IAT-2001 and WI-2001 Joint Keynote Speakers

Benjamin Wah (2001 IEEE Computer Society President)University of Illinois at Urbana-Champaign
Edward A. Feigenbaum (Turing Award Winner)Stanford University



IAT-2001 Invited Speakers

Toyoaki NishidaUniversity of Tokyo, Japan
Zbigniew W. RasUniversity of North Carolina, USA
Andrzej SkowronWarsaw University, Poland
Katia SycaraCarnegie Mellon University, USA



Intelligent Agents for Market-Trend Prediction

Benjamin Wah
Department of Electrical and Computer Engineering
and the Coordinated Science Laboratory
University of Illinois at Urbana-Champaign
Urbana, IL 61801, USA
http://manip.crhc.uiuc.edu/
(2001 IEEE Computer Society President)


In this presentation we discuss the role of intelligent agents in market-trend predictions. Market-trend data, such as stock- market data, are characterized by non-stationary time series that may depend on non-numeric and non-quantifiable measures. The prediction of market trends, therefore, should consist of prediction of non-stationary time series and the abstraction and integration of non-numeric information in prediction. In this talk, we survey various prediction techniques for and mining of market-trend data. We propose to use intelligent agents in the abstraction of non-numeric information, the decomposition of non- stationary time series into multiple stationary time series, and the prediction of trends using artificial neural networks.
Finally, we illustrate our techniques in predicting stock-market data.


Benjamin W. Wah received his Ph.D. degree in computer science from the University of California, Berkeley, CA, in 1979. He is currently the Robert T. Chien Professor of Electrical and Computer Engineering, and a Research Professor of the Coordinated Science Laboratory and the Beckman Institute at the University of Illinois, Urbana- Champaign, Urbana, IL. Previously, he had served on the faculty of Purdue University (1979-85), as a Program Director at the National Science Foundation (1988-89), as Fujitsu Visiting Chair Professor of Intelligence Engineering, University of Tokyo (1992), and as McKay Visiting Professor of Electrical Engineering and Computer Science, University of California, Berkeley (1994). In 1989, he was awarded a University Scholar of the University of Illinois; in 1998, he received the IEEE Computer Society Technical Achievement Award; and in 2000, the IEEE Millennium Medal.

Wah's current research interests are in nonlinear search and optimization, knowledge engineering, multimedia signal processing, and parallel and distributed processing.

Wah was the Editor-in-Chief of the IEEE Transactions on Knowledge and Data Engineering between 1993-1996, and is the Honorary Editor-in-Chief of Knowledge and Information Systems. He currently serves on the editorial boards of Information Sciences, International Journal on Artificial Intelligence Tools, Journal of VLSI Signal Processing, Parallel Algorithms and Applications, and Neural Processing Letters. He had chaired a number of international conferences and was the International Program Committee Chair of the IFIP World Congress in 2000. He has served in the IEEE in various capacities and is currently the 2001 President of the IEEE Computer Society. He is a Fellow of the IEEE and the Society for Design and Process Science.



Spinning the Semantic Web

Edward A. Feigenbaum
(Turing Award Winner), Stanford University




Edward Feigenbaum is one of the pioneers of Artificial Intelligence research and its applications. For this work, he was awarded the ACM Turing Award in 1995, the highest award given for research in Computer Science.

He received his B.S. degree in Electrical Engineering in 1956 and his Ph.D. in 1960, both from Carnegie Mellon University. After a Fulbright Scholarship year in the UK, he taught at University of California, Berkeley until moving to Stanford University in 1965.

He has been Chairman of the Computer Science Department and Director of the Computer Center at Stanford University. In 1965 he founded the well-known laboratory now known as the Stanford Knowledge Systems Laboratory. For many years, he was Co-Principal Investigator of the NIH-sponsored national computer facility for applications of Artificial Intelligence to Medicine and Biology known as SUMEX-AIM.

He is the Past President of the American Association for Artificial Intelligence. His public service includes: NSF Computer Science Advisory Board; ISAT, an ARPA study committee for Information Science and Technology; the NRC's Computer Science and Technology Board. He has been a member of the Board of Regents of the National Library of Medicine, and Air Force Scientific Advisory Board. From 1994-97 he served at the Pentagon as Chief Scientist of the Air Force.

He was co-author of: the encyclopedia, The Handbook of Artificial Intelligence; the early book, Computers and Thought; and the book Applications of Artificial Intelligence in Organic Chemistry: The DENDRAL Program. He was also the founding editor of the McGraw-Hill Computer Science Series. He is co-author of the books: The Fifth Generation: Artificial Intelligence and Japan's Computer Challenge to the World; and The Rise of the Expert Company (about corporate successes in the use of expert systems).

