%2Epdf engineering risk risk assessment risk quantification knowledge portal risk ontology Adverse Drug Event Prevention in Neonatal Care: A Rule-Based Approach 2 European Federation of Medical Informatics special topic conference on Data and Knowledge for Medical Decision Support (EFMI STC 2013), Studies in Health Technology and Informatics IOS Press B. Blobel et al. Vol. 186 2013 0 170-174 Adverse drug events (ADE) in a neonatal unit can be of great importance due to the underlying nature and the special characteristics of the patients. This paper presents our work on the development of a knowledge base (KB) for supporting the identification and prevention of ADEs. First, a literature review was conducted to identify ADEs observed through the use of the most commonly-used drugs in a specific neonatal unit. Then, the acquired knowledge was encoded according to an ontological data model developed for the representation of the specific facts for the neonatal unit. Finally, a rule-based prototype consisting of 164 rules was implemented in order to represent and simulate the inference procedure about preventing ADEs. Prague, Czech Republic, April 17th-19th 2013 http%3A%2F%2Febooks%2Eiospress%2Enl%2Fpublication%2F32786 Adverse drug event (ADE) knowledge base rules neonatal unit Enabling Agent Reasoning over the Web 2 6th Balkan Conference on Informatics (BCI 2013) ACM 2013 8 259-266 The aim of this paper is to introduce a preliminary system that provides reasoning services over the Web. The system is based on extending the EMERALD framework for agent based reasoning services with a Web service interface. The approach is exemplified using an intelligent brokering sample scenario that involves the defeasible reasoners included into the EMERALD framework. Thessaloniki, Greece, 19-21 Sep 2013 http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1145%2F2490257%2E2490270 Web service reasoning software agent UML DEiXTo: A web data extraction suite 2 6th Balkan Conference on Informatics (BCI 2013) ACM 2013 4 9-12 Web data extraction (or web scraping) is the process of collecting unstructured or semi-structured information from the World Wide Web, at different levels of automation. It is an important, valuable and practical approach towards web reuse while at the same time can serve the transition of the web to the semantic web, by provid-ing the structured data required by the latter. In this paper we present DEiXTo, a web data extraction suite that provides an arsenal of features aiming at designing and deploying well-engineered extraction tasks. We focus on presenting the core pattern matching algorithm and the overall architecture, which allows programming of custom-made solutions for hard extraction tasks. DEiXTo consists of both freeware and open source components. BCI2013-kokkoras.pdf http%3A%2F%2Fdeixto%2Ecom DEiXTo Thessaloniki, Greece, 19-21 Sep 2013 http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1145%2F2490257%2E2490297 web data extraction web scraping pattern matching A Rule Based Personalized Location Information System for the Semantic Web 2 14th International Conference on Electronic Commerce and Web Technologies (EC-Web '13) Springer C. Huemer and P. Lops Lecture Notes in Business Information Processing, Volume 152 2013 12 27-38 In this paper, an innovative Personalized Location Information System for the Semantic Web (called SPLIS) is presented. The proposed system adopts sche-ma.org ontology and combines it with rule-based policies, to deliver fully contextualized information to the user of a location-based system. Owners of points of interest can add their own rule-based policies to SPLIS to expose and deploy their marketing strategy on special offers, discounts, etc. These rules are combined at run-time with information about relevant place properties and user (people) profiles. Additionally, owners of points of interest can extend the ontology by adding dynamically specific properties. Rules are encoded in RuleML for interchangeability and to Jess in order to be executed. All data and rules are stored in the form of triples, using Sesame. Rules are evaluated on-the-fly to deliver personalized information according to the rules that fired within the current user-location-time context. In the paper, a demonstration of SPLIS is given using data from Google Places API and Google map for visualization. splis-ECWEB2013.pdf http%3A%2F%2Ftinyurl%2Ecom%2Fsplis%2Dlogin SPLIS Prague, Czech Republic, 26-29 Aug 2013 http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%252F978%2D3%2D642%2D39878%2D0%5F3 Rules Ontologies Location Based Services Context Knowledge-based e Contract Negotiation among Agents Using Semantic Web Technologies 2 5th International Conference on Computational Collective Intelligence Technologies and Applications (ICCCI 2013) Springer C. Badica, N.T. Nguyen, and M. Brezovan LNAI 8083 2013 10 215-224 E-Commerce enabled new ways of transactions. Companies and individuals negotiate and make contracts every day. Practically, contracts are agreements between parties that must be kept. These agreements affect the involved parties irretrievably. Hence, negotiating them efficiently is proved vital. To this end we propose the use of intelligent agents, which benefit from Semantic Web technologies, such as RDF and RuleML, for data and policy exchanges. Each agent encounter is characterized by the interaction or negotiation protocol and each party’s strategy. This study defines a knowledge-based negotiation procedure where protocols and strategies are separated enabling reusability and thus enabling agent participation in interaction processes without the need of reprogramming. In addition, we present the integration of this methodology into a multi-agent knowledge-based framework and next a use case scenario using the contract net protocol that demonstrates the added value of the approach. ICCCI2013.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald%2Femerald%2Ehtml EMERALD Craiova, Romania, 11 - 13 Sep 2013 http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%252F978%2D3%2D642%2D40495%2D5%5F22 Semantic Web Agents e-Contract negotiation Reaction RuleML Congestion Management for Urban EV Charging Systems 2 4th IEEE International Conference on Smart Grid Communications (SmartGridComm 2013) IEEE 2013 6 We consider the problem of managing Electric Vehicle (EV) charging at charging points in a city to ensure that the load on the charging points remains within the desired limits while minimizing the inconvenience to EV owners.We develop solutions that treat charging points and EV users as self-interested agents that aim to maximize their profit and minimize the impact on their schedule. In particular, we propose variants of a decentralised and dynamic approach as well as an optimal centralised static approach. We evaluated these solutions in a real setting based on the road network and the location of parking garages of a UK city and show that the optimal centralised (nondynamic) solution manages the congestion the best but does not scale well, while the decentralised solutions scale to thousands of agents. paperSmartGridComm.pdf Vancouver, Canada, 21-24 Oct 2013 http%3A%2F%2Fsgc2013%2Eieee%2Dsmartgridcomm%2Eorg%2Fcontent%2Fieee%2Dsmartgridcomm Electric Vehicles Intelligent Agents Congestion Management Large-Scale Semantic Indexing of Biomedical Publications at BioASQ 2 BioASQ Workshop 2013 6 Automated annotation of scientific publications in real-world digital libraries requires dealing with challenges such as large number of concepts and training examples, multi-label training examples and hierarchical structure of concepts. BioASQ is a European project that contributes a large-scale biomedical publications corpus for working on these challenges. This paper documents the participation of our team to the large-scale biomedical semantic indexing task of BioASQ. tsoumakas-bioasq.pdf Valencia, Spain, September 27, 2013 Evaluating Feature Selection Methods for Multi-Label Text Classification 2 BioASQ Workshop 2013 12 Multi-label text classification deals with problems in which each document is associated with a subset of categories. These documents often consist of a large number of words, which can hinder the performance of learning algorithms. Feature selection is a popular task to find representative words and remove unimportant ones, which could speed up learning and even improve learning performance. This work evaluates eight feature selection algorithms in text benchmark datasets. The best algorithms are subsequently compared with random feature selection and classifiers built using all features. Results agree with literature by finding that well-known approaches, such as maximum chi-squared scoring across all labels, are good choices to reduce text dimensionality while reaching competitive multi-label classification performance. spolaor-bioasq.pdf Valencia, Spain, September 27, 2013 Multi-Label Classification Methods for Multi-Target Regression 5 arXiv preprint arXiv:1211.6581 arXiv 2014 16 Real world prediction problems often involve the simultaneous prediction of multiple target variables using the same set of predictive variables. When the target variables are binary, the prediction task is called multi-label classification while when the target variables are real-valued the task is called multi-target regression. Although multi-target regression attracted the attention of the research community prior to multi-label classification, the recent advances in this field motivate a study of whether newer state-of-the-art algorithms developed for multi-label classification are applicable and equally successful in the domain of multi-target regression. In this paper we introduce two new multi-target regression algorithms: multi-target stacking (MTS) and ensemble of regressor chains (ERC), inspired by two popular multi-label classification approaches that are based on a single-target decomposition of the multi-target problem and the idea of treating the other prediction targets as additional input variables that augment the input space. Furthermore, we detect an important shortcoming on both methods related to the methodology used to create the additional input variables and develop modified versions of the algorithms (MTSC and ERCC) to tackle it. All methods are empirically evaluated on 12 real-world multi-target regression datasets, 8 of which are first introduced in this paper and are made publicly available for future benchmarks. The experimental results show that ERCC performs significantly better than both a strong baseline that learns a single model for each target using bagging of regression trees and the state-of-the-art multi-objective random forest approach. Also, the proposed modification results in significant performance gains for both MTS and ERC. spyromitrosArxiv2014.pdf http%3A%2F%2Farxiv%2Eorg%2Fabs%2F1211%2E6581 mutli-target regression multi-output regression multivariate regression multi-label classification regressor chains stacking The Smart IHU Project Architecture for Energy and Ambient Intelligence Applications 5 LPIS Technical Reports 2013 0 This paper presents a complete architecture for a Smart University Building. The real-world deployment is based on a wide range of wireless sensor and actuator networks, inte- grated by a middleware based on the Service-Oriented Architec- ture of Web Services. The middleware provides the necessary basis for various energy monitoring, management and savings applications as well as Intelligent Agents in the context of Ambi- ent Intelligence. thanosSmart1TR.pdf S2REd: A Semantic Web Rule Editor 5 Technical Report TR-LPIS-409-13 2013 8 A key factor in the further progress of the Semantic Web is the development and wide-spread usage of rule- and logic-based applications. However, there is an evident lack of software tools that can assist end-users in developing