Oct 20: Mathukumalli Vidyasagar: A Metric between Probability Distributions of Different Sizes

Thursday, October 20, 2011
Gilman 132

There are many ways to compare two probability distributions defined on a common set, for instance
the total variation metric.   However, in problems of reduced‐order modeling, one has to compare
probability distributions on sets of different cardinality.    In this talk a “Variation of Information”
metric is defined for such a purpose, and the problem of optimal order reduction in this metric is also
studied.   It is shown that the problems of computing the metric as well as order reduction are both
closely related to a problem in computer science known as bin‐packing with over‐stuffing.

Mathukumalli Vidyasagar received the B.S., M.S. and Ph.D. degrees in electrical engineering from the
University of Wisconsin in Madison, in 1965, 1967 and 1969 respectively. Between 1969 and 1989, he
was a Professor of Electrical Engineering at various universities in the USA and Canada. In 1989 he
returned to India as the Director of the newly created Centre for Artificial Intelligence and Robotics
(CAIR), which he built up into a leading research laboratory with about 40 scientists. In 2000 he
moved to the Indian private sector as an Executive Vice President of India’s largest software
company, Tata Consultancy Services (TCS). He retired from TCS in 2009 at the age of 62, and joined
the Erik Jonsson School of Engineering & Computer Science at the University of Texas at Dallas, as a
Cecil & Ida Green Chair in Systems Biology Science. In March 2010 he was named the Founding Head
of the newly created Bioengineering Department. His current research interests are in the application
of stochastic processes and stochastic modeling to problems in computational biology, control
systems and quantitative finance

Comments are closed.