Conselho Nacional de Redes Neurais

V Congresso Brasileiro de Redes Neurais

Palestrantes Internacionais

[Teuvo Kohonen] [Paul Werbos] [Yaser S. Abu-Mostafa]
[Mohamed A. El-Sharkawi] [Louis A. Wehenkel] [Tharam S. Dillon]

Os seguintes convidados:

  • Prof. Timo Teräsvirta (Stockholm School of Economics),
  • Dr. Pratap Sondhi (City Bank), e
  • Prof. John Moody (Oregon Graduate Institute),

    • também já foram confirmados como palestrantes durante o V CBRN.


    Teuvo Kohonen
    Dr. Eng., Professor of the Academy of Finland; Head, Neural Networks Research Centre, Helsinki University of Technology, Finland.

    His research areas are the theory of self-organization, associative memories, neural networks, and pattern recognition, in which he has published over 200 research papers and four monography books. His fifth book is on digital computers.

    Since the 1960s, Professor Kohonen has introduced several new concepts to neural computing: fundamental theories of distributed associative memory and optimal associative mappings, the learning subspace method, the self-organizing feature maps (SOMs), the learning vector quantization (LVQ), novel algorithms for symbol processing like the redundant hash addressing and dynamically expanding context, and recently, emergence of invariant-feature filters in the Adaptive-Subspace SOM (ASSOM). A new two-level SOM architecture (WEBSOM) has been developed in his laboratory for exploratory textual data mining. In the largest WEBSOM implemented so far, over one million documents have been organized: see the demo at http://websom.hut.fi/websom/.

    The best known application by Prof. Kohonen is the neural speech recognition system. He has also done design work for electronics industries.

    He is recipient of:

  • the Honorary Prize of Emil Aaltonen Foundation in 1983,
  • the Cultural Prize of the Finnish Commercial Television (MTV) in 1984,
  • the IEEE Neural Networks Council Pioneer Award in 1991,
  • the International Neural Network Society Lifetime Achievement Award in 1992,
  • Prize of the Finnish Cultural Foundation in 1994,
  • 1995 Technical Achievement Award of the IEEE Signal Processing Society,
  • Centennial Prize of the Finnish Association of Graduate Engineers (TEK) in 1996,
  • King-Sun Fu Prize in 1996, and others.

    • He is Honorary Doctor of the University of York in U.K. and Åbo Akademi in Finland, member of Academia Scientiarum et Artium Europaea, titular member of the Academié Européenne des Sciences, des Arts et des Lettres, member of the Finnish Academy of Sciences and the Finnish Academy of Engineering Sciences, IEEE Fellow, and Honorary Member of the Pattern Recognition Society of Finland as well as the Finnish Society for Medical Physics and Medical Engineering. He was elected the First VicePresident of the International Association for Pattern Recognition for the period 1982 - 84, and acted as the first President of the European Neural Network Society during 1991 - 92.

    The more recent work of Professor Kohonen is expounded in his newest book Self-Organizing Maps (Springer Series in Information Sciences, Vol. 30, 1995; Second, extended edition, 1997).


    Paul Werbos
    Paul J. Werbos holds four degrees from Harvard and the London School of Economics in: (1) economics; (2) international political systems, emphasizing European economic institutions; (3) applied mathematics, with a major in quantum physics and a minor in decision and control; (4) applied mathematics, towards an interdisciplinary Ph.D. thesis. Prior to that, during high school, he obtained an FCC First Class Commercial Radiotelephone License, and took undergraduate and graduate mathematics courses at Princeton and the University of Pennsylvania.

    For about four years after the PhD, he taught courses at Maryland in quantitative methods and global futures, and performed research in intelligent systems for policy application. Then for nine years he worked at the Department of Energy evaluating and developing a wide range of energy forecasting models. In 1989 he joined NSF as a program director in the ECS Division with emphasis on Neuroengineering. He also initiated the SBIR topic 26 which emphasizes fuel-cell automobiles, for which he is Technical Coordinator. Within the Knowledge Modeling and Computational Intelligence (KMCI) area, his main goal is to maximize the development and dissemination of step-by-step advances in systems design which will lead to an understanding and replication of the general kind of learning-based intelligence.