such applications. Consequently, users usually resort to more generic tools that offer support at a syntactic level, but prove inadequate in semantically supporting the user. This paper presents S2REd, a Semantic Web rule editor that introduces a supplementary layer of semantic assistance during rule base development. The tool offers semantic assistance via: (i) The Semantic Tag Mapping window that provides a meta-modeling facility for generating schemas over various rule language versions and, (ii) the Namespace Dialog window, for loading ontologies that serve as the underlying vocabulary for expressing rule atoms. S2REd assists in developing RuleML rule bases, but is equally suitable for any other XML-based syntax for representing rule sets. TR-S2RED.pdf http%3A%2F%2Fsourceforge%2Enet%2Fprojects%2Fs2red%2F S2Red Semantic Web RuleML Rule Bases Rule Editor Ontologies Choreographing agent encounters in the Semantic Web using rules 1 Journal of Intelligent and Fuzzy Systems IOS Press Vol. 27, No. 2 2014 12 625-640 In order for automated agent-based e-Commerce transactions to blossom, well-defined, analyzable and easily customizable interaction protocols or choreographies of involved parties need to be developed. Although, several domain-depended protocols have already been developed, efficient methodologies and technologies for facilitating the definition, deployment, reuse and maintenance of interaction protocols should be developed. This paper proposes a rule-based, reusable, analyzable and easily comprehensible by the user choreography definition methodology, called K-SWAN. Τhe proposed choreography scheme separates the definition of the agent shared interaction protocol from the private agent interaction strategy and enables agents to choose the appropriate protocol for the transaction, from a library of re-usable interaction protocols, and automatically combine it with their personal strategy, from a private library, by using SW technologies for both. Complying with K-SWAN methodology will let agents participate seamlessly in different interaction processes and/or modify their behavior with a minimal programming effort. Finally, this paper presents the integration of the K-SWAN methodology into EMERALD, a multi-agent knowledge-based framework based on SW standards, which maximizes reusability and interoperability of behavior between agents. IOS-JIFS-final.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fprojects%2Fk%2Dswan%2F K%2DSWAN http%3A%2F%2Fiospress%2Emetapress%2Ecom%2Fcontent%2Fw103375072567438%2F%3Fp%3De3a9a5bc0414460e965e0fc30088ca96%26pi%3D5 Semantic Web Intelligent Agents Reactive Rules Agent Interaction Choreographies Agent Policies The 9th annual MLSP competition: New methods for acoustic classification of multiple simultaneous bird species in a noisy environment 2 Proc. 2013 IEEE International Workshop on Machine Learning for Signal Processing (MLSP) IEEE 2013 0 Birds have been widely used as biological indicators for ecological research. They respond quickly to environmental changes and can be used to infer about other organisms (e.g., insects they feed on). Traditional methods for collecting data about birds involves costly human effort. A promising alternative is acoustic monitoring. There are many advantages to recording audio of birds compared to human surveys, including increased temporal and spatial resolution and extent, applicability in remote sites, reduced observer bias, and potentially lower cost. However, it is an open problem for signal processing and machine learning to reliably identify bird sounds in real-world audio data collected in an acoustic monitoring scenario. Some of the major challenges include multiple simultaneously vocalizing birds, other sources of non-bird sound (e.g., buzzing insects), and background noise like wind, rain, and motor vehicles. Using RuleML for Representing and Prolog for Simulating Fuzzy Cognitive Maps 3 Fuzzy Cognitive Maps for Applied Sciences and Engineering Springer Elpiniki I. Papageorgiou Intelligent Systems Reference Library, Volume 54 2014 22 65-87 Fuzzy Cognitive Map (FCM) technique is broadly used for decision making and predictions by experts and scientists of a wide range of disciplines. The use of the FCMs would be even wider if a standardized representation of FCMs was developed and a system that would simulate them was constructed. Having such a system, decision makers would be able to create and examine their own developed Fuzzy Cognitive Maps, and also distribute them e.g. through Internet. In this chapter, a) we propose a RuleML representation of FCMs and b) we present the design and implementation of a system that assists experts to simulate their own FCMs. This system, which is developed using the Prolog programming language, makes the results of the FCM simulation directly available to other co-operative systems because it returns them in standard RuleML syntax. In the chapter, the design choices of the implemented system are discussed and the capabilities of the RuleML representation of FCM are presented. The use of the system is exhibited by a number of examples concerning an e-business company. tsadiras04.pdf http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%2F978%2D3%2D642%2D39739%2D4%5F4 Fuzzy Cognitive Maps RuleML Prolog Model-based reinforcement learning for humanoids: A study on forming rewards with the iCub platform 2 Computational Intelligence, Cognitive Algorithms, Mind, and Brain (CCMB), 2013 IEEE SSCI IEEE 2013 0 87-93 Technological advancements in robotics and cognitive science are contributing to the development of the field of cognitive robotics. Modern robotic platforms are able to exhibit the ability to learn and reason about complex tasks and to follow behavioural goals in complex environments. Nevertheless, many challenges still exist. One of these great challenges is to equip these robots with cognitive systems that allow them to deal with less constrained situations, beyond constrained scenarios as in industrial robotics. In this work we explore the application of the Reinforcement Learning (RL) paradigm to study the autonomous development of robot controllers without a priori supervised learning. Such a model-based RL architecture is discussed for the cognitive implications of applying RL in humanoid robots. To this end we show a developmental framework for RL in robotics and its implementation and testing for the iCub robotic platform in two novel experimental scenarios. In particular we focus on iCub simulation experiments with comparisons between internal perception-based reward signals and external ones, in order to compare learning performance of the robot guided by its own perception of action's outcomes with the one when the robot has its actions externally evaluated. CCMB.pdf Singapore Pattern discovery for microsatellite genome analysis 1 Computers in Biology and Medicine Elsevier Edward John Ciaccio 2014 8 Microsatellite loci comprise an important part of eukaryotic genomes. Their applications in biology as genetic markers are related to numerous fields ranging from paternity analyses to construction of genetic maps and linkage to human disease. Existing software which offer pattern discovery algorithms for the correct identification and downstream analysis of microsatellites are scarce and are proving to be inefficient to analyse the large, exponentially increasing, sequenced genomes. Moreover, such analyses can be very difficult for bioinformatically inexperienced biologists. In this paper we present MiGA (Microsatellite Genome Analysis) software for the detection of all microsatellite loci in genomic data through a user friendly interface. The algorithm searches exhaustively and rapidly for most microsatellites. Contrary to other applications, MiGA takes into consideration the following three most important aspects: The efficiency of the algorithm, the usability of the software and the plethora of offered summary statistics. All the above, help biologists to obtain basic quantitative and qualitative information regarding the presence of microsatellites in genomic data as well as downstream processes, such as selection of specific microsatellite loci for primer design and comparative genome analysis. http://dx.doi.org/10.1016/j.compbiomed.2014.01.002i http%3A%2F%2Fmlkd%2Ecsd%2Eauth%2Egr%2Fbio%2Fmiga%2Findex%2Ehtml Applications+website http%3A%2F%2Fauthors%2Eelsevier%2Ecom%2Fsd%2Farticle%2FS0010482514000043 Genome Analysis Bioinformatics software Mining methods Pattern Discovery Microsatellites Simple sequence repeats Polyadenylation site prediction using PolyA-iEP method 3 Polyadenylation Method and Protocols Springer Joanna Rorbach and Agnieszka Bobrowicz Methods In Molecular Biology, 1125 2014 0 131-140 This chapter presents a method called PolyA-iEP that has been developed for the prediction of polyadenylation sites. More precisely PolyA-iEP is a method that recognizes mRNA 3’ends which contain polyadenylation sites. It is a modular system which consists of two main components. The first exploits the advantages of emerging patters and the second is a distance-based scoring method. The outputs of the two components are finally combined by a classifier. The final results reach very high scores of sensitivity and specificity. Data Mining Machine Learning Classification Emerging Patterns Bioinformatics Polyadenylation Autonomous Selection of Inter-Task Mappings in Transfer Learning 2 2013 AAAI Spring Symposium Series AAAI 2013 0 AAAI.pdf Stanford, U.S.A. Reinforcement Learning Agents Providing Advice in Complex Video Games 1 Connection Science Taylor & Francis (in press) 2014 0 This paper introduces a teacher-student framework for reinforcement learning. In this frame- work, a teacher agent instructs a student agent by suggesting actions the student should take as it learns. However, the teacher may only give such advice a limited number of times. We present several novel algorithms that teachers can use to budget their advice effectively, and we evaluate them in two experimental domains: Mountain Car and Pac-Man. Our results show that the same amount of advice, given at different moments, can have different effects on student learning, and that teachers can significantly affect student learning even when students use different learning methods and state representations. CONNECTION.pdf A holistic virtual laboratory on wireless communications and sensor networks 1 Journal of Interactive Mobile Technologies International Association of Online Engineering 2013 0 holistic.pdf Learning to play Monopoly: A Reinforcement Learning approach 2 AISB 2014: AI & GAMES (in press) 2014 0 Reinforcement Learning is a rather popular machine learning paradigm which relies on an agent interacting with an environment and learning through trial and error to maximize the cummulative sum of rewards received by it. In this paper, we are proposing a novel representation of the famous board game Monopoly as a Markov Decision Process and a Reinforcement Learning agent capable of playing and learning winning strategies. The conclusions drawn from the experiments are particularly positive, since the proposed agent demonstrated intelligent behavior and high win rates against different types of agent-players. AISB.pdf London, U.K. Agents Teaching Humans in Reinforcement Learning Tasks 2 Adaptive Learning Agents 2014 (in press) 2014 0 This paper extends our existing teacher-student framework to allow a knowledgeable agent to teach human students. An agent teacher instructs a human student by suggesting actions the student should take as it learns. This paper extends previous algorithms, used for agents teaching other agents, to develop several new algorithms for agents teaching humans. Our results in the Pac-Man domain show that our new approaches can indeed be effectively used to improve human learning. Moreover, some of these human-teaching approaches