BCIG SPEAKER EVENT: “Predictive Learning, Induction and Philosophy of Science”

ABSTRACT: The field of Predictive Learning (aka Soft Computing, Machine Learning, Pattern Recognition and Data Mining) is concerned with estimating ‘good’ or useful models from available data. Such problems can be usually stated in the framework of inductive learning, where the goal is to come up with a model (generalization) from several known observations (data samples). Since the number of data samples is finite, such problems are ill-posed, and this leads to the development of numerous machine learning algorithms. Induction is commonly (albeit subconsciously) used by all humans for reasoning and coping with life’s uncertainties. The classical philosophy of science is also concerned with principles of inductive learning and epistemology. So there is a clear connection between predictive learning, induction and the philosophy of science. I will investigate this connection, and relate the main concepts developed in Vapnik-Chervonenkis (VC) learning theory to similar concepts and principles in the philosophy of science. In particular, I will discuss the connection between the VC-dimension and two philosophical principles, parsimony (Occam’s razor) and Popper’s falsifiability. Also, we discuss a novel interpretation of the concept of margin (in Support Vector Machine methods) using Popper’s falsifiability.

DINNER TO FOLLOW: A self-hosted (you pay for what you consume) dinner with Professor Cherkassky will follow the lecture. Most likely this will be at one of BCIG’s big three favorites: Maggiano’s, Guapos or Haandi, depending upon the peoples’ choice.

3:00 - 4:30 pm February 14, 2008

\

Vladimir Cherkassky

University of Minnesota

Vladimir Cherkassky is Professor of Electrical and Computer Engineering at the University of Minnesota. He received Ph.D. in Electrical Engineering from University of Texas at Austin in 1985. His current research is on methods for predictive learning from data, and he has co-authored a monograph Learning From Data published by Wiley in 1998. Professor Cherkassky has served