Dr. Feigenbaum is a co-founder of three start-up firms in applied artificial intelligence, IntelliCorp, Teknowledge and Design Power Inc. He also was a member of the Board of Directors of Sperry Corporation. He is a member of the Advisory Council of the Japanese Kansai Silicon Valley Venture Forum.

He was elected to the National Academy of Engineering in 1986 and the American Academy of Arts and Sciences in 1991. He was selected for the Productivity Hall of Fame of the Republic of Singapore. He is an elected Fellow of the American Association for Artificial Intelligence; the American College of Medical Informatics; and the American Institute of Medical and Biological Engineering; and the American Association for the Advancement of Science. He is the first recipient of the Feigenbaum Medal, an award established in his honor by the World Congress of Expert Systems. He received the U.S. Air Force Exceptional Civilian Service Award in 1997.



Social Intelligence Design for Knowledge Creating Communities

Toyoaki Nishida
University of Tokyo, Japan


Communities play an important role in knowledge creation by providing people with opportunities to continually learn from others, find partners to collaborate with, and demonstrate the significance of their disciplines. In education or business, it is relatively easy to find typical examples of knowledge creating communities for sharing and exchanging specialized knowledge among knowledge workers. In other domains such as NPO or local communities, people are naturally practicing mutual learning and invaluable knowledge is built as a result, even if knowledge creation is not deemed a primary goal of the community.

In this paper, I present an interdisciplinary approach to augmenting the community knowledge creating process by integrating insights from social psychology, cognitive psychology, and advanced information technology. I emphasize the role of conversations and stories as a means of establishing a common background in a community.

I describe several systems that primarily use the conversational modality to mediate community communication. EgoChat allows the user to make conversation with virtualized egos responding on behalf of other users. It allows the user to take an initiative by interrupting the conversation and changing its flow. VoiceCafe allows artifacts to make conversation with people or other artifacts. It stimulates creative thinking by bringing about utterances from the physical object's point of view, which might be strikingly different from humans' view.

These engineering approaches should be tightly coupled with sociological and cognitive approaches, to predict and assess the effects of community communication mediation systems on the human society. I discuss issues on designing a constructive framework of interaction for achieving practical goals without being caught by known pathological pitfalls of group interactions.


Toyoaki Nishida is a professor of Department of Information and Communication Engineering, School of Engineering, The University of Tokyo. He received the B.E., the M.E., and the Doctor of Engineering degrees from Kyoto University in 1977, 1979, and 1984 respectively. His research centers on artificial intelligence in general. His current research focuses on community computing and support systems, including knowledge sharing, knowledge media, and the agent technology. He has been leading the Breakthrough 21 Nishida Project, sponsored by Ministry of Posts and Telecommunications, Japan, aiming at understanding and assisting networked communities. He is an area editor (intelligent systems) of New Generation Computing and an editor of Autonomous Agents and Multiagent Systems.



Query Answering based on Distributed Knowledge Mining

Zbigniew W. Ras
University of North Carolina, USA




Zbigniew W. Ras received his Ph.D. degree from University of Warsaw, Poland in 1973. Currently, he is a Professor of Computer Science in the College of Information Technology at the University of North Carolina, Charlotte and an Overseas Associate in the Institute of Computer Science at the Polish Academy of Sciences. Also, he is a member of the Senat at the Polish-Japanese Institute of Information Technology in Warsaw, Poland. Previously, he had served on the faculty of University of Tennessee, University of Florida, University of Warsaw, Columbia University. Also, he was Summer Associate at Lockheed Missiles and Space Research Laboratory at Palo Alto and a visiting professor at several universities in western Europe including University of Bonn and Linkoping University. He is Editor-in-Chief of the Journal of Intelligent Information Systems (Kluwer), Deputy Editor-in-Chief of Fundamenta Informaticae Journal (IOS Press) and the General Chair of ISMIS symposium. He is the author/co-author of about 100 papers and editor/co-editor of 23 books in intelligent systems area.


Approximate Reasoning in Distributed Environment of Agents

Andrzej Skowron
Institute of Mathematics
Warsaw University, Poland
e-mail: skowron@mimuw.edu.pl

Information processing in intelligent systems needs new computing paradigms. One of the recently emerging ones is Granular Computing (Computing with Words). Granular computations are performed on information granules representing vague and complex concepts delivered by agents, involved, among others, in knowledge representation, communication with other agents, and reasoning.