perform better than some of the original algorithms when one agent teaches another agent. ALA14.pdf Paris, France An Autonomous Transfer Learning Algorithm for TD-Learners 2 Proc. 8th Hellenic Conference on Artificial Intelligence (SETN 2014) (in press) 2014 0 The main objective of transfer learning is to use the knowledge ac- quired from a source task in order to boost the learning procedure in a target task. Transfer learning comprises a suitable solution for reinforcement learning algo- rithms, which often require a considerable amount of training time, especially when dealing with complex tasks. This work proposes an autonomous method for transfer learning in reinforcement learning agents. The proposed method is empirically evaluated in the keepaway and the mountain car domains. The results demonstrate that the proposed method can improve the learning procedure in the target task. SETN_TD2014.pdf Ioannina, Greece Rule-based Approaches for Energy Savings in an Ambient Intelligence Environment 1 Journal of Pervasive and Mobile Computing Elsevier Vol. 19 2015 0 1-23 This paper presents a novel real-world application for energy savings in a Smart Building environment. The proposed system unifies heterogeneous wireless sensor networks under a Semantic Web Service middleware. Two complementary and mutually exclusive rule-based approaches for enforcing energy-saving policies are proposed: a reactive agent based on production rules and a deliberative agent based on defeasible logic. The system was deployed at a Greek University, showing promising experimental results (at least 4% daily savings). Although the percentage of energy savings may seem low, the greatest merit of the method is ensuring no energy is wasted by constantly enforcing the policies. thanosStratosJPMC2014.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Epmcj%2E2014%2E05%2E001 defeasible logic smart building energy savings ambient intelligence semantic web services Feature Evaluation Metrics for Population Genomic Data 2 Proceedings of 8th Hellenic Conference on Artificial Intelligence (SETN 2014). Springer A. Likas, K. Blekas and D. Kalles Artificial intelligence: Methods and Applications, LNCS, 8445 2014 0 436-441 Single Nucleotide Polymorphisms (SNPs) are considered nowadays one of the most important class of genetic markers with a wide range of applications with both scientific and economic interests. Although the advance of biotechnology has made feasible the production of genome wide SNP datasets, the cost of the production is still high. The transformation of the initial dataset into a smaller one with the same genetic information is a crucial task and it is performed through feature selection. Biologists evaluate features using methods originating from the field of population genetics. Although several studies have been performed in order to compare the existing biological methods, there is a lack of comparison between methods originating from the biology field with others originating from the machine learning. In this study we present some early results which support that biological methods perform slightly better than machine learning methods. Kavakiotis_SETN14.pdf Ioannina, Greece, May 15-17, 2014 Machine learning Bioinformatics SNPs Single nucleotide polymorphism Feature selection A Comprehensive Study over VLAD and Product Quantization in Large-scale Image Retrieval 1 IEEE Transactions on Multimedia IEEE 2014 16 This paper deals with content-based large-scale image retrieval using the state-of-the-art framework of VLAD and Product Quantization proposed by Jegou et al. [1] as a starting point. Demonstrating an excellent accuracy-efficiency trade-off, this framework has attracted increased attention from the community and numerous extensions have been proposed. In this work, we make an in-depth analysis of the framework that aims at increasing our understanding over its different processing steps and boosting its overall performance. Our analysis involves the evaluation of numerous extensions (both existing and novel) as well as the study of the effects of several unexplored parameters. We specifically focus on a) employing more efficient and discriminative local features, b) improving the quality of the aggregated representation, and c) optimizing the indexing scheme. Our thorough experimental evaluation provides new insights into extensions that consistently contribute and others that do not to performance improvement, and sheds light into the effects of previously unexplored parameters of the framework. As a result, we develop an enhanced framework that significantly outperforms the previous best reported accuracy results on standard benchmarks and is more efficient. spyromitrosIEEETMM2014.pdf image retrieval indexing image classification A Multi-Agent Coordination Framework for Smart Building Energy Management 2 Proc. 3nd International Workshop on Artificial Intelligence Techniques for Power Systems and Energy Markets (IATEM) in conjunction with DEXA 2014 IEEE ISBN 978-1-4799-5721-7 2014 6 126-130 This paper presents a novel energy management framework for multi-agent coordination in smart buildings. The framework builds on top of an existing Service-Oriented middleware for Ambient Intelligence, which offers sensor and actuator functions of wireless devices. The middleware also provides a semantics infrastructure that assists in authoring agent policies for reducing energy consumption and maximizing user comfort. Each agent within the framework is responsible for monitoring the environmental context and controlling the electrical appliances of a specific room. However, the collective behavior of the multi-agent system is controlled by a Coordinator Agent that approves or rejects the allocation of building resources in time, aiming at more “long-term” goals that are out of the reach and scope of the individual Room Agents. The agents’ underlying logic is expressed via defeasible logics, a formalism offering intuitive knowledge representation and advanced conflict resolution mechanisms. thanosIATEM2014.pdf Munich, Germany, September 1 - 4, 2014 http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1109%2FDEXA%2E2014%2E39 multi-agent systems ambient intelligence smart environmnets web services context-aware systems An Applied Energy Management Approach in Intelligent Environments based on a Hybrid Agent Architecture 2 Proc. 3rd International Workshop on AI Problems and Approaches for Intelligent Environments (AI4IE) in conjunction with ECAI 2014 http://ceur-ws.org/ (to appear) 2014 6 This paper presents a novel energy management framework for multi-agent coordination in smart buildings. The framework builds on top of an existing Service-Oriented middleware for Ambient Intelligence, which offers sensor and actuator functions of wireless devices. The middleware also provides a semantics infrastructure that assists in authoring agent policies for reducing energy consumption and maximizing user comfort. Each agent within the framework is responsible for monitoring the environmental context and controlling the electrical appliances of a specific room. However, the collective behavior of the multi-agent system is controlled by a Coordinator Agent that approves or rejects the allocation of building resources in time, aiming at more “long-term” goals that are out of the reach and scope of the individual Room Agents. The agents’ underlying logic is expressed via defeasible logics, a formalism offering intuitive knowledge representation and advanced conflict resolution mechanisms. thanosAI4IE2014.pdf Prague, Czech Republic, August 18-22, 2014 Proceedings of the 4th International Conference on Web Intelligence, Mining and Semantics (WIMS14) 4 The ACM International Conference Proceedings Series ACM, New York, USA ISBN: 978-1-4503-2538-7 2014 0 It is our great pleasure to welcome you at the International Conference on Web Intelligence, Mining and Semantics (WIMS’14) taking place on June 2nd – 4th at Thessaloniki, Greece. The WIMS conference series was initiated by the Western Norway Research Institute, as a part of the celebration of their 25th anniversary. The 2nd international conference in this series WIMS’12 was organized at the University of Craiova, Romania, and the 3rd conference (WIMS’13) – at Autonomous University of Madrid, Spain. This year, the 4th WIMS conference (WIMS’14) is organized under the auspices of the Aristotle University of Thessaloniki, in Greece. This conference is intended to stimulate discussions on the forefront of research concerned with intelligent approaches to transform the World Wide Web into a global reasoning and semantics-driven computing machine. We believe that the scope of the WIMS’14 will serve the interest of the scientific community as well as relevant industries and general public. 27 full papers, 13 short papers and 1 discussion paper were accepted for the presentation at the conference, in 12 sessions, and publication in the electronic proceedings and the ACM Digital Library. This conference also features 3 distinguished keynote addresses, presented in 3 plenary sessions, 2 workshops and 1 tutorial, presented in 6 more sessions. http%3A%2F%2Fwims14%2Ecsd%2Eauth%2Egr%2F WIMS%2D14+web+site 2-4 June 2014, Thessaloniki, Greece http%3A%2F%2Fdl%2Eacm%2Eorg%2Fcitation%2Ecfm%3Fid%3D2611040 Scalable Web and Data Architectures and Infrastructures Web Intelligence Web Mining, Information and Knowledge Extraction Web Semantics and Reasoning Evaluation and Validation of WIMS Technologies and Applications WIMS Applications Agent Reasoning on the Web using Web Services 1 Computer Science and Information Systems Vol. 11, No. 2 2014 24 697-721 In this paper we present an approach for reusing agent-based reasoning capabilities by making them available for invocation as Web services. In this way, we provide the missing link between the highly interoperable Web services and the autonomicity and intelligence of agent-based systems, so that the latter can be seamlessly integrated into the knowledge-rich SemanticWeb environment without being compromised by isolated communication platforms and languages or restricted to only one or just few reasoning formalisms. We have achieved this by extending the EMERALD framework for agent based reasoning with a Web service interface. Our approach is exemplified by the development of an online system for intelligent brokering of apartment rentals. The broker intelligence is captured as a defeasible knowledge base, while its problem solving process involves the invocation of third party defeasible reasoning Web services included into the EMERALD framework. comsis-paper-revised.