    He has served as President of the International Neural Network Society, where he is still on the Governing Board. He also serves on the AdCom of the IEEE Systems, Man and Cybernetic Society, and the Neural Networks Technical Committee of the IEEE Neural Networks Council.


    Yaser S. Abu-Mostafa
    Professor of Electrical Engineering and Computer Sci. California Institute of Technology CALTECH. B.Sc., Cairo University, 1979; M.S.E.E., Georgia Institute of Technology, 1981; Ph.D., Caltech, 1983. Garrett Research Fellow in Electrical Engineering, 1983; Assistant Professor, 1983-89; Associate Professor, 1989-94; Professor, 1994-.

    AREAS OF EXPERTISE: Learning systems, neural networks, financial applications, information theory, and pattern recognition.

    PUBLICATIONS: More than 60 publications, including two articles in Scientific American.

    PROFESSIONAL ACTIVITIES:

  • Founding Program Chairman, First Annual IEEE Conference on Neural Information Processing Systems (NIPS), Denver, Colorado, November 1987.
  • Founding Member, IEEE Neural Networks Council, 1988-90.
  • Current and former member of the editorial boards of several technical journals, including IEEE Transactions on Information Theory, IEEE Transactions on Circuits and Systems, Neural Networks, and Journal of Complexity.
  • Technical Consultant, Citicorp+Citibank, 1988-present. Technology evaluation at the Citicorp Corporate Technology Office (New York), and supervising the application of neural networks to forecasting FX markets and risk analysis (Hong Kong).
  • Technical Consultant, Digital Persona, Inc., 1995-present. Member of the technical advisory board, and advising on technical issues of electronic security.
  • Technical Consultant, Simplex Risk Management (Tokyo, Hong Kong and Singapore), 1997-present. Technical analysis of interest-rate derivatives for risk management and arbitrage.
  • Chairman, International Conference on Computational Finance, New York City, January 1999.
  • Chairman, International Conference on Neural Networks in the Capital Markets (NNCM), 1994 and 1996.

    • HONORS AND AWARDS:

  • The Richard P. Feynman Teaching Award, 1996.
  • The Feynman-Hughes Fellowship, 1990-93.
  • The Milton and Francis Clauser Doctoral Prize (most original doctoral thesis at Caltech), June 1983.
  • Honorary Medal from President Sadat of Egypt for distinguished graduation, October 1979.


    • Mohamed A. El-Sharkawi
    Mohamed A. El-Sharkawi is a Fellow of IEEE. He received his B.Sc. in Electrical Engineering in 1971 from Cairo High Institute of Technology, Egypt. His M.A.Sc and Ph.D. in Electrical Engineering were from the University of British Columbia in 1977 and 1980 respectively.

    In 1980, he joined the University of Washington as a faculty member where he is presently a Professor of Electrical Engineering and the Associate Chair. He has also served as the Chairman of Graduate Studies and Research.
     

  •  Founder of the international conference on the Application of Neural Networks to Power Systems (ANNPS)
  •  Co-founder of the international conference on Intelligent Systems Applications to Power (ISAP).
  •  Member of the administrative committee of the IEEE Neural Networks Council representing the Power Engineering Society.
  •  Video Tutorial Chair of the IEEE Continuing Education Committee and the neural network council.
  •  Founding Chairman of several IEEE task forces and working groups and subcommittees, including the task force on Application of Neural Networks to Power Systems, the working group on Advanced Control Strategies for dc-type Machines, and the task force on Intelligent Systems Application to Dynamic Security Assessment.
  •  Co-founder of the IEEE Subcommittee on Intelligent Systems.
  •  Member of the editorial board and associate editor of several journals including the IEEE Transactions on Neural Networks, the Engineering Intelligent Systems, and the International Journal of Neurocomputing.
  •  Co-editor of the IEEE tutorial book on the applications of NN to power systems.
  •  Author of textbook on Electric Drives to be published by PWS-Kent.
  •  Chairman of the IEEE International Electric Machines and Drives to be held in Seattle, May 1999.
  •  Organized and taught several international tutorials on intelligent systems applications, power quality and power systems.
  •  Organized and chaired numerous panel and special sessions in IEEE and other international conferences.
  •  His major areas of funded research include Intelligent systems applications, High performance precision drives and Power electronics applications to power systems
  •  Published over 120 papers and book chapters in these areas.

