From: Rosa Teorell (rosat@microsoft.com)
Date: Wed Jan 14 2004 - 11:51:51 PST
You are invited to attend...
************************************************************************
*****************************
WHO: Victor Eliashberg
AFFILIATION: Stanford University, Department of Electrical
Engineering & Avel Elecronics
TITLE: Concept of E-machine: How does a "dynamical" brain
learn to process "symbolic" information?
WHEN: Fri 2/6/2004
WHERE: 113/1021 Research Lecture Room, Microsoft Research
TIME: 10:30AM-12:00PM
HOST: Michael H. Freedman
************************************************************************
******************************
ABSTRACT:
The human brain has many remarkable information processing
characteristics that deeply puzzle scientists and engineers. Among the
most important and the most intriguing of these characteristics are the
brain's broad universality as a learning system and its mysterious
ability to dynamically change (reconfigure) its behavior depending on a
combinatorial number of different contexts. This talk discusses a class
of hypothetically brain-like dynamically reconfigurable associative
learning systems that shed light on the possible nature of these brain's
properties. The systems are arranged on the general principle referred
to as the concept of E-machine (Eliashberg, 1967, 1979, 1981, 1979,
1990).
The talk addresses the following questions:
1. How can "dynamical" (analog) neural networks function as universal
programmable "symbolic" machines?
2. What kind of a universal programmable symbolic machine can form
arbitrarily complex software in the process of programming similar to
the process of biological associative learning?
3. How can a universal learning machine dynamically reconfigure its
software depending on a combinatorial number of possible contexts? What
is working memory and mental set? Why does the data need to "decay" in
our short-term memory?
The talk explains the concept of E-machine and outlines a broad range of
its potential applications. These applications include:
context-sensitive associative memory, context-dependent pattern
classification, context-dependent motor control, imitation, simulation
of complex "informal" environments and natural language.
BIO:
Dr. Victor Eliashberg is a system engineer with a broad hands-on
experience in the design and implementation of different types of
digital, analog and mixed signal systems (hardware and software). He has
an eclectic background that includes computer science, electrical
engineering, applied mathematics, physics, neurobiology and psychology.
He was involved in mathematical simulation of distributed chemical
reactors, design and implementation of industrial control systems,
physical and biomedical instrumentation, and the simulation of linear
electron accelerators.
Victor received his MS in electrical engineering from the Leningrad
Electrical Engineering Institute, and his Ph.D. in automatic control
from the Leningrad Technological Institute (former USSR). For many
years, he combined the work as an industrial system engineer with the
study of the problem of information processing in the brain. He has
several publications related to the latter problem.
Victor is a current member of the International Neural Network Society
and the former member of the American Artificial Intelligence Society
and the Cognitive Science Society. He worked as a principal investigator
on a neural network project sponsored by DOD (with Prof. B. Widrow, Dr.
T.M. Hoff, and Dr. M Gluck). He was a visiting scholar at the Tel Aviv
University, the University of Maryland, the UC Berkeley, the IHES
(France), and the New York University. He is currently a consulting
professor at the Department of Electrical Engineering at the Stanford
University and the president of a small consulting company, Avel
Elecronics, that is involved in the development of educational and
research software for brain modelling and cognitive modelling.
_______________________________________________
Theory-group mailing list
Theory-group@cs.washington.edu
http://mailman.cs.washington.edu/mailman/listinfo/theory-group
This archive was generated by hypermail 2.1.6 : Wed Jan 14 2004 - 11:52:38 PST