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fprojects%2Fk%2Dswan%2F K%2DSWAN+Project http%3A%2F%2Fwww%2Ecomsis%2Eorg%2Farchive%2Ephp%3Fshow%3Dpprbci%2D1303 Web service reasoning software agent UML Research and Applications in Web Intelligence, Mining, and Semantics 2 Proc. 4th International Conference on Web Intelligence, Mining and Semantics (WIMS’14) ACM, New York, NY, USA DOI=10.1145/2611040.2611045 2014 6 Article 0 The Web has an enormous influence on our everyday life. Thus, more efficient intelligent approaches and technologies are needed to realize the Web's full potential. Intelligence can be achieved by making the Web aware of the semantics of its own structures and content and by applying intelligent techniques to effectively access web resources. The Semantic Web was one of the significant steps towards bringing Intelligence to the Web. Based on this starting point, the Web Intelligence, Mining, and Semantics (WIMS) community works toward researching and implementing the next generation of the intelligent Web for humans and machines. In this editorial, opening the volume of the proceedings of WIMS'14, we review the topics of interest for the WIMS community, analyze the response of this year's authors to these topics, and present the program of the conference. We hope that this material will be useful for a reader as a key for the structure and content of these proceedings. http%3A%2F%2Fwims14%2Ecsd%2Eauth%2Egr%2F WIMS%2714 Thessaloniki, Greece, June 2-4 2014 http%3A%2F%2Fdl%2Eacm%2Eorg%2Fcitation%2Ecfm%3Fdoid%3D2611040%2E2611045 Web intelligence web semantics web mining scalable web data architectures information extraction knowledge extraction reasoning application case study evaluation validation methodology Providing a context-aware location based web service through semantics and user-defined rules 2 Proc. 4th International Conference on Web Intelligence, Mining and Semantics (WIMS’14) ACM, New York, NY, USA DOI=10.1145/2611040.261 2014 11 Article 9 In this paper, the design and the implementation of a novel context-aware location based service is presented, called "Geo SPLIS/Geographic Semantic Personalized Location Information System". Geo SPLIS offers users the capability to add their own contextualized preferences regarding Points of Interests (POIs) and combines them with POI owners group targeted offers to deliver high quality personalized information. In order to achieve this, the presented system a) collects data from external sources such as Google Places API, POIs' websites and Google+b) adopts the schema.org ontology to represent people and places profiles, c) provides a user friendly web editor for adding rules at run time, d) uses RuleML and Jess compatible rules to model user preferences and group-targeted place offers and make them machine executable, e) stores data and rules in the Sesame RDF triple store and f) evaluates these data and rules on-the-fly so that to deliver POIs and offers matching user context, presented on Google Maps. Geo SPLIS aims to address some issues regarding knowledge-based personalization in location based services and provide a collaborative knowledge creation platform for other systems in the web. http%3A%2F%2Ftinyurl%2Ecom%2FGeoSPLIS GeoSPLIS Thessaloniki, Greece, June 2-4 2014 http%3A%2F%2Fdl%2Eacm%2Eorg%2Fcitation%2Ecfm%3Fdoid%3D2611040%2E2611083 Semantic web Ontologies Rules Context Location Based Services Points of Interest Preferences Group-Targeted Offers Using Rules to Develop a Personalized and Social Location Information System for the Semantic Web 2 Proc. 8th International Web Rule Symposium (RuleML 2014), co-located with the 21st European Conference on Artificial Intelligence (ECAI 2014) Springer A. Bikakis et al. LNCS 8620 2014 15 82-96 In this work, the design and implementation of an innovative context-aware location based social networking service is presented. The proposed system, called “Geosocial SPLIS”, utilizes Semantic Web technologies to deliver personalized information to the end user. It addresses some drawbacks of knowledge-based personalization systems and aims to provide a collaborative knowledge creation platform for other systems. To achieve this, it a) collects data from external sources such as Google Places API and Google+ b) adopts the schema.org ontology to represent people and places profiles, c) provides a web editor for adding rules (modeling user preferences and group-targeted place offers) at run time, d) uses RuleML and Jess rules to represent these rules, e) combines at run-time the above to match user context with up to date information, presented on Google Maps and f) matches user’s preferences with those of his/her nearby friends to present POI’s that are suitable to all of them. All data and rules are stored in the Sesame RDF triple store in order to be shared among various systems. RuleML2014-paper_32.pdf http%3A%2F%2Ftinyurl%2Ecom%2FGeoSPLIS GeoSPLIS Prague, Czech Republic, August 18-20, 2014 http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%252F978%2D3%2D319%2D09870%2D8%5F6 Semantic Web Ontologies Rules Context Location Based Services Points of Interest Preferences Group-Targeted Offers Geosocial SPLIS: A Rule-Based Service for context-aware point of interest exploration 2 8th International Rule Challenge at RuleML 2014, co-located with the 21st European Conference on Artificial Intelligence (ECAI 2014) CEUR Workshop Proceedings Th. Patkos, A. Wyner, A. Giurca 1211 2014 14 This paper presents the design and implementation of a novel geosocial semantic service called “Geosocial SPLIS” (GeoSocial Semantic Per-sonalizing Location Information Service). The service a) gets data about points of interest (POIs) and user profiles from external sources such as Google Places API and Google+, b) adopts the well known schema.org ontology, c) supports a user friendly web editor so as regular users to be able to insert and customize their daily preferences about points of interest (POIs), and POI owners their group targeted offers, d) uses RuleML and Jess rules to make this rules machine comprehensible, e) presents contextualized information on Google Maps. ruleml2014challenge_submission_5.pdf http%3A%2F%2Ftinyurl%2Ecom%2FGeoSPLIS Geosocial+SPLIS Prague, Czech Republic, August 18-20, 2014 http%3A%2F%2Fceur%2Dws%2Eorg%2FVol%2D1211%2F Semantic Web Ontologies Rules Context Location Based Services Points of Interest Preferences Group-Targeted Offers Branty: A Social Media Ranking Tool For Brands 2 Proceedings of ECML PKDD 2014 Springer 2014 4 In the competitive world of popular brands, strong presence in social media is of major importance for customer engagement and products advertising. Up to now, many such tools and applications enable end-users to observe and monitor their company's web profile, their statistics, as well as their market outreach and competition status. This work goes beyond the individual brands statistics since it automates a brand ranking process based on opinions emerging in social media users' posts. Twitter streaming API is exploited to track micro-blogging activity for a number of famous brands with emphasis on users' opinions and interactions. The social impact is captured from 3 different perspectives (objective counts, opinion reckoning, influence analysis), which estimate a score assigned to each brand via a multi-criteria algorithm. The results are then exposed in a Web application as a list of the most social brands on Twitter. But, are conventional metrics, such as followers, enough in order to measure the social impact of a brand? Different usage scenarios of our application reveal that the social presence of a brand is more complex than current social impact frameworks care to admit. branty.pdf Nancy, France Multi-Target Regression via Random Linear Target Combinations 2 Proceedings of ECML PKDD 2014 Springer 2014 16 Multi-target regression is concerned with the simultaneous prediction of multiple continuous target variables based on the same set of input variables. It arises in several interesting industrial and environmental application domains, such as ecological modelling and energy forecasting. This paper presents an ensemble method for multi-target regression that constructs new target variables via random linear combinations of existing targets. We discuss the connection of our approach with multi-label classification algorithms, in particular RA$k$EL, which originally inspired this work, and a family of recent multi-label classification algorithms that involve output coding. Experimental results on 12 multi-target datasets show that it performs significantly better than a strong baseline that learns a single model for each target using gradient boosting and compares favourably to multi-objective random forest approach, which is a state-of-the-art approach. The experiments further show that our approach improves more when stronger unconditional dependencies exist among the targets. tsoumakas-ecml-pkdd-2014.pdf Nancy, France Ensemble Approaches for Large-Scale Multi-Label Classification and Question Answering in Biomedicine 2 Proceedings BioASQ 2014 Workshop 2014 0 This paper documents the systems that we developed for our participation in the BioASQ 2014 large-scale bio-medical semantic indexing and question answering challenge. For the large-scale semantic indexing task, we employed a novel multi-label ensemble method consisting of support vector machines, labeled Latent Dirichlet Allocation models and meta-models predicting the number of relevant labels. This method proved successful in our experiments as well as during the competition. For the question answering task we combined different techniques for scoring of candidate answers based on recent literature. bioasq2014.pdf Sheffield, UK Drawing Parallels between Multi-Label Classification and Multi-Target Regression 2 ECML PKDD 2014 Workshop on Multi-Target Prediction 2014 1 Learning from multi-label data has recently received increased attention by researchers working on machine learning and data mining for two main reasons. The first one is the ubiquitous presence of multi-label data in application domains ranging from multimedia information retrieval to tag recommendation, query categorization, gene function prediction, medical diagnosis, drug discovery and marketing. The other reason is a number of challenging research problems involved in multi-label learning, such as dealing with label rarity, scaling to large number of labels and exploiting label relationships (e.g. hierarchies), with the most prominent one being the explicit modelling of label dependencies. Multi-label learning is closely related to multivariate regression, also known as multi-output or multi-target regression, which aims at predicting multiple real-valued target variables instead of binary ones. Despite that multi-target regression is a less popular task, it still arises in several interesting domains, such as predicting the wind noise of vehicle components, stock price prediction, ecological modelling and more recently energy-related forecasting, such as wind and solar energy production forecasting and load/price forecasting. Multi-label learning is often treated as a special case of multi-target regression in statistics. However, we could more precisely state that both are instances of the more general learning task of predicting multiple targets, which could be real-valued, binary, ordinal, categorical or even of mixed type. The baseline approach of learning a separate model for each target applies to both learning tasks. Most importantly, they share the same core challenge of modelling dependencies among the different targets. In this talk we will emphasize this tight relationship between multi-label classification and multi-target regression by discussing their similarities, while also highlighting their differences. We will further discuss techniques that can inherently handle both tasks. Motivated from their tight connection, we will discuss pathways for transferring recent advances in the more popular multi-label learning task to multi-target regression and will give examples of our recent work in this direction. Finally, we will present our recent extension of the Mulan librarywith methods for multi-target regression. tsoumakas-mtp.pdf Nancy, France Label Construction for Multi-label Feature Selection 2 Proceedings of the 2014 Brazilian Conference on Intelligent Systems (BRACIS) 2014 6 Multi-label learning handles datasets where each instance is associated with multiple labels, which are often correlated. As other machine learning tasks, multi-label learning also suffers from the curse of dimensionality, which can be mitigated by dimensionality reduction tasks, such as feature selection. The standard approach for multi-label feature selection transforms the multi-label dataset into single-label datasets before using traditional feature selection algorithms. However, this approach often ignores label dependence. This work proposes an alternative method, LCFS, which constructs new labels based on relations between the original labels to augment the label set of the original dataset. Afterwards, the