Intelligent Systems: An Assesment of the Past and the Prospects for the Future

Okyay Kaynak, Professor & PhD, IEEE Fellow, EIC of IEEE TII

Bogazici University, Bebek, 34342 Istanbul, Turkey

Email: o.kaynak@ieee.org

Personal Website: http://mecha.ee.boun.edu.tr/kaynak.html

��

Abstract:  The last decade of the last millennium is characterized by what might be called the intelligent systems revolution, as a result of which, it is now possible to have man­�Cmade systems that exhibit ability to reason, learn from experience and make rational decisions without human intervention. Prof. Zadeh has coined the word MIQ (machine intelligence quotient) to describe a measure of intelligence of man-made systems. In this perspective, an intelligent system can be defined as a system that has a high MIQ.

In the presentation the state-of-art reached in intelligent systems is overviewed with examples and a perspective on the future is given. The reasons behind the slow pace of developments are discussed. The talk closes with a consideration of the possible research directions in mechatronics and robotics as driving forces behind the development of intelligent systems.

��

��

Bio-Sketch: Okyay Kaynak received the B.Sc. degree with first class honors and Ph.D. degrees in electronic and electrical engineering from the University of Birmingham, UK, in 1969 and 1972 respectively.

From 1972 to 1979, he held various positions within the industry. In 1979, he joined the Department of Electrical and Electronics Engineering, Bogazici University, Istanbul, Turkey, where he is presently a Full Professor. He has served as the Chairman of the Computer Engineering and the Electrical and Electronic Engineering Departments and as the Director of Biomedical Engineering Institute at this university. Currently, he is the UNESCO Chair on Mechatronics and the Director of Mechatronics Research and Application Centre. He has hold long-term (near to or more than a year) Visiting Professor/Scholar positions at various institutions in Japan, Germany, U.S. and Singapore. His current research interests are in the fields of intelligent control and mechatronics. He has authored three books and edited five and authored or coauthored more than 200 papers that have appeared in various journals and conference proceedings.

Dr. Kaynak is a fellow of IEEE. He served as the President of the IEEE Industrial Electronics Society (2002-2003) and as the Vice President for Conferences of the IEEE Computational Intelligence Society (2004-2005). He is the Editor-in-Chief of IEEE Transactions on Industrial Informatics. Additionally he is on the Editorial or Advisory Boards of a number of scholarly journals.

��

BIZSIM: The World of Business in a Box

John L Casti, Professor & PhD, EIC of AMC

The Wissenschaftzentrum Wien and IIASA, Vienna, Austria
Email: castiwien@cs.com
Personal Website: http://internet.cybermesa.com/~roger_jones/johncasti.htm or www.wzw.at

��

Abstract: Almost every business environment involves a collection of agents---shoppers, traders, firms---interacting in ways to serve each individual agent's "best interests." These interactions generate collective effects, often termed _emergent properties_ such as price changes in a financial market or "fads" for a particular kind of clothing or restaurant. Agent-based models can be employed to study how these properties arise (or disappear) by populating a computer landscape with many such agents and putting them into various types of interactions.

This talk addresses the use of agent-based models and simulations to study particular questions in the world of business and finance. The general ideas are illustrated by examples from the global catastrophe insurance industry, supermarkets, and stock exchanges.

��

Bio-Sketch: John L Casti received my Ph.D. in mathematics under Richard Bellman at the University of Southern California in 1970. I worked at the RAND Corporation in Santa Monica, CA, and served on the faculties of the University of Arizona, NYU and Princeton before becoming one of the first members of the research staff at the International Institute for Applied Systems Analysis (IIASA) in Vienna, Austria. In 1986, I left IIASA to take up a position as a Professor of Operations Research and System Theory at the Technical University of Vienna. I am also a member of the External Faculty of the Santa Fe Institute in Santa Fe, New Mexico, USA, where I worked extensively on the application of biological metaphors to the mathematical modeling of problems in economics, finance and road-traffic networks, as well as on large-scale computer simulations for the study of such networks.

In 2000 I formed two companies in Santa Fe and London, Commodicast, Inc. and SimWorld, Ltd, devoted to the employment of tools and concepts from modern system theory for the solution of problems in business and finance. In early 2005 I returned to Vienna as a Research Fellow at the Wissenschaftzentrum Wien, where I am in the process of establishing a new research division devoted to questions at the interface of the arts/humanities/social sciences, natural sciences, and philosophy/mathematics. In addition, I have co-founded The Kenos Circle with Dr. Michael Zillner, a professional society that aims to make use of complexity science in order to gain a deeper insight into the future than that offered by more conventional statistical tools.