Specifications of complex tasks are often formulated in words, phrases or more complex texts of a natural language. Hence, the following main problem arises: if and how can an information granule, in a sense, sufficiently close to the target information granule Gt representing the task specification, be constructed from input information granules (e.g. representing sensor measurements).

One of the important problems is related to the construction of an interface allowing knowledge acquisition agents (KA-agents) to acquire knowledge from customer-agents (CA-agents), who specify the task. The aim is to induce a satisfactory approximation Gk of the target information granule Gt in the language of KA-agents, i.e., an information granule Gk sufficiently close to the target information granule Gt. Hence, some tools for expressing inclusion and proximity of information granules measured by the degree of proximity are needed. For this purpose we use rough sets and rough mereology. The interface construction should be supported by background knowledge (in particular by ontology of concepts) and experimental data.

An information granule G sufficiently close to the information granule Gk delivered by KA-agents should be constructed from input information granules (representing, e.g., sensor measurements). In the search for the granule G relevant operations and inclusion (closeness) measures on information granules should be discovered and used. The granule G is constructed from basic components defined by information granule calculi. Any such calculus consists of components like

(1) elementary input information granules;
(2) operations on information granules;
(3) relations of inclusion and proximity measured by the proximity degree between information granules;
(4) schemes of information granule construction (IG-schemes) which can be treated as approximate reasoning schemes on information granules.

Elementary information granules together with inclusion and proximity relations between such granules are primitive constructs in granule construction. Higher level constructs, like information granules and inclusion (closeness) relations related to them, can be defined from the previously constructed lower level constructs using relevant operations.

One kind of important operations on information granules are the fusion operations, based, e.g., on negotiation schemes for resolving conflicts between agents, delivering arguments of operations. More complex operations, called transducers, are defined by robust IG-schemes. Such schemes are obtained by approximate reasoning rules and methods for their composition, dependent on available data and background knowledge. The robustness of IG-schemes means that the closeness (inclusion) of constructed granules is preserved in a satisfactory degree under small deviations of input granules (or operation parameters used for the granule construction). The robustness of the target construction can be deduced from the robustness of their sub-constructions, if some constraints for composition are satisfied. The robust IG-schemes should rather be extracted from experimental (e.g. sensory) data or/and background knowledge than obtained by means of classical deduction mechanisms.

The IG-schemes are parameterized. Relevant information granules are constructed by tuning IG-scheme parameters. There are several kinds of parameters tuned in the process of searching for relevant information granules. Some of them are come from approximation spaces of agents and allow to obtain a proper generalization degree of the granule constructed in the inductive reasoning. The other ones are related to agent teams and are used to tune measures of inclusion (closeness) between information granules and to tune propagation mechanisms of the inclusion (closeness) degrees along the IG-schemes. The IG-schemes can be treated as higher order neural networks performing operations on information granules instead of numbers. One of the main problems of a new Rough-Neuro Computing paradigm is to develop methods for inducing parameterized IG-schemes and methods for tuning their parameters.

The presented approach is based on the foundations of a calculus on information granules developed on the basis of rough set and rough mereological approaches. Its aim is to create a methodology and tools for solving a wide class of complex problems, from the identification of road traffic situations by an unmanned aerial vehicle to problems of text data mining in the Internet.


Andrzej Skowron holds a Ph.D. degree in Mathematical Foundations of Computer Science from the University of Warsaw in Poland, Doctor of Science (Habilitation) degree in Mathematical Foundations of Computer Science from the University of Warsaw in Poland. In 1991 he received the Scientific Title of Professor.

Andrzej Skowron is Full Professor in the Faculty of Mathematics, Computer Science and Mechanics at Warsaw University. He is the head of the Logic Section in the Institute of Mathematics. He is also connected with the Polish Academy of Sciences, Institute of Computer Science. From 1988 to 1990, ha was the Deputy Dean of the Faculty of Mathematics, Computer Science and Mechanics at Warsaw University. From 1994 to 1999, he was also the Head of the Senate in the Polish-Japanese Institute of Information Technology.

Prof. Skowron is the author and co-author of about 200 publications, including several edited books. His areas of expertise include soft computing methods and applications, reasoning with incomplete information, approximate reasoning, rough sets, rough mereology, granular computing, synthesis and analysis of complex objects, intelligent agents, knowledge discovery systems, and advanced data mining techniques, decision support systems, adaptive systems. Currently, his research is focused on rough set theory and its applications.