augmented dataset is submitted to the standard multi-label feature selection approach. Experiments using Information Gain as a measure to evaluate features were carried out in 10 multi-label benchmark datasets. For each dataset, the quality of the features selected was assessed by the quality of the classifiers built using the features selected by the standard approach in the original dataset, as well as in the dataset constructed by four LCFS settings. The results show that setting LCFS with simple strategies using pairs of labels gives rise to better classifiers than the ones built using the standard approach in the original dataset. Moreover, these good results are accomplished when a small number of features are selected. bracis2014.pdf Sao Carlos, Brazil Collecting University Rankings for Comparison Using Web Extraction and Entity Linking Techniques 3 ICT in Education, Research and Industrial Applications Springer-Verlag V. Ermolayev et al. CCIS, Vol. 469 2014 25 23-46 University rankings are rankings of institutions in higher education, ordered by combinations of factors. Rankings are conducted by various organizations, such as news media, websites, governments, academics and private corporations. Due to huge financial and other interests, the rankings of universities worldwide recently received increasing attention. The rankings are based on different criteria and collect data in various ways. As a result, there is a large divergence in the specific rankings of different institutions. In order to compare rankings so that safe conclusions about their reliability are drawn, data from the sites of different such ranking lists must be collected. In this paper we present this first step for university ranking comparison, namely we discuss in detail how we have developed a Prolog application, called URank, that collects the data, by a) extracting them from the various ranking list web sites using web data extraction techniques, b) uniquely identifying the University entities within the above lists by linking them to the DBpedia linked open data set, and c) constructing a combined data set by merging the individual ranking list data sets using their DBpedia URI as a primary key. icteri-bassiliades.pdf http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%252F978%2D3%2D319%2D13206%2D8%5F2%23 University rankings Web data extraction Entity linking Linked open data Semantic Web A Survey of Agent Platforms 1 Journal of Artificial Societies and Social Simulation SIMSOC Consortium 18 (1) 2015 0 11 From computer games to human societies, many natural and artificial phenomena can be represented as multi-agent systems. Over time, these systems have been proven a really powerful tool for modelling and understanding phenomena in fields, such as economics and trading, health care, urban planning and social sciences. However, although, intelligent agents have been around for years, their actual implementation is still in its early stages. Since the late nineties many agent platforms have been developed. Some of them have already been abandoned whereas others continue releasing new versions. On the other hand, the agent-oriented research community is still providing more and more new platforms. This vast amount of platform options leads to a high degree of heterogeneity. Hence, a common problem is how people interested in using multi-agent systems should choose which platform to use in order to benefit from agent technology. This decision was usually left to word of mouth, past experiences or platform publicity, lately however people depend on solid survey articles. To date, in most cases multi-agent system surveys describe only the basic characteristics of a few representatives without even providing any classification of the systems themselves. This article presents a comparative up-to-date review of the most promising existing agent platforms that can be used. It is based on universal comparison and evaluation criteria, proposing classifications for helping readers to understand which agent platforms broadly exhibit similar properties and in which situations which choices should be made. http%3A%2F%2Fjasss%2Esoc%2Esurrey%2Eac%2Euk%2F18%2F1%2F11%2Ehtml Intelligent Agents Multi-agent Systems Agent platforms SocialSensor: Finding Diverse Images at MediaEval 2014 2 CEUR Workshop Proceedings CEUR-WS.org Vol-1263 2014 2 2 This paper describes the participation of the SosialSensor team in the Retrieving Diverse Social Images Task of MediaEval 2014. All our entries are produced by a different instantiation (set of features, parameter configuration) of the same diversification algorithm that optimizes a joint relevance-diversity criterion. All our runs are automated and use only resources given by the task organizers. Our best results in terms of the social ranking metric (F1@20 ~ 0.59) came by the runs that combine visual and textual information, followed by the visual-only run. spyromitrosMediaeval2014.pdf Barcelona, Catalunya, Spain http%3A%2F%2Fceur%2Dws%2Eorg%2FVol%2D1263%2Fmediaeval2014%5Fsubmission%5F36%2Epdf diversification image retrieval relevance detector greedy optimization visual features Dynamic Ensemble Pruning based on Multi-Label Classification 1 Neurocomputing Elsevier Volume 150, Part B 2015 12 501-512 Dynamic (also known as instance-based) ensemble pruning selects a (potentially) different subset of models from an ensemble during prediction based on the given unknown instance with the goal of maximizing prediction accuracy. This paper models dynamic ensemble pruning as a multi-label classification task, by considering the members of the ensemble as labels. Multi-label training examples are constructed by evaluating whether ensemble members are accurate or not on the original training set via cross-validation. We show that classification accuracy is maximized when learning algorithms that optimize example-based precision are used in the multi-label classification task. Results comparing the proposed framework against state-of-the-art dynamic ensemble pruning approaches in a variety of datasets using a heterogeneous ensemble of 200 classifiers show that it leads to significantly improved accuracy. markatopoulou-2015.pdf http%3A%2F%2Fhttp%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eneucom%2E2014%2E07%2E063 Semantically Aware Web Service Composition Through AI Planning 1 International Journal on Artificial Intelligence Tools World Scientific Nikolaos Bourbakis 24 2014 25 1450015 Web service composition is a significant problem as the number of available web services increases; however, manual composition is not an efficient option. Automated web service composition can be performed using AI Planning techniques, utilizing descriptions of available atomic web services, enhanced with semantic awareness and relaxation. This paper discusses a unified , semantically aware approach, handling both semantic (OWL-S&SAWSDL) and non - semantic (WSDL) web service descriptions. In the first case, ontology analysis is adopted to semantically enhance the planning domains and problems, in order to deal with cases where exact syntactic input - to - output matching is not feasible. In the non-semantic descriptions case, semantic information is acquired utilizing alternative sources such as lexical thesauri. Concept similarity measures are applied and utilized to achieve the desired degree of semantic relaxation. The solution to a web service composition problem is a plan describing the desired composite service . To support the proposed approach, the PORSCE framework has been implemented. The framework is modular, integrating discrete web service description languages and semantic relaxation techniques. Based on the similarity measures suggested in the paper, performance issues are also explored. http%3A%2F%2Fwww%2Eworldscientific%2Ecom%2Fdoi%2Fabs%2F10%2E1142%2FS0218213014500158 Policies Production System for Ambient Intelligence Environments 2 8th Hellenic Conference on Artificial Intelligence (SETN 2014) Springer International Publishing A. Likas, K. Blekas, D. Kalles LNAI 8445 2014 13 251-263 This paper presents a tool for designing Policies that govern the operation of an Ambient Intelligence (AmI) environment in order to minimize energy consumption and automate every-day tasks in smart settlements. This tool works on top of a semantic web services middleware and interacts with the middleware’s ontology in order to facilitate the designing, monitoring and execution of user defined rules that control the operation of a network of heterogeneous sensors and actuators. Furthermore, it gives the user the capability to organize these rules in tasks, in order to aggregate and discern relative rules. The main objective of this system is to provide a better monitoring and management of the resources, so as to achieve energy efficiency and reduce power consumption. The work presented in this paper is part of the Smart IHU project, which is developed at International Hellenic University. Ioannina http%3A%2F%2Fwww%2Egoogle%2Egr%2Furl%3Fsa%3Dt%26rct%3Dj%26q%3D%26esrc%3Ds%26source%3Dweb%26cd%3D1%26ved%3D0CCUQFjAA%26url%3Dhttp%253A%252F%252Flink%2Espringer%2Ecom%252Fchapter%252F10%2E1007%25252F978%2D3%2D319%2D07064%2D3%5F20%26ei%3DfTOIVPWYIYrcywOc74KYBw%26usg%3DAFQjCNF2y8CnYtPKsWkKOOsT5c13%2Dm8LBA%26sig2%3Dgkt8XOv1SDVn8XxDR0v7Cw%26bvm%3Dbv%2E81456516%2Cd%2EbGQ%26cad%3Drja Managing Electric Vehicles in the Smart Grid Using Artificial Intelligence: A Survey 1 IEEE Transactions on Intelligent Transportation Systems IEEE (to appear) 2015 17 Along with the development of Smart Grids, the wide adoption of Electric Vehicles (EVs) is seen as a catalyst to the reduction of CO2 emissions and more intelligent transportation systems. In particular, EVs augment the grid with the ability to store energy at some points in the network and give it back at others and therefore help optimise the use of energy from intermittent renewable energy sources and let users refill their cars in a variety of locations. However, a number of challenges need to be addressed if such benefits are to be achieved. On the one hand, given their limited range and costs involved in charging EV batteries, it is important to design algorithms that will minimise costs while avoid users being stranded. On the other hand, collectives of EVs need to be organized in such a way as to avoid peaks on the grid that may result in high electricity prices and overload local distribution grids. In order to meet such challenges, a number of technological solutions have been proposed. In this paper, we focus on those that utilise artificial intelligence techniques to render EVs and the systems that manage collectives of EVs smarter. In particular, we provide a survey of the literature and identify the commonalities and key differences in the approaches. This allows us to develop a classification of key techniques and benchmarks that can be used to advance the state-of-the art in this space. ITS_rigas.pdf http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1109%2FTITS%2E2014%2E2376873 A Context-Aware Web-Mapping System for Group-Targeted Offers Using Semantic Technologies 1 Expert Systems with Applications Elsevier Vol. 42, No. 9 2015 0 4443-4459 Existing location based systems that propose offers to their users do not provide points of interest (POIs) owners with the capability to flexibly customize their target groups of people / customers based on their context but they simply rely on the pre-determined application’s methods to approach them. These systems also suffer from information overload, often providing offers to a user that are neither valid nor interesting because they do not match his/her context. Moreover, these offering strategies are not interoperable among different systems. In this paper, we present the design and implementation of an innovative web-mapping context-aware system called “SPLIS” (Semantic