    • Holds 5 licensed patents: three on Adaptive Var Controller for distribution systems and two on Adaptive Sequential Controller for circuit breakers. TRENCH ELECTRIC is currently manufacturing and marketing the equipment.


    Louis Wehenkel
    Former research associate of the Belgian National Fund for Scientific Research (F.N.R.S.) and presently professor of stochastic methods in the Department of Electrical Engineering and Computer Science, at the University of Liège.

    Main publications:
    1) Books
    Automatic Learning Techniques in Power Systems -THE KLUWER INTERNATIONAL SERIES IN ENGINEERING AND COMPUTER SCIENCE, Volume 429.

    2) Papers

  • An artificial intelligence framework for on-line transient stability assessment of power systems; L. Wehenkel, Th. Van Cutsem, M. Ribbens-Pavella. IEEE Trans. on PS, Vol. PWRS-4, No.2, 1989, pp. 789-800.
  • Inductive inference applied to on-line transient stability assessment of electric power systems; L. Wehenkel, Th. Van Cutsem, M. Ribbens-Pavella. Automatica, Vol.25, No.3, 1989, pp. 445-451.
  • Decision trees and transient stability of electric power systems; L. Wehenkel, M. Pavella. Automatica, Vol.27, No.1, pp. 115-134, 1991.
  • Extended equal area criterion revisited; Y. Xue, L. Wehenkel, R. Belhomme, P. Rousseaux, M. Pavella, E. Euxibie, B. Heilbronn, J.F. Lesigne. IEEE Trans. on PS, Vol. PWRS-7, No.3, 1992, pp. 1012-1022.
  • Decision tree approach to power systems security assessment; L. Wehenkel, M. Pavella. EPES, Vol. 15, No. 1, 1993, pp. 13-36.
  • Decision tree approaches to voltage security assessment; T. Van Cutsem, L. Wehenkel, M. Pavella, B. Heilbronn, M. Goubin. Proc. of the IEE, Part C, Vol. 140, No. 3, 1993, pp. 189-198.
  • Decision tree pruning using an additive information quality measure; L. Wehenkel.Partie du chapitre 5 : Measures of information, Intelligent Systems with Uncertainty, Eds. B. Bouchon-Meunier, L. Valverde, R.R. Yager, Elsevier, 1993 , pp. 397-411. (ELSEVIER93.ps.gz 77962 bytes)
  • Decision tree based transient stability method - A case study; L. Wehenkel, M. Pavella, E. Euxibie, B. Heilbronn. IEEE Trans. on PS, Vol. 9, No. 1, 1994, pp. 459-469. (IEEEWM93.ps.gz 110709 bytes)
  • Machine learning, neural networks and statistical pattern recognition for voltage security : a comparative study; (Extended version) L. Wehenkel, T. Van Cutsem, M. Pavella, Y. Jacquemart, B. Heilbronn, and P. Pruvot.
  • Engineering Intelligent Systems for Electrical Engineering and Communications, Vol. 2, No 4, 1994, pp. 233-245. (EIS94.ps.gz 54885 bytes)
  • Contingency severity assessment for voltage security using non-parametric regression techniques; L. Wehenkel. IEEE Trans. on PS, Vol. 11, No. 1, 1996, pp. 101-111. (IEEEWM95.ps.gz 156654 bytes)
  • Automatic learning. Overview of methods and potentials; L. Wehenkel. Numéro spécial de la Revue E - SRBE, ``Apprentissage automatique - Applications aux réseaux d'énergie électrique'', décembre 1996, pp. 6-18.
  • SIME : A hybrid approach to fast transient stability assessment and contingency selection; Y. Zhang, L. Wehenkel, P. Rousseaux, M. Pavella. EPES, Vol. 19, No. 3, 1997, pp. 195-208.
  • Machine-learning approaches to power-system security assessment; L. Wehenkel. IEEE Expert, Intelligent Systems & their Applications, Vol. 12, No. 5, 1997, pp. 60-72.
  • Use of Kohonen feature maps for the analysis of voltage security related electrical distances; L. Wehenkel, Y. Jacquemart.Industrial Applications of Neural Netzorks, F. Fogelman Soulié and P. Gallinari (eds.) World Scientific, 1997, pp. 101-110.
  • SIME : a comprehensive approach to fast transient stability assessment; Y. Zhang, L. Wehenkel, M. Pavella. Transactions of the IEE of Japan (Section B), Vol. 118-B, Jan. 1998.
  • Electric power system dynamic security assessment; M. Pavella, L. Wehenkel. Revue Internationale Générale d'Electricité, Vol. 1, 1998.
  • Automatic induction of fuzzy decision trees and its application to power system security assessment;X. Boyen, L. Wehenkel. Int. Journal on Fuzzy Sets and Systems, Vol. 102, No 1, pp. 3-19, 1999.
  • Probabilistic design of power-system special stability controls; L. Wehenkel, C. Lebrevelec, M. Trotignon, J. Batut., Control Engineering Practice, Vol. 7, No. 2, 1999, pp. 183-194.
  • Transient stability-constrained maximum allowable transfer;  A. L. Bettiol, L. Wehenkel and M. Pavella.  IEEE Transactions on Power Systems, Vol. 14, Number 2, May 1999, pp. 654-659.
  • Data Mining. Tutorial paper;  C. Olaru and L. Wehenkel. IEEE Magazine on Computer Applications in Power, Vol. 12, Number 3, July 1999.
  • On neurofuzzy and fuzzy decision trees approaches; C. Olaru, L. Wehenkel (ivited contribution) in Information, Uncertainty and Fusion Eds.: B. Bouchon-Meunier, R. R. Yager, L. A. Zadeh, Kluwer Academic Publishers 2000, pp. 131-145.