Since 1995 he is the Editor-in-Chief of Fundamenta Informaticae journal and a member of Editorial Boards of several others journals including Knowledge Discovery and Data Mining and Studia Logica.

Prof. Skowron was the President of the International Rough Set Society from 1996 to 2000, and has served or is currently serving on the program committees of over 40 international conferences and workshops, including ISMIS '97-99 (program chair), RSCTC '98-00 (program chair), and RSFDGrC '99 (program chair). He has delivered several invited talks at international conferences including a plenary talk at the 16th IFIP World Computer Congress (Beijing, 2000).

Throughout his career Andrzej Skowron has won many awards for his achievements, including awards from the Ministry of Science, the Rector of Warsaw University, the Ministry of Education, Mazur's Award of the Polish Mathematical Society, and Janiszewski's Award of the Polish Mathematical Society.



Multi-agent Infrastructure for Agent Interoperation in Open Computational Environments

Katia Sycara
The Robotics Institute
School of Computer Science
Carnegie Mellon University, USA
e-mail: katia@cs.cmu.edu
http://www.cs.cmu.edu/~softagents/

Multi-agent Systems (MASs) are becoming increasingly important: as a scientific discipline, as a software engineering paradigm, and as a commercially viable and innovative technology. Despite the considerable research that has gone into the formation of theories, scientific principles and guidelines for MASs, there is relatively little experience with the building, fielding and routine use of them.. To achieve this goal, a stable, widely used, widely accessible and extensible MAS infrastructure is crucial. Various standards bodies (e.g. FIPA) are attempting to define standards for various aspects of MAS infrastructure, such as Agent Communications Languages.. However, there is no coherent account of what constitutes a MAS infrastructure, what functionality it supports, what characteristics it should have in order to enable various value-added abilities, and what its possible relation with and requirements it may impose on the design and structure of single agents.

Another equally important aspect of MAS that operate in an open world like the Internet, where communication links, informaiton sources, services and agents can appear and disappear dynamically is the issuse of discovery and interoperation of agents. White pages and yellow page registries of companies, for example have been proposed and implemented (e.g. Yahoo business categories) for human understndability. We have coined the term middle agents [Decker&Sycara, IJCAI97] to describe various agent intermediaries that can act as brokers and discovery services for agents on the Internet. These domain independent intermediaries facilitate the finding and matching of agents and services with desirable functionalities (e.g. an agent that finds weather information).

Such intermediaries start having appeal for industry. For example, industrial organizations (e.g. SUN) are developing and making accessible software that could constitute a part of a MAS infrastructure, such as JINI for service discovery. Protocols such as UDDI (www.uddi.org), SOAP (www.soapware.org) and languages such as WSDL (www.wsdl.org) ebXML and e-speak are receiving increased visibility.

In this talk, we will present a model of MAS infrastructure, and our implemented RETSINA system that is an example of the general infrastructure model. We will also discuss various applications that we have implemented using RETSINA.


Dr. Sycara is a Research Professor in the School of Computer Science at Carnegie Mellon University. She is also the Director of the Advanced Technology Laboratory. She holds a B.S in Applied Mathematics from Brown University, M.S. in Electrical Engineering from the University of Wisconcin and PhD in Computer Science from Georgia Institute of Technology. She has given numerous invited talks, and has authored or co-authored more than 150 technical papers dealing with Multiagent Systems, Software Agents, Negotiation, Case-Based Reasoning and the application of these techniques to manufacturing, crisis action planning, scheduling and financial planning.

She has served as the General Chair of the Second International Conference on Autonomous Agents (Agents 98), as the Chair of the Steering Committee of the Agents Conference (1999-2001) and a member of the AAAI Executive Council (1997-99). She is a founding member and member of the Board of Directors of the International Foundation of Multiagent Systems (IFMAS), and is the Scholarship Chair of the American Association for Artificial Intelligence. She is a Founding Editor-in-Chief of the journal gAutonomous Agents and Multiagent Systemsh; an Editor-in-Chief of the Springer Series on Agents; on the Editorial Board of the Kluwer book series on gMultiagent Systems, Artificial Societies and Simulated Organizationsh; the Area Editor for AI and Management Science of the journal gGroup Decision and Negotiationh and on the editorial board of ETAI journal on the Semantic Web. She has served on the Editorial Board of gIEEE Intelligent Systems and their Applicationsh, gAI in Engineeringh and gConcurrent Engineering: Research and Applicationsh She is a member of AAAI, ACM, and Senior Member of IEEE.