Personalized Location Information System) that utilizes Semantic Web technologies for delivering group-targeted offers from POI owners to users/potential customers. The presented system a) adopts the schema.org ontology, b) uses RuleML-compatible rules to represent group-targeted POI offers, c) combines at run-time the above to match user context with suitable offers, and finally, d) visualizes offers in an intuitive way. The paper also reports on a user evaluation of the system. http%3A%2F%2Ftinyurl%2Ecom%2Fsplis%2Dlogin SPLIS+service http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Eeswa%2E2015%2E01%2E039 Rules Ontologies Semantic Web Location Based Services Context Points of Interest Group-Targeted Offers Integrating Multiple Immunogenetic Data Sources For Feature Extraction and Mining Mutation Patterns: The Case of Chronic Lymphocytic Leukemia Shared Mutations 2 Statistical Methods for Omics Data Integration and Analysis. Heraklion, Crete, Greece, November 10-12 2014 0 The aim of this work is to extract features and create high quality datasets through integration of multiple information resources for somatic hypermutation (SHM) analysis in the clonotypic immunoglobulin (IG) receptors of patients with Chronic Lymphocytic Leukemia (CLL). This can set the basis for an in-depth investigation of a series of as yet unanswered biological questions, through data mining analysis, which is clinically relevant given the great prognostic value of SHM in CLL (Damle et al, 1999). The virtue of the proposed approach is illustrated via the case of “towards analysis” which is our attempt to identify potential developmental transformation or movement of IG gene germlines towards other IG gene germlines through SHM. Kavakiotis_et_al_SMODIA14.pdf data integration feature extraction list aggregation mutation patterns somatic hypermutation SHM chronic lymphocytic leukemia CLL Improving Diversity in Image Search via Supervised Relevance Scoring 2 International Conference on Multimedia Retrieval ACM 2015 8 Results returned by commercial image search engines should include relevant and diversified depictions of queries in order to ensure good coverage of users' information needs. While relevance has drastically improved in recent years, diversity is still an open problem. In this paper we propose a reranking method that could be implemented on top of such engines in order to provide a better balance between relevance and diversity. Our method formulates the reranking problem as an optimization of a utility function that jointly considers relevance and diversity. Our main contribution is the replacement of the unsupervised definition of relevance that is commonly used in this formulation with a supervised classification model that strives to capture a query and application-specific notion of relevance. This model provides more accurate relevance scores that lead to significantly improved diversification performance. Furthermore, we propose a stacking-type ensemble learning approach that allows combining multiple features in a principled way when computing the relevance of an image. An empirical evaluation carried out on the datasets of the MediaEval 2013 and 2014 "Retrieving Diverse Social Images" (RDSI) benchmarks confirms the superior performance of the proposed method compared to other participating systems as well as a state-of-the-art, unsupervised reranking method. spyromitrosICMR2015.pdf Shanghai, China image retrieval diversity relevance feedback image classification image reranking The Tomaco Hybrid Matching Framework for SAWSDL Semantic Web Services 1 IEEE Transactions on Services Computing IEEE, IF: 1.985 (to appear) 2015 0 This work aims to advance Web Service retrieval, also known as Matching, in two directions. Firstly, it introduces a matching algorithm for SAWSDL, which adapts and extends known concepts with novel strategies. Effective logic-based and syntactic strategies are introduced and combined in a novel hybrid strategy, targeting an envisioned well-defined, real-world scenario for matching. The algorithm is evaluated in a universal environment for matching algorithms, SME2, in an objective, reproducible manner. Evaluation ranks Tomaco high amongst state of the art, especially for early recall levels (first in macro-averaging precision, up to 0.7 recall). Secondly, this work introduces the Tomaco web application, which aims to promote wide-spread adoption of Semantic Web Services while targeting the lack of user-friendly applications in this field, by integrating a variety of configurable matching algorithms proposed in this paper. It, finally, allows discovery of both existing and user-contributed service collections and ontologies, serving also as a service registry. thanosIEEETSC2015.pdf (to appear) http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1109%2FTSC%2E2015%2E2430328 Web Services Discovery Intelligent Web Services and Semantic Web Internet Reasoning Services Web-based Services A smart university platform for building energy monitoring and savings 1 Journal of Ambient Intelligence and Smart Environments IOS Press, IF 2014: 1.082 (to appear) 2015 0 This paper presents a novel, integrated platform for energy monitoring, management and savings in the context of a Smart Uni-versity Building. Namely, the proposed Smart IHU platform integrates an intelligent, rule-based agent that enforces savings, while a variety of applications offers user interaction with the system and the means for monitoring and management. The applica-tion layer is built over a common Web Service middleware, incorporating semantic interoperability. Monitoring applications visu-alize raw and aggregated sensor readings, such as building energy disaggregation, environmental measurements and data center efficiency. Extensive monitoring capabilities allow users to take immediate action and devise policies towards energy-savings. Such policies are, then, autonomously enforced by the intelligent, hybrid agent, which is capable of both deliberative (long-term) and reactive (immediate) actions. The agent is also integrated with the OpenADR standard for receiving provider instructions in future Smart Grids. A pilot deployment of the agent, with expert-formulated policies, has managed to reduce the total daily con-sumption of a typical university office by approximately 16%. thanosJAISE2015.pdf Smart grids sensor networks web services semantic web ambient intelligence Editorial of Special issue on Web Intelligence, Mining and Semantics 1 International Journal on Artificial Intelligence Tools World Scientific 24 (2) 2015 6 1502002 This special issue of the International Journal on Artificial Intelligence Tools (IJAIT) focused on Web Intelligence, Mining and Semantics (WIMS). Web intelligence (WI), as a research direction, has a broad agenda dealing with the issues that arise around the WWW phenomenon [1]. It is a converging research area bringing together several research communities, such as Databases (DB), Information Retrieval (IR), Semantic Technologies (ST), Artificial Intelligence (AI) to mention a few. To facilitate this convergence, the WIMS community started the conference series in 2011 on web intelligence, web mining and web semantics. For the Web to reach its full potential and gain intelligence, we need to enhance its services, make it more comprehensible, and increase its usability. The WIMS community is interested in novel data architectures and infrastructures especially suited to meet the challenges of the Web, namely the enormous volumes of data, their dynamic nature, their explicit or implicit semantics, their integrity and provenance. Web intelligence is about applying artificial intelligence and information technology techniques on the Web in order to create novel, adaptive and smarter web-based products, services and frameworks. Especially for WIMS’14, the focus is on either using some new AI techniques for providing general purpose web intelligence, such as semantic agent systems or nature-inspired models, or using AI techniques for more specialized web research and application areas. The facet of web mining and extracting information and knowledge from the Web within the WIMS landscape covers several interrelated and vibrant research directions providing enabling technologies for web intelligence. In particular, the methods and technologies for mining content spanning across different modalities (multimedia) and being very dynamic in its nature or constrained in access time (data streams) gain increased attention in the research communities and more demand in industries. The Semantic Web vision emerged from the confluence of several communities – artificial intelligence, hypertext, web developers – and so there are a number of ways to realize its motivation and goals. The Semantic Web is an approach to encoding and publishing information in ways that makes it easier for computers to understand, thus making the Web agent-friendly [2]. The Semantic Web offers an ambitious vision of an internet populated with intelligent agents and services able to exchange information, tasks and knowledge using simple protocols coupled with a rich knowledge representation language. This special issue focuses on intelligent approaches, supported by tools, to transform the World Wide Web into a global reasoning and semantics-driven computing machine. This issue includes the articles which are the extended versions of the selected best papers from the 4th International Conference on Web Intelligence, Mining and Semantics (WIMS’14), Thessaloniki, Greece held on 02-04 June 2014. http%3A%2F%2Fwims14%2Ecsd%2Eauth%2Egr%2F WIMS%2714 http%3A%2F%2Fwww%2Eworldscientific%2Ecom%2Fdoi%2Fabs%2F10%2E1142%2FS0218213015020029 Web Intelligence Web Mining Web Semantics On Discovering Relationships in Multi-Label Learning via Linked Open Data 2 4th Workshop on Knowledge Discovery and Data Mining Meets Linked Open Data (Know@LOD 2015) CEUR Workshop Proceedings Johanna Volker, Heiko Paulheim, Jens Lehmann, Vojtech Svatek Vol-1365 2015 6 In multi-label learning, each instance can be related with one or more binary target variables. Multi-label learning problems are commonly found in many applications, e.g. in text classification where a news article is possible to be both on politics and finance. The main motivation of multi-label learning algorithms is the exploitation of label dependencies in order to improve prediction accuracy. In this paper, we present ongoing work on a method that uses the linked open data cloud to detect relationships between labels, enriches the set of labels with new concepts which are super classes of two or more labels, trains a model on the enhanced training set and finally, makes predictions on the enhanced test set in order to improve the prediction accuracy of the initial labels. http%3A%2F%2Fceur%2Dws%2Eorg%2FVol%2D1365%2Fpaper4%2Epdf multi-label learning linked open data semantics WordNet Building a logic for a public administration service transformation algorithm 2 E-Democracy, Security, Privacy and Trust in a Digital World, 5th International Conference, E-Democracy 2013 Springer Communications in Computer and Information Science, Vol. 441 2014 0 73-79 This paper presents a rationale for establishing a Public Administration Service Transformation Algorithm. It introduces an abstract top level reusable model representing how Public Administration (PA) operates in providing services to the citizens, based on an input-output model. The approach adopted here is a goal-oriented one, placing the administrative act at the centre of PA’s operation, as act is the means of expressing PA’s will. In this way an algorithm which may identify malfunctions, propose services and conceptualize systems to remedy failings in service provision could be built. Using PA’s performance, both in effectiveness and efficiency to spot problems, and based on the improvement of these features, suggestions of services and systems could be made. Athens, Greece, Dec 5-6, 2013 http%3A%2F%2Flink%2Espringer%2Ecom%2Fchapter%2F10%2E1007%2F978%2D3%2D319%2D11710%2D2%5F7 Transformational government Public Administration operational model Public Administration effectiveness Public Administration efficiency On the Design of a Knowledge Base for Adverse Drug Event Prevention in Neonatal Intensive Care Units 2 poster presented at 5th Panhellenic Conference on Biomedical Technologies 2013 0 Athens, Greece, April 4-6, 2013 New Trends in Database and Information Systems II 4 Selected papers of the 18th East European Conference on Advances in Databases and Information Systems and Associated Satellite Events, ADBIS 2014, Proceedings II Springer Advances in Intelligent Systems and Computing, Vol. 312 2015 0 340 This volume contains the papers of 3 workshops and the doctoral consortium, which are organized in the framework of the 18th East-European Conference on Advances in Databases and Information Systems (ADBIS’2014). The 3rd International Workshop on GPUs in Databases (GID’2014) is devoted to subjects related to utilization of Graphics Processing Units in database environments. The use of GPUs in databases has not yet received enough attention from the database community. The intention of the GID workshop is to provide a discussion on popularizing the GPUs and providing a forum for discussion with respect to the GID’s research ideas and their potential to achieve high speedups in many database applications. The 3rd International Workshop on Ontologies Meet Advanced Information Systems (OAIS’2014) has a twofold objective to present: new and challenging issues in the contribution of ontologies for designing high quality information systems, and new research and technological developments which use ontologies all over the life cycle of information systems. The 1st International Workshop on Technologies for Quality Management in Challenging Applications (TQMCA’2014) focuses on quality management and its importance in new fields such as big data, crowd-sourcing, and stream databases. The Workshop has addressed the need to develop novel approaches and technologies, and to entirely integrate quality management into information system management. Ohrid, Macedonia, September 7-10, 2014 http%3A%2F%2Fwww%2Espringer%2Ecom%2Fus%2Fbook%2F9783319105178 Databases Information Systems Rule Technologies: Foundations, Tools, and Applications, Proceedings of 9th International Symposium, RuleML 2015 4 Lecture Notes in Computer Science Springer 9202 2015 0 474 The annual International Web Rule Symposium (RuleML) is an international conference on research, applications, languages and standards for rule technologies. RuleML is the leading conference to build bridges between academia and industry in the field of rules and its applications, especially as part of the semantic technology stack. It is devoted to rule-based programming and rule-based systems including production rules systems, logic programming rule engines, and business rule engines/business rule management systems; Semantic Web rule languages and rule standards (e.g., RuleML, SWRL, RIF, PRR, SBVR, DMN, CL, Prolog); rule-based event processing languages (EPLs) and technologies; and research on inference rules, transformation rules, decision rules, production rules, and ECA rules. This book constitutes the refereed proceedings of the 9th International RuleML Symposium, RuleML 2015, held in Berlin, Germany, in August 2015. The 25 full papers, 4 short papers, 2 full keynote papers, 2 invited research track overview papers, 1 invited paper, 1 invited abstracts presented were carefully reviewed and selected from 63 submissions. The papers cover the following topics: general RuleML track; complex event processing track, existential rules and datalog+/- track; legal rules and reasoning track; rule learning track; industry track. RuleML+2015 http%3A%2F%2F2015%2Eruleml%2Eorg%2F Berlin, Germany, August 2-5, 2015 http%3A%2F%2Flink%2Espringer%2Ecom%2Fbook%2F10%2E1007%252F978%2D3%2D319%2D21542%2D6 rules RuleML complex event processing existential rules Datalog legal rules and reasoning rule learning rule industry Joint Proceedings of the 9th International Rule Challenge, the Special Track on Rule-based Recommender Systems for the Web of Data, RuleML2015 Industry Track and the 5th RuleML Doctoral Consortium 4 CEUR Workshop Proceedings, ISSN 1613-0073 CEUR.org 1417 2015 0 The 9th International Web Rule Challenge at the 9th International Web Rule Symposium (RuleML 2015) in Berlin is a forum where new ways of using rule-based systems are presented and practical experiences about implementing these systems are reported. The 9th edition features a broad set of rule-based applications: toolkits for querying existential rule knowledge bases, UIs for rule modelling and rule editors, rule translators, agent-based systems as well as the RuleML 1.02 family of rule markup languages. As in the past four editions, the RuleML Symposium hosts the 5th Doctoral Consortium attracting Ph.D. researchers in the area of Web Rules from different backgrounds and encouraging an interdisciplinary research approach. This year there are five PhD papers introducing research into data specification and extraction, software bridges between Prolog and general Java development, rule induction techniques and integration between business processes and business rules. The doctoral students benefit from the dynamic and friendly setting at RuleML, as well as from fruitful interactions with their dedicated academic mentors and the attending researchers and industrial experts in the field, who can evaluate their research projects from both theoretical and practical perspectives. Following the previous successful RuleML2014 track on “Learning business Rules from Data”, this year the RuleML Challenge features “Rule-based recommender systems for the web of data” Challenge (RecSysRules’15). The aim of the challenge was to evaluate performance of rule learning algorithms applied on recommender problems and to track progress in the use of linked open data cloud for feature set extension in machine learning. There were three participating teams in the RecSysRules challenge and 37 submitted runs. Short papers describing two of the participating systems are included in the proceedings, along with a paper describing framework that can be used to create datasets for future editions of the challenge. A highlight of this year's event is the Industry Track introducing three papers describing research work in innovative companies, from rule-based applications in financial industry to systems that can generate large quantity of natural language text in different languages, as well as management applications for agricultural policies data. The work submitted by Jean-François Baget, Alain Gutierrez, Michel Leclère, Marie-Laure Mugnier, Swan Rocher and Clément Sipieter entitled “ Datalog+, RuleML and OWL 2: Formats and Translations for Existential Rules” has been awarded with the RuleML2015 Challenge Best Paper Award. The KTIML team submission to the RecSysRules’15 challenge entitled "Transformation and aggregation preprocessing for top-k recommendation GAP rules induction" has been awarded prize for the best recommender performance. Thanks to all authors, students, supervisors, referees, co-chairs, members of the program committee and the organizing team that made the RuleML2015 Symposium, 9th International Web Rule Challenge, and 5th Doctoral Consortium great successes. RuleML+2015 http%3A%2F%2F2015%2Eruleml%2Eorg%2F Berlin, Germany, August 2-5, 2015 http%3A%2F%2Fceur%2Dws%2Eorg%2FVol%2D1417%2F rules RuleML Rule Modeling Rule Authoring Rule Engines Rule Applications Rule-based Recommender Systems rule industry Algorithms for Electric Vehicle Scheduling in Mobility-on-Demand Schemes 2 2015 IEEE 18th International Conference on Intelligent Transportation Systems IEEE (accepted for presentation) 2015 6 We study a setting where electric vehicles (EVs) can be hired to drive from pick-up to drop-off points in a mobility-on-demand (MoD) scheme. Each point in the MoD scheme is equipped with a battery swap facility that helps cope with the EVs’ limited range, while the goal of the system is to maximise the number of customers that are serviced. Thus, we first model and solve this problem optimally using Mixed-Integer Programming (MIP) techniques and show that the solution scales up to medium sized problems. Given this, we develop a greedy approach that is shown to output solutions that are close to the optimal and can scale to thousands of consumer requests and EVs. Both algorithms are evaluated in a setting using data of actual locations of shared vehicle pick-up and dropoff stations in Washington DC, USA and the greedy algorithm is shown to be on average 90% of the optimal in terms of average task completion. sharedEVsITS.pdf Las Palmas de Gran Canaria, Spain, 15-18 Sep 2015 http%3A%2F%2Fwww%2Eitsc2015%2Eorg%2F Electric Vehicles Energy Efficiency Smart Mobility Transportation Electrification An Ontology-based Decision Support Tool for Optimizing Domestic Solar Hot Water System Selection 1 Journal of Cleaner Production Elsevier (accepted) 2015 31 In an effort to tackle climate change most countries utilize renewable energy sources. This is more pronounced in the building sector, which is currently one of the major consumers of energy, mostly in the form of heat. In order to further promote the use of domestic solar hot water systems in buildings, an ontology-based decision support tool has been developed and is presented in this paper. The proposed tool aids non-technical consumers to select a domestic solar hot water system tailored to their needs, containing up-to-date information on its components and interrelationships, installation costs etc., in the form of an ontology formulated in OWL (Web Ontology Language). The optimum system configurations are computed based on various specific parameters, such as number of occupants, daily hot water requirements and house location. The backbone of the proposed system is an ontology that represents the application domain and contains information regarding the various domestic solar hot water system components along with their interrelationships. Ontologies are a rapidly evolving knowledge representation paradigm that offers various advantages and, when deployed specifically in the domestic solar hot water systems domain, deliver improved representation, sharing and re-use of the relevant information. As a conclusion, this paper presents an ontology-driven decision support system for facilitating the selection of domestic solar hot water system, which delivers certain advantages, such as sustainability of the decision support system itself, due to its open and interoperable knowledge-base, and its adaptability/flexibility in decision making policies, due to is semantic (ontological) nature. OntologyDSHWS.pdf http%3A%2F%2Frad%2Eihu%2Eedu%2Egr%2FDSHWS%2FDSHWS%2Eowl Ontology http%3A%2F%2Fdx%2Edoi%2Eorg%2F10%2E1016%2Fj%2Ejclepro%2E2015%2E08%2E088 Sustainable development Ontology-based selection DSHWS building integration DISARM: A Social Distributed Agent Reputation Model based on Defeasible Logic 1 Journal of Systems and Software Elsevier accepted for publication 2016 45 Agents act in open and thus risky environments with limited or no human intervention. Making the appropriate decision about who to trust in order to interact with is not only necessary but it is also a challenging process. To this end, trust and reputation models, based on interaction trust or witness reputation, have been proposed. Yet, they are often faced with skepticism since they usually presuppose the use of a centralized authority, the trustworthiness and robustness of which may be questioned. Distributed models, on the other hand, are more complex but they are more suitable for personalized estimations based on each agent’s interests and preferences. Furthermore, distributed approaches allow the study of a really challenging aspect of multi-agent systems, that of social relations among agents. To this end, this article proposes DISARM, a novel distributed reputation model. DISARM treats Multi-agent Systems as social networks, enabling agents to establish and maintain relationships, limiting the disadvantages of the common distributed approaches. Additionally, it is based on defeasible logic, modeling the way intelligent agents, like humans, draw reasonable conclusions from incomplete and possibly conflicting (thus inconclusive) information. Finally, we provide an evaluation that illustrates the usability of the proposed model. DISARM-JSS-final-revised.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald +%09EMERALD Multi-agent Systems Agent Reputation Distributed Trust Management Logic-Based Approach Defeasible Reasoning Semantic Web A Policy-based B2C e-Contract Management Workflow Methodology Using Semantic Web Agents 1 Artificial Intelligence and Law Springer accepted for publication 2016 25 Since e-Commerce has become a discipline, e-Contracts are acknowledged as the tools that will assure the safety and robustness of the transactions. A typical e-Contract is a binding agreement between parties that creates relations and obligations. It consists of clauses that address specific tasks of the overall procedure which can be represented as workflows. Similarly to e-Contracts, Intelligent Agents manage a private policy, a set of rules representing requirements, obligations and restrictions, additionally to personal data that meet their user’s interests. In this context, this study aims at proposing a policy-based e-Contract workflow management methodology that can be used by semantic web agents, since agents benefit from Semantic Web technologies for data and policy exchanges, such as RDF and RuleML that maximize interoperability among parties. Furthermore, this study presents the integration of the above methodology into a multi-agent knowledge-based framework in order to deal with issues related to rules exchange where no common syntax is used, since this framework provides reasoning services that assist agents in interpreting the exchanged policies. Finally, a B2C e-Commerce scenario is presented that demonstrates the added value of the approach. kravari-et-al-AI-Law.pdf http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fsystems%2Femerald EMERALD Semantic Web Intelligent Agents e-Contracts Workflows Policies Defeasible Reasoning Logic Programming Techniques: The Prolog language (in Greek - Τεχνικές Λογικού Προγραμματισμού: Η Γλώσσα Prolog) 4 [ebook], ISBN 978-960-603-246-2 Hellenic Academic Libraries Link Available Online at: http://hdl.handle.net/11419/777 2016 327 Ο Λογικός Προγραμματισμός (ΛΠ) ανήκει στις πλέον ενδιαφέρουσες σχολές προγραμματισμού, σημαντικά διαφορετική από τις "κλασικές" σχολές του προστακτικού και του αντικειμενοστραφούς προγραμματισμού. Η χρήση της Μαθηματικής Λογικής ως εργαλείο αφαίρεσης για περιγραφή υπολογισμών και η εκμετάλλευση των αποδεικτικών διαδικασιών της, οδηγούν σε συμπαγή προγράμματα τα οποία βρίσκουν εφαρμογές σε πολύπλοκα και ενδιαφέροντα πεδία, όπως είναι η Τεχνητή Νοημοσύνη και το Σημασιολογίκο Ιστό. Όμως, αυτή η υψηλού επιπέδου προσέγγιση στον προγραμματισμό δημιουργεί προβλήματα σε εκείνους που πρώτη φορά έρχονται σε επαφή με τον ΛΠ , αφενός γιατί απαιτεί να υιοθετήσουν μια δηλωτική προσέγγιση στην ανάπτυξη προγραμμάτων "ξένη" προς αυτή των κυρίαρχων γλωσσών, και αφετέρου, γιατί απαιτείται η εκμάθηση τεχνικών αναπαράστασης και επίλυσης προβλημάτων που αν και είναι γενικά εφαρμόσιμες στον προγραμματισμό, δεν χρησιμοποιούνται συχνά στις υπόλοιπες σχολές, με κλασικό παράδειγμα την αναδρομή. Το παρόν βιβλίο φιλοδοξεί να καλύψει τις παραπάνω απαιτήσεις και να αποτελέσει το βασικό σύγγραμμα για οποιονδήποτε επιθυμεί να εντρυφήσει στην τέχνη και στις τεχνικές του ΛΠ. Έχοντας ως όχημα τον κυριότερο εκπρόσωπο της σχολής ΛΠ, τη γλώσσα Prolog, το βιβλίο έχει στόχο: (α) να αναφέρει σύντομα τα θεωρητικά θεμέλια του ΛΠ, δηλαδή την Κατηγορηματική Λογική Πρώτης Τάξης και την αρχή της ανάλυσης, (β) να παρουσιάσει σε βάθος την γλώσσα προγραμματισμού Prolog, τα διαθέσιμα κατηγορήματα και πως αυτά εντασσόμενα σε τεχνικές ΛΠ προγραμματισμού αποτελούν ισχυρά εργαλεία για την επίλυση προβλημάτων, (γ) να παρουσιάσει εφαρμογές όπως είναι η ανάπτυξη ευφυών συστημάτων στις οποίες ο Λογικός Προγραμματισμός προσφέρει σημαντικά πλεονεκτήματα, και τέλος, (δ) να αναπτύξει προγραμματιστικές δεξιότητες του αναγνώστη, όπως είναι η αναδρομή και η ιεραρχική ανάπτυξη προγραμμάτων, εφαρμόσιμες σε όλες τις σχολές προγραμματισμού. http%3A%2F%2Fusers%2Euom%2Egr%2F%7Eiliass%2FLPTechniques%2F %C9%F3%F4%FC%F4%EF%F0%EF%F2+%E2%E9%E2%EB%DF%EF%F5 Athens http%3A%2F%2Fhdl%2Ehandle%2Enet%2F11419%2F777 ΛΟΓΙΚΟΣ ΠΡΟΓΡΑΜΜΑΤΙΣΜΟΣ ΛΟΓΙΚΗ ΤΕΧΝΙΚΕΣ ΠΡΟΓΡΑΜΜΑΤΙΣΜΟΥ PROLOG ΕΠΕΞΕΡΓΑΣΙΑ ΣΥΜΒΟΛΩΝ ΑΝΑΠΑΡΑΣΤΑΣΗ ΓΝΩΣΗΣ ΚΑΙ ΣΥΛΛΟΓΙΣΤΙΚΗ Multi-Target Regression via Input Space Expansion: Treating Targets as Inputs 1 Machine Learning Springer Johannes Furnkranz 2016 44 In many practical applications of supervised learning the task involves the prediction of multiple target variables from a common set of input variables. When the prediction targets are binary the task is called multi-label classification, while when the targets are continuous the task is called multi-target regression. In both tasks, target variables often exhibit statistical dependencies and exploiting them in order to improve predictive accuracy is a core challenge. A family of multi-label classification methods address this challenge by building a separate model for each target on an expanded input space where other targets are treated as additional input variables. Despite the success of these methods in the multi-label classification domain, their applicability and effectiveness in multi-target regression has not been studied until now. In this paper, we introduce two new methods for multi-target regression, called stacked single-target and ensemble of regressor chains, by adapting two popular multi-label classification methods of this family. Furthermore, we highlight an inherent problem of these methods—a discrepancy of the values of the additional input variables between training and prediction—and develop extensions that use out-of-sample estimates of the target variables during training in order to tackle this problem. The results of an extensive experimental evaluation carried out on a large and diverse collection of datasets show that, when the discrepancy is appropriately mitigated, the proposed methods attain consistent improvements over the independent regressions baseline. Moreover, two versions of Ensemble of Regression Chains perform significantly better than four state-of-the-art methods including regularization-based multi-task learning methods and a multi-objective random forest approach. http%3A%2F%2Farxiv%2Eorg%2Fabs%2F1211%2E6581 Arxiv+version+of+the+paper%2E http%3A%2F%2Flink%2Espringer%2Ecom%2Farticle%2F10%2E1007%252Fs10994%2D016%2D5546%2Dz Multi-target regression Multi-label classification Stacking Chaining F Kokkoras http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Ffotis%2Ehtml kokkoras%40csd%2Eauth%2Egr I Vlahavas http%3A%2F%2Fplase%2Ecsd%2Eauth%2Egr%2Fvlahavas%2F vlahavas%40csd%2Eauth%2Egr G Tsoumakas http%3A%2F%2Fmlkd%2Ecsd%2Eauth%2Egr%2Fgreg%2Ehtml greg%40csd%2Eauth%2Egr C Berberidis http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Fchristos%2Ehtml berber%40csd%2Eauth%2Egr D Vrakas http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Fdvrakas%2Ehtml dvrakas%40csd%2Eauth%2Egr N Bassiliades http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Fnbassili%2F nbassili%40csd%2Eauth%2Egr I Sakellariou http%3A%2F%2Flpis%2Ecsd%2Eauth%2Egr%2Fpeople%2Felias%2Ehtml iliass%40csd%2Eauth%2Egr I Refanidis http%3A%2F%2Fmacedonia%2Euom%2Egr%2F%257Eyrefanid%2F yrefanid%40uom%2Egr K Paparizos W G Aref M Atallah A K Elmagarmid D Sampson V Papaioannou I Futo Z Pasztor J Szeredi A G Walid H Jiang E Houstis S Demetriadis G Palaigeorgiou A Barbas A Molohides L Tsoukalas D Psillos E Hatzikraniotis I Lefkos G Bisdikian V Dimitroula S Dimitrakos P Argirakis K Korobilis I Stamelos P Katsaros A Tsoukias A Pombortsis F Milcent P Linardis P Kefalas http%3A%2F%2Fwww%2Ecity%2Eacademic%2Egr%2Fmaterial%2Facademic%5Fstaff%2Fcomputer%5Fscience%2Fkefalas%2Findex%2Ehtml C Halatsis M Karamanlidou P Tsarchopoulos S Gregory A Likartsis N Bourbakis D Goldman R Fermatt K Mamzoridi M Pirpasopoulos P Gray E Tzelalis C Maciazek S M Embury T J Reynolds D Stamatis C Konidaris L Gallos E Petridou I Nikolaidis D Katsoulis et al. L Angelis D Anagnostopoulos G Tzanis http%3A%2F%2Fusers%2Eauth%2Egr%2F%7Egtzanis gtzanis%40csd%2Eauth%2Egr I Katakis http%3A%2F%2Fusers%2Eauth%2Egr%2Fkatak katak%40csd%2Eauth%2Egr G Meditskos http%3A%2F%2Fusers%2Eauth%2Egr%2Fgmeditsk gmeditsk%40csd%2Eauth%2Egr G Antoniou T Skylogiannis I Partalas partalas%40csd%2Eauth%2Egr A Bikakis G Governatori S Konias N Maglaveras A Alexandridou E Kontopoulos http%3A%2F%2Fusers%2Eauth%2Egr%2Fskontopo skontopo%40csd%2Eauth%2Egr I Chouvarda S Diplaris P Mitkas E V Hatzikos L Anastasakis G Eleftherakis S Bibi M Doerr E Banos K Kravari L Asmanis O Hatzi A Symeonidis A Papakonstantinou A Nestoras I Feneris ifeneris%40csd%2Eauth%2Egr G Ηondris M Golias D Angelides S Marnas G Paliouras G Hondris G Gogou P D Bamidis K Trohidis G Kalliris A Seridou E Spyromitros-Xioufis http%3A%2F%2Fusers%2Eauth%2Egr%2Fespyromi espyromi%40csd%2Eauth%2Egr I Kavakiotis http%3A%2F%2Fwww%2Eikavakiotis%2Egr ikavak%40csd%2Eauth%2Egr I Avguleas K Gkirtzou S Triantafilou E Kargioti A Athanassiades K Tzevanidis A Dimou V Mezaris I Kompatsiaris I Savvas J Hatonen E Fournou G Valkanas P Natsiavas G Nasierding A Kouzani E Loza Mencia S Park J Furnkrnaz I P Antoniades G Koutitas K Mokos V Vassiliadis M Perrotin A Paschke M Befa mmpefa%40csd%2Eauth%2Egr T G Stavropoulos http%3A%2F%2Fusers%2Eauth%2Egr%2Fathstavr athstavr%40csd%2Eauth%2Egr L Iliadis I Maglogiannis M Bramer A Tsioliaridou G Koutitas M Ioannou G Sakkas F Markatopoulou J Vilcek E Ziaka E Lampridou K Ntonas M Spiliopoulou M Spiliopoulou D L Settas M Nikolaidou V Katos J Dix C Malliarakis G Kastori T Osmun H Boley K Papatheodorou A Fachantidis http%3A%2F%2Fusers%2Eauth%2Egr%2F%7Eafa%2F afa%40csd%2Eauth%2Egr K Sechidis A Arvanitidis E Stachtiari K Chatzidimitriou G Boutsioukis M Taylor D Vlachava T Zetta S Papadopoulos S Papadopoulos A Tsadiras E Myrgioti A Miliou C Badica I Viktoratos E S Rigas E Pimenidis A B Sideridis G Papadopoulos C Liaskos C Spyropoulos (eds.) K Gottis T Dergiades N Alechina M Dastani M de Vos B Logan S Mera A Morris-Martin F Schapachnik Y Xenidis E Loukogeorgaki A Taflanidis S Arnaouti K Lazou M Farini V Koutkias V Drossou S Ilie C Papavasileiou S D Ramchurn J Argyriou M Laliotis N Markantonatos N Spolaor D Corney C Martin A Goker L Aiello B Thomee W Groves G Kordopatis--Zilos s s F Briggs Yonghong Huang R Raich K Eftaxias Zhong Lei W Cukierski S F Hadley A Hadley M Betts X Z Fern J Irvine L Neal A Thomas G Fodor Hong Wei Ng Thi Ngoc Tho Nguyen H Huttunen P Ruusuvuori T Manninen A Diment T Virtanen J Marzat J Defretin D Callender C Hurlburt K Larrey M Milakov A Di Nuovo A Cangelosi A Triantafyllidis P Samaras A Voulgaridis N Karaiskou E Konstantinidis N Charboni L Torrey P Bailis Y Zhan S Andreadis J Argyriou D Sfounis R Akerkar J Davies V Ermolayev A Arvanitidis A Serafi A Vakali A Vrekou I Papanikolaou D Dimitriadis M Monard H Lee N Kotsiopoulos A Xochelli A Agathangelidis K Stamatopoulos A Hadzidimitriou A L Ginsca A Popescu E Papagiannopoulou M Ivanovic M Kon-Popovska Y Manolopoulos T Palpanas G Trajcevski A Vakali G Gottlob F Sadri D Roman P Fodor A Giurca T Kliegr G J Nalepa M Palmirani M Proctor N Stojanovic G Martinopoulos D Lazarou A Seitaridis