    • Tharam S. Dillon BE Monash, PhD Monash, FIEEE, FIEAust, FACS, FsaRS.
    Professor Dillon has published six books - two authored and four edited. He currently has another book being reviewed for publication. He also has over 320 scientific papers published as book chapters, in international and national journals or refereed conference proceedings.

    Intelligent Systems
    In the area of Neural Networks, he has developed:
    i. proofs of convergence of the graded response Hopfield net with asymmetrical connection matrices, both in the absence and presence of noise
    ii. proofs of convergence for the self adaptive Cottrell and Forte nets which are a generalisation of unsupervised learning techniques, and
    iii. methods of incorporating prior knowledge into the neural network.

    He was also the first researcher to develop a technique for load forecasting in power systems based on neural networks, as far back as 1975 . This technique is now widely used in industry. Its importance is reflected in its inclusion as one of the subjects in an IEEE video, presented by Professor Dillon in 1996.

    In the area of Neuro-Expert Systems, he has developed a technique of combining the neural net paradigm with the symbolic knowledge based paradigm in an integrated fashion to produce a generic architecture for hybrid systems (The GENEUS Architecture).

    In addition, he has developed a technique for learning in such Hybrid Neuro-Expert Systems . He has also developed an approach that combines Fuzzy techniques, Neural Nets and Case Based Reasoning.

    He is currently writing a book on these Hybrid Neuro-Expert Systems which has been commissioned by Kluwer. The Hybrid Neuro-Expert System has also been applied to the practical problem of Alarm Handling in a computer control centre and shows much better performance than a system based on Symbolic Expert Systems.

    In the field of Intelligent Systems, he works on the following areas:
    1. Automated Knowledge Acquisition
    2. Knowledge Based Systems
    3. Neural Networks
    4. Neuro-Expert Systems.

    In the area of Automated Knowledge Acquisition, he has developed three important techniques:
    i. a method of feature selection for decision trees based on the generalisation of the Goodman-Kruskall Tau measure of association from statistics
    ii. a method of learning examples using a supervised neural net followed by analysis of the net leading to extraction of symbolic knowledge in the form of concepts, concept hierarchies and production rules (The BRAINNE System), and
    iii. extracting symbolic knowledge from unsupervised nets].

    In the area of Symbolic Knowledge Based Systems, he has developed:
    i. a technique for including time-related features in knowledge based systems to permit the modelling of real-time systems
    ii. a method of validation and verification of expert systems based on a new class of high level nets, namely State Controlled Petri Nets and
    iii. second generation expert systems that combine a Petri net based qualitative model and a frame/production rule system

    Software Engineering
    In the field of Software Engineering, his work covers five areas:
    1. Object Oriented Conceptual Modelling
    2. Software Metrics
    3. User Interface Design
    4. Knowledge Based Software Engineering
    5. Theoretical Foundations of Object Oriented Systems.

    Conceptual Modelling is essentially the first step of development of a software system. It consists of developing a model of that portion of the real world that the system is meant to address. His work covers the development of a software engineering methodology for three classes of systems:
    i. traditional procedurally oriented software
    ii. databases
    iii. knowledge based systems.

    The stages involved in the development of software for these three classes of systems can be broken down into:
    a. development of a conceptual model
    b. development of a software structure model
    c. implementation.

    He shows that, with suitable extensions, the object oriented paradigm can be used to develop a conceptual model for all three classes of systems.

    In the area of Software Metrics, he has developed a software metric based on information theory for measuring the complexity of software. This is useful in characterising the "goodness" of a design based on the Structured Analysis and Design Methodology. He has also developed a metric for the useability of software based on Fuzzy Systems Theory. In the area of User Interface development, a methodology of design of user interfaces for object oriented environments has been developed. This divides the design into two parts; the logical design and the perceptual design. Perspectives of the conceptual model are used for developing the logical design. A prototype based method coupled with a systematic method of useability evaluation (based on a fuzzy system useability metric) are used for the perceptual design. In the Knowledge Based Software Engineering area, a knowledge based system has been developed for computer aided design for parts of the software development process.

    Two specific areas have been successfully addressed so far, using a knowledge based system to carry out:
    i. transformation of an object oriented conceptual model into a relational logical model suitable for implementation in a Relational Database Management System, and
    ii. transformation of an object oriented logical model of a user interface into a perceptual representation suitable for an initial prototype of the user interface. Here the knowledge base is derived from guidelines for a particular environment; for example, CUA for the IBM PC.

    A theoretical framework for Object Oriented Systems that properly characterises the notion of inheritance, including overriding has been developed.

    Computer Networks and Protocols
    In the field of Computer Networks and Protocols, his work is in four areas:
    1. Fault Detection in Computer Networks
    2. Formal Description of ISO Protocols
    3. Object Oriented Design of Protocols
    4. Object Oriented Generic Simulator for Networks and Architectures.

    The work on Fault Detection utilises an expert system for detection of faults on the Domain Name Server (DNS) system. The work on Formal Description of Protocols uses high level Petri nets to model several ISO Protocols, including]:
    i. File Transfer and Access Mode (FTAM) Protocol, and
    ii. Common Application Service Elements (CASE).

    Several errors in these protocols were uncovered and the results fed back to the relevant ISO Working Group, allowing corrections before the final versions of the documents for the protocols were released.

    Currently specification of ISO protocols attracts a considerable degree of effort, and then an ASN 1 engine is used to implement the Protocol Data Units associated with a particular protocol. There is little in the way of a systematic design methodology. We have proposed a systematic design methodology based on the object oriented paradigm. This required extension of the work on object oriented systems to properly characterise inter-object.

    The last area involves the development of a generic object oriented simulator capable of modelling computer networks and architectures. It utilises the facility of polymorphism to allow quick interchange of modules for the system being modelled.

    Reliable, Fault Tolerant Systems and Intrusion Detection
    In the field of Reliable Fault Tolerant Systems and Intrusion Detection, the work covers three areas:
    1. Modelling of Fault Tolerant Systems
    2. Architecture of Fault Tolerant Systems
    3. Neural Net Based Approach to Intrusion Detection and Mathematical Modelling.

    In the area of Modelling of Fault Tolerant Systems, the work consists of extending the existing mathematical models for such systems to cater for reconfiguration and repair. Specifically, the extensions allow for modelling of effects of imperfect and possibly deleterious repair as well as modelling latent faults.

    In the area of architecture, a distributed software implemented fault tolerant architecture was developed and implemented using microprocessors. It was also used as the basis for development of an ultra-reliable Signalling Transfer Point (STP) for the No. 7 Signalling System for a telecommunication network. Work is also done on software fault tolerant architectures and techniques.

    In the last area, a Neural Net Based Approach for Intrusion Detection using an unsupervised neural network is utilised to learn normal patterns of operating system calls. This net is then used to detect abnormal situations. Further mathematical modelling of intrusion detection and virus attacks has been carried out.


    Data de atualização desta página: 16/Nov/2000, 12/Mar/2001
    Data de instalação desta página: 16/Novembro/2000
    Página instalada e atualizada por: Cairo L. Nascimento Jr.