Keynote Speakers
We are very pleased to announce a number of world-class keynote speakers for KES2007. The speakers and the titles of their talks are shown below.
Walter J Freeman
University of California, Berkeley, USAThermodynamic model of knowledge retrieval in brain dynamics for information processing
More details ...
Yoshiteru Ishida
Toyohashi University of Technology, JapanThe immune system offered a glimpse: what makes biological systems distinct from artificial ones
More details ...
Hans-Andrea Loeliger
ETH, Zurich, SwitzerlandThe factor graph approach to model-based signal processing
More details ...
Jean Francois Cardoso
Ecole Nationale Superieure de TelecommunicationsIndependent components analysis: concepts and applications
More details ...
Mario Gerla
University of California, Los Angeles, USAProbing and mining the urban environment using the vehicular sensor network
More details ...
Mario Lauria
The Telethon Institute of Genetics and Medicine (TIGEM), Naples, ItalyBiologically inspired computing and self-organizing computation
More details ...
Walter J Freeman
University of California, Berkeley, USAThermodynamic model of knowledge retrieval in brain dynamics for information processing
Abstract:
Computational models for the retrieval of knowledge as it occurs in brain dynamics fall short of brain performance in speed and robustness of pattern recognition, especially in detecting minute but highly significant pattern fragments embedded in noise. Here a non-computational model is proposed that is based on studies of thermodynamic systems operating far from equilibrium, such as brains constructing order by dissipating energy. Brains create knowledge through reinforcement learning of stimuli and store it in the form of landscapes of attractors and their basins. Brains do this by forming nerve cell assemblies in cortical connectivity; each assembly governs an attractor serving a category. Retrieval of a desired category of stored knowledge is by a phase transition from a background receiving state of cortex to an active transmitting state that is induced by the information given by a learned stimulus. The prestimulus background cortical activity resembles Rayleigh noise, because it contains aperiodic null spikes at which analytic amplitude nears zero. Phase transitions in recognition and recall occur at the null spikes, because the high signal to noise ratio at the null spikes enables capture of cortex by even a very weak stimulus, which selects an appropriate basin of attraction. The approach to modeling this operation is done in terms of thermodynamics, complemented with many-body physics and quantum field theory, using concepts of spontaneous symmetry breaking and emergence of unitarily inequivalent ground states corresponding to the actualized forms of knowledge that are retrieved and deployed momentarily at rapid rates from the brain store.
Freeman WJ [2006] Definitions of state variables and state space for brain-computer interface. Part 1. Multiple hierarchical levels of brain function. Cognitive Neurodynamics 1(1): 1871-3080 (Print) 1871-4099
(0nline) http://dx.doi.org/10.1007/s11571-006-9001-x http://www.springerlink.com/content/1871-4099/?Content+Status=Accepted
Freeman WJ, Holmes MD, West GA, Vanhatalo S [2006] Fine spatiotemporal structure of phase in human intracranial EEG. Clin. Neurophysiol. 117, 6, 2006, pp 1228-1243.
http://repositories.cdlib.org/viewcontent.cgi?article=5624&context=postprints
Freeman WJ, Vitiello G (2006) Nonlinear brain dynamics as macroscopic manifestation of underlying many-body field dynamics. Physics of Life Reviews 3: 93-118.
http://dx.doi.org/10.1016/j.plrev.2006.02.001 http://repositories.cdlib.org/postprints/1515
Biography:
Walter J Freeman studied physics and mathematics at M.I.T., electronics in the Navy in World War II, philosophy at the University of Chicago, medicine at Yale University, internal medicine at Johns Hopkins, and neuropsychiatry at UCLA. He has taught brain science in the University of California at Berkeley since 1959, where he is Professor of the Graduate School. He received his M.D. cum laude in 1954, the Bennett Award from the Society of Biological Psychiatry in 1964, a Guggenheim in 1965, the MERIT Award from NIMH in 1990, the Pioneer Award from the Neural Networks Council of the IEEE in 1992, the Premió Calabria, Universitį Mediterraneo, Reggio Calabria, 2002, and the Helmholtz Lifetime Achievement Award from the International Neural Network Society 2005. He was President of the International Neural Network Society in 1994, and is Life Fellow of the IEEE. He has authored over 450 articles and 4 books: "Mass Action in the Nervous System" 1975, "Societies of Brains" 1995, "Neurodynamics" 2000, and "How Brains Make Up Their Minds 2001.
Yoshiteru Ishida
Toyohashi University of Technology, JapanThe Immune System Offers a Glimpse: what makes biological systems distinct from artificial ones
Abstract:
We discuss the intrinsic difference between biological systems and artificial systems by focusing on the immune systems as a typical biological system and computer systems as the most sophisticated man-made systems. Although both the immune system and computer systems share the information-intensive character, they are quite different in their way of handling the information. The differences can be conspicuous through the viewpoints of: serial/circular causality; redundancy/diversity; and direct/indirect information transfer. Some of the in characters dealing with the information in the immune systems may be adopted for the design principle of the new generation information systems. Other than self-maintenance, distributed, and adaptive characters, self-defining process will be stressed.
Biography:
Yoshiteru Ishida received Dr. Eng. in Applied Mathematics and Physics from Kyoto University in 1986. He served as an assistant professor at Kyoto University from 1983 to 1993. From 1994 to 1998, he has been an associate professor at Nara Institute of Science and Technology. Since 1998, he has been a professor at Department of Knowledge-based Information Engineering at Toyohashi University of Technology. He had been a visiting researcher at School of Computer Science, Carnegie-Mellon University (1986-1987), Department of Psychology, Carnegie-Mellon University (1993-1994) and Santa Fe Institute (1997-1998). His research interest includes biological complexity typified by the immune system (Immunity-Based Systems: A Design Perspective, as in his book from Springer); self-organization by a game theoretic approach; and qualitative theory on large-scale dynamical networks. Recent activities and publications can be found at:http://www.sys.tutkie.tut.ac.jp/~ishida /en/index.html
Hans-Andrea Loeliger
ETH, Zurich SwitzerlandThe factor graph approach to model based signal processing
Abstract:
Factor graphs (or similar graphical models) allow a unified approach to system modeling and algorithms in coding theory, signal processing, machine learning, and other fields. After a brief elementary introduction, the talk focuses on the systematic development of estimation algorithms in model-based signal processing. The factor graph approach encourages to mix and match different algorithmic techniques, and it allows to compose algorithms by putting together tabulated message computation rules for the building blocks of the system model.
Biography:
Hans-Andrea Loeliger received a diploma in electrical engineering in 1985 and a Ph.D. in 1992, both from ETH Zurich, Switzerland. From 1992 to 1995 he was with Linkoping University, Sweden. From 1995 to 2000, he was with Endora Tech AG, Basel, Switzerland, of which he is a cofounder. Since 2000, he has been a Professor at ETH Zurich. His research interests lie in the broad areas of signal processing, information theory, communications, and electronics. He is a fellow of the IEEE.
Jean Francois Cardoso
Ecole Nationale Superieure de Telecommunications, FranceIndependent components analysis: concepts and applications
Abstract:
Decomposing signals or images into "components" often is a key step in data processing. Fourier Analysis is the most important example of a data-independent decomposition and Principal Component Analysis (PCA) is the best known data-dependent decomposition. The last decade has seen the development of Independent Component Analysis (ICA) which analyzes signals and images into linear components which are "as independent as possible".
In contrast to PCA which yields orthogonal, uncorrelated components, ICA tries very hard to express statistical independence *beyond* decorrelation. Statistical independence being a much stronger notion than decorrelation, ICA is in particular able to recover sets of underlying components from observations using only a minimal set of assumptions (the so-called "blind" separation of components).
In this lecture, I will recall the basics of ICA and show how "upgrading" from decorrelation to independence can be achieved by resorting to non-Gaussian or non stationary components models. The lecture will demonstrate, based on examples on real data, that it is possible to keep component models strong enough to express independence beyond decorrelation but simple enough to lead to simple algorithms.
Mario Gerla
University of California, Los Angeles, USAProbing and mining the urban environment using the vehicular sensor network
Abstract:
There has been growing interest in urban surveillance using vehicles that monitor the environment, classify the events, e.g., license plate readings, and exchange metadata with neighbours in a peer-to-peer fashion. The idea is to create a totally distributed index of all the events, which can be accessed by authorized users. For instance, the Department of Transportation extracts traffic congestion statistics; the Department of Health monitors pollutants, and; the Police carries out forensic crime investigations.
In this talk we describe different techniques that can be used to maintain such a distributed index within a mobile structure like the vehicle grid. We then focus on MobEyes, a Peer to Peer middleware solution that diffuses data summaries via epidemic dissemination to create a distributed index of the massive sensor data base. We discuss the search performance of this index in different motion, density and diffusion conditions. We review the design challenges, from information dissemination to harvesting, routing and security.
Mario Lauria
The Telethon Institute of Genetics and Medicine (TIGEM), Naples, ItalyBiologically inspired computing and self-organizing computation.
Abstract:
Computer science and biology are two fields of research with a long history of mutual influence. Continuing with this tradition, two new and thought-provoking topics at the intersection of the two disciplines will be presented in this talk. In the "nature-inspired technology" part of the talk, a biologically inspired and fully-decentralized form of organization of the computation enables a population of mobile agents to colonize a peer-to-peer network. Through the careful design of agent behavior, the emerging organization of the computation can be customized for a wide variety of applications domains. In the "technology-inspired nature" part of the talk, the basic concepts of synthetic biology will be presented, along with some discussion of the challenges and possible implications of the technology.
Biography:
Current PositionVisiting Senior Scientist - Systems Biology Lab, Telethon Institute of Genetics and Medicine (TIGEM) Via P. Castellino 111, 80131 Napoli, ItalyEmail: lauria@tigem.it
Education:
Universitą di Napoli "Federico II" Electrical Engineering Laurea 1992
Universitą di Napoli "Federico II" Electrical Eng. and Computer Science PhD 1997
University of Illinois at Urbana-Champaign Computer Science M.S. 1996
Professional Experience:
Assistant Professor (adjunct), Dept of Computer Science and Eng., Ohio State University Jan 07 - present
Assistant Professor (adjunct), Dept of Biomedical Informatics, Ohio State University Sept 01 - present
Assistant Professor Dept of Computer Science & Engineering, Ohio State University Mar 00 - Jan 07
Visiting scholar Dept of Computer Science, Vrije Universiteit, Amsterdam Jan 00 - Mar 00
Postdoctoral Res. Associate Dept of Computer Science, University of California, San Diego Aug 98 - Sept 99
Postdoctoral Res. Associate Dept of Computer Science, Univ of Illinois, Urbana-Champaign Sep 97 - Aug 98
Professore a contratto, Facoltą di Ingegneria, Universitą di Napoli Jan 97 - Sept 97
Computer Systems Analysts, Ansaldo Trasporti S.p.A., Napoli, Italy Jun 93 - Mar 94
Honors:
1994 Fulbright scholarship
1997 NATO Advanced Science Fellowship
2007 Elevation to IEEE Senior Member