Over the past few years, I have written a numerous articles and seven technical monographs and textbooks on mathematical modeling. In addition, I am the editor of the journals Applied Mathematics & Computation (Elsevier, New York) and Complexity (Wiley, New York). In 1989 my text/reference work Alternate Realities: Mathematical Models of Nature and Man (Wiley, 1989) was awarded a prize by the Association of American Publishers in a competition among all scholarly books published in mathematics and the natural sciences. In 1992, I also published Reality Rules (Wiley, New York), a two-volume text on mathematical modeling.

In addition to these technical volumes, I have written several popular books on science: Paradigms Lost: Images of Man in the Mirror of Science (Morrow, 1989), which addresses several of the most puzzling controversies in modern science, Searching for Certainty: What Scientists Can Know About the Future (Morrow, 1991), a volume dealing with problems of scientific prediction and explanation of everyday events like the weather, stock market price movements and the outbreak of warfare, and Complexification (HarperCollins, 1994), a study of complex systems and the manner in which they give rise to counterintuitive, surprising behavior. I have also written two popular volumes on mathematics: Five Golden Rules: Great Theories of 20th-Century Mathematics---and Why They Matter, and its sequel, Five More Golden Rules (1995, 2000) both published by John Wiley & Sons (New York). My next work of popular science was Would-Be Worlds, a volume on computer simulation and the way it promises to change the way we do science. It was also published by John Wiley & Sons (New York) in1996. In 1998 I published a volume of ``scientific fiction,'' involving Ludwig Wittgenstein, Alan Turing, J.B.S. Haldane, C.P. Snow and Erwin Schrödinger in a dinner-party conversation on the question of the uniqueness of human cognition and the possibility of thinking machines. This book was published under the title The Cambridge Quintet by Little, Brown (UK) in December 1997 and by Addison-Wesley in the US in early 1998.

More recently, my published books include Art & Complexity (Elsevier, Amsterdam, 2003), a volume edited with A. Karlqvist, as well as a short volume on the life of the Austrian logician, Kurt Gödel, the book Gödel: A Life of Logic (Perseus Books, Cambridge, MA, 2002). In the same year I published the volume, The One, True, Platonic Heaven (Joseph Henry Press, Washington, DC, 2003), which addresses in a fictional format the question of the limits to scientific knowledge.

My current research interests have also shifted somewhat to the exploration of questions in the social and behavioral realm and the relationship between social ��moods�� and their consequent social actions and behaviors. In this direction, my latest book, Destiny��s Design: Why Human Events Happen will be published by John Wiley & Sons, London and New York, in Fall 2006. It addresses the directions and patterns of social causation and their implications for future trends and collective social events, such as styles in popular culture, the outcome of political processes, and even the rise and fall of civilizations.

��

Mapping Evolutionary Pathways of HIV-1 Drug Resistance Using Conditional Selection Pressure

Christopher John Lee, Professor & PhD

Dept. of Chemistry & Biochemistry, UCLA, Los Angeles, CA 90095.

Email: leec@mbi.ucla.edu,

Personal Website: http://www.bioinformatics.ucla.edu/leelab

��

Abstract: Can genomics provide a new level of strategic intelligence about rapidly evolving pathogens? We have developed a new approach to measure the rates of all possible evolutionary pathways in a genome, as a conditional Ka/Ks network, and have applied this to several datasets, including clinical sequencing of 50,000 HIV-1 samples. These data reveal specific accessory mutations that greatly accelerate the evolution of multi-drug resistance, and other ��kinetic trap�� mutations that block it. Our analysis was highly reproducible in four independent datasets, and can decipher a pathogen��s evolutionary pathways to multi-drug resistance even while such mutants are still rare.

Bio-Sketch: Christopher John Lee is an associate professor in the Department

of Chemistry and Biochemistry at the University of California, Los Angeles. He

received his B.A. in biochemistry and molecular biology from Harvard University (1988),

and his Ph.D. training with Michael Levitt at Stanford University (1989-1993).

His main interests include genomics and bioinformatics of alternative splicing,

new approaches for multiple sequence alignment and analysis, databases for
genome analysis, and viral genome evolution. For more information, please visit
http://www.bioinformatics.ucla.edu/leelab.

��

Neuroengineering Hard Problems: Modularity, Scale, Complexity, and Robustness

Donald Wunsch,  Professor & PhD, IEEE Fellow

Department of Electrical and Computer Engineering, University of Missouri at Rolla, USA,

Email: dwunsch@ece.umr.edu

Personal Website: http://ece.umr.edu/dwunsch.html

��

Abstract: The types of hard engineering problems that are potentially solveable by computational intelligence tend to have predictable challenges.  They may require a number of interacting solutions, each of which is challenging to develop.  Thus, they require modular approaches.  They may scale poorly and thus tax computational resources, regardless of the advances in those resources.  Scaling is of course a source of complexity, but other complexities, such as lack of appropriate metrics for data, can also make a problem difficult.  Furthermore, the need for a system to be robust to errors can often raise a problem to a higher plane of difficulty.

This presentation will motivate neuroengineering approaches by discussing several examples of solved and unsolved hard problems, each of which has one or more of the properies discussed above.  A comparison and contrast of the techniques applied is discussed, in order to motivate a new direction for neural architecture research.

��

Bio-Sketch: Donald Wunsch is the Mary K. Finley Missouri Distinguished Professor of Computer Engineering at the University of Missouri - Rolla, where he has been since July 1999.   His prior positions were Associate Professor and Director of the Applied Computational Intelligence Laboratory at Texas Tech University, Senior Principal Scientist at Boeing, Consultant for Rockwell International, and Technician for International Laser Systems.  His education includes an Executive MBA from Washington University in St. Louis in 2006, the Ph.D. in Electrical Engineering from the University of Washington (Seattle) in 1991, the M.S. in Applied Mathematics from the same institution in 1987, the B.S. in Applied Mathematics from the University of New Mexico in 1984, and he also completed a Humanities Honors Program at Seattle University in 1981.  He has over 200 publications in his research field of computational intelligence, and has attracted over $5 million in research funding.  He has produced seven Ph.D.'s in Electrical Engineering, four in Computer Engineering, and one in Computer Science.  He is an IEEE Fellow, a recipient of the Halliburton Award for Excellence in Teaching and Research and the National Science Foundation CAREER Award.  His research interests are in neural networks, and their applications in: reinforcement learning, approximate dynamic programming, the game of Go, financial engineering, graph theory, risk assessment, representation of knowledge and uncertainty, collective robotics, computer security, critical infrastructure protection, biomedical applications of computational intelligence, telecommunications, and smart sensor networks.  He served as voting member of the IEEE Neural Networks Council, Technical Program Co-Chair for IJCNN 02, General Chair for IJCNN 03, International Neural Networks Society Board of Governors Member, and is now Past-President of the International Neural Networks Society.

��

A Unified Perspective  and New Results on RHT Computing, Mixture Based Learning, and  A General Problem Solving Paradigm

Lei Xu, Professor & PhD, IEEE& IAPR Fellow, MEAS

Computer Science & Engineering Dept., The Chinese University of Hong Kong, HK

Email: lxu@cse.cuhk.edu.hk

Personal Website: http://www.cse.cuhk.edu.hk/~lxu/

��

Abstract: On one hand,  multiple   object detection approaches of Hough transform (HT) and Randomized HT types   have been extended into an evidence accumulation featured general framework for problem solving, with five key  mechanisms elaborated and several  extensions of HT and RHT presented. On the other hand, another framework is proposed to integrate  typical multi-learner based approaches for  problem solving, particularly on  Gaussian mixture based  data clustering and local subspace learning, multi-sets mixture based object detection and motion estimation, and multi-agent coordinated problem solving. Typical learning algorithms, especially those based on Rival Penalized Competitive Learning (RPCL) and  Bayesian Ying-Yang (BYY) learning, are summarized from  a unified perspective  with new extensions. Furthermore, the two different frameworks are not only examined with one viewed crossly from a perspective of the other, with new insights and extensions, but also  further unified into a general problem solving paradigm that   consists of five basic mechanisms in terms of   acquisition, allocation, amalgamation, admission, and affirmation}, or shortly  A5 paradigm.

��

Bio-Sketch: Lei Xu is a chair professor of Chinese Univ Hong Kong where he  joined in 1993. He completed his PhD thesis at Tsinghua Univ in  1986, and worked at universities during 1987-93, including Peking  Univ, Harvard and MIT. Prof. Xu has published around 100 journal  papers and book chapters on statistical learning and pattern recognition, with a number of well-cited contributions. Up to Feb, 2006, his papers have got over 1500 citations according to  SCI-Expended, and over 2500 citations according to Google Scholar  (GS).  His 10 most frequently cited papers have got over 900 (SCI) and 1700 (GS). One single paper  has scored over 301 (SCI) and 650 (GS) , each of the other nine are scored between 41-103(SCI) and 55-177(GS), respectively. Also, he is  ranked by /CiteSeer / at the 2061-th among 10,000 most cited authors  (of 773109 total authors). He has given near 50  keynote/ plenary/ invited/ tutorial talks in international conferences  such as IJCNN, WCNN, ICONIP, IEEE ICNN, etc. served or has been  serving as associate editor for several international journals, a  governor of International Neural Network Society (01-03), and a past president of Asian-Pacific Neural Networks Assembly. Also, Prof. Xu  has also been serving as a member of engineering panel, Hong Kong  RGC research committee (01-06 ), a member of selection committee, Chinese NSFC/HK RGC Joint research scheme (02 -05 ), external expert  for Chinese NSFC information science panel (04 - ) and a nominator  for the prestigious Kyoto prize (03-04 ), also served as a general  chair, Program Committee chair, as well as program /organizing  committee members on many major international conferences in his  fields, including WCNN, ICNN, IJCNN, IEEE WCCI, ICONIP, NIPS, ICANN, IEEE CIFER, and Intl. Conf. on Computational Finance, etc. Moreover, He has received several national prestigious academic awards  (including 1993 National Nature Science Award) and international  awards (including 1995 INNS Leadership Award). He is an IEEE Fellow  (01) and a Fellow of International Association for Pattern Recognition (02), and a member of European Academy of Sciences (02).

��

Neural Mechanism for Reinforcement Learning and Decision Making

Aike Guo, Professor & PhD, MCAS

Institute of Neuroscience, Shanghai Institutes for Biological Sciences, Key Laboratory of Neurobiology, Chinese Academy of Sciences, Shanghai 200031, China

State Key Laboratory of  Brain and Cognitive Science, Institute of Biophysics, Chinese Academy of Sciences, Beijing 100101, China

Email: akguo@ion.ac.cn

Personal Website: http://bcslab.ibp.ac.cn/keyandw/expert/GuoAK.asp

��

Abstract: To explore the neural mechanisms of decision making under uncertainty is essential because it is a fundamental intelligent activity at every societal level (Hsuet al., 2005). The shared interest in understanding decision making has resulted in the emergence of a new interdisciplinary field of research, as ��Neuroeconomics��, the goal of which is to understand the neural correlates of our choice behavior (Sugrue et al., 2005). To make adaptive decisions, animals must evaluate the costs and benefits of available array of options. Dopamine neurons acting as ��common reward currency�� are sensitive to the uncertainty of both the occurrence and the time of reward. Due to its sophisticated genetics, relatively simple anatomy, behavioral richness, and its remarkable molecular similarity to mammals, fruit flies Drosophila became the ��Jack of all trades�� in Life Science.

The goal of our study is to search the primary events or principles of the high brain functions. We have discovered a simple behavior of decision-making among competing alternatives in Drosophila. Flies in the ��paradoxical�� situation resolve the dilemma by taking into account the reliability of the current information during retrieval, but flies lacking mushroom bodies, a prominent part in the Drosophila brain involved in many behaviors, seem to have difficulties in resolving ��conflicting�� situations. Thus, Drosophila has greater cognitive processing abilities than they are usually given credit for (Tang & Guo, 2001).

We quest for, does decision circuit in flies share similar circuit principle with that in high-level animals? We will focus in near future on the reinforcement learning principles mediated by dopamine systems because converging evidences from higher vertebrates and Drosophila fit well with a major role of dopamine system in decision-making, as well as a central role in guiding our behaviors and thoughts as well. We believe that the enhancing and focusing effect of dopamine may be combined with neuronal plasticity to provide mechanism for reinforcement learning.

��

Bio-Sketch: Aike Guo is the Associate Director of ION �� a Senior Investigator and Head of the Laboratory of Learning and Memory. He was the chief scientist for the ��973 Program�� project "Basic Research of Brain Development and Plasticity" sponsored by the Ministry of Science and Technology of China (2000-2005). He became an academician of the Chinese Academy of Sciences in 2003. Dr. Guo's current research interest is to understand the molecular, cellular, and integrative mechanisms underlying learning, memory and higher cognitive functions, e.g., decision-making, crossmodal learning and memory, and selective attention in Drosophila. In addition to the main line mentioned above, he has also pursued several side-projects, including thermal nociception,, circadian clock, motion perception, male-male courtship, drug addiction as well as the neurodegenerations in flies. His long-term goal in these projects is also to elucidate the circuit mechanisms and relate the findings to the reinforcement learning that may be essential for addictive and courtship behaviors.

��

Intelligent Use of Data for Condition Monitoring and Applications

George W Irwin, Professor & PhD, IEEE & IEE Fellow, MRAE & MRIA

Intelligent Systems and Control Lab, School of Electronics, Electrical Engineering and Computer Science, Queen��s University Belfast, UK

Email: g.irwin@qub.ac.uk

Personal Website: http://www.ee.qub.ac.uk//staff/academic/gwi.htm

��

Abstract: The field of process control provides a useful perspective on the ICIC ��06 theme of emerging intelligent computing technology and applications.

The current generation of distributed control systems, as found on chemical, petrochemical and pharmaceutical plants use single workstations to supervise a number of individual control loops, loosely connected by a data communications network. Sensor data is then used for automatic control at the lowest level, with unit performance information being provided by a workstation at the next level up. Relatively little provision is made for processing the raw data into higher level knowledge, arguable a necessary feature of intelligent computing. Thus for example, while a localised event will trigger an alarm, operator intervention is needed to isolate the source and provide a fault diagnosis.

Autonomous fault diagnosis (and fault tolerant control) demands an integrated multidisciplinary solution characterised by a synergy between communications, computing (and control) while offering also considerable scope for intelligent computing. To illustrate these two aspects, the talk will first present some new research on reduced communication control followed by a brief introduction to conventional multivariate statistical process control (MSPC). Online monitoring results from a DuPont plant, supplied by Annex6 Ltd a former spin-out company from our group, will be given. Recent theoretical contributions, specifically our work on nonlinear MSPC using auto-associative neural networks and its extension using the statistical local approach will also be described. The talk will conclude by discussing a unique application to automotive condition monitoring, including experimental results from a 4 cylinder, 1.9 litre Volkswagen diesel engine, before looking to future research directions and challenges.

��

Bio-Sketch: George Irwin leads the Intelligent Systems and Control group and is Director of the Virtual Engineering Centre at Queen University Belfast. His research covers identification, monitoring, and control, including neural networks, fuzzy neural systems and multivariate statistics. He is currently working on wireless networked control systems, fault diagnosis of internal combustion engines and novel techniques for fast temperature measurement. Much of his work has involved industrial collaboration and he was Technical Director of Anex6 Ltd, a spin out company from his group specialising in process monitoring. He has published over 350 research papers and 6 edited books.

George Irwin has been awarded a number of prizes including four IEE Premiums, a Best Paper award from the Czech Academy of Sciences and the 2002 Honeywell International Medal from the UK Institute of Measurement and Control. International recognitions include Honorary Professor at Harbin Institute of Technology (1999) and Shandong University (2005) and Visiting Professor at Shanghai University (2005 - 2008). His group collaborates with the Institute of Intelligent Machines, Chinese Academy of Sciences as Co-General Chair of the 2^nd International Conference on Intelligent Computing Kunming China (06). In 2006 he also serves on the International  Programme Committees of the IEEE International Conference on Intelligent Systems and the IEEE Machine Learning for Signal Processing Workshop.

He is a former Editor-in-Chief of the IFAC Journal Control Engineering Practice and chaired the UK Automatic Control Council. He currently chairs the IFAC Publications Committee and serves on the editorial boards of several journals. Prof Irwin has been elected Fellow of the Royal Academy of Engineering and Member of the Royal Irish Academy. He is a Chartered Engineer, an IEEE Fellow, a Fellow of the IEE and a Fellow of the Institute of Measurement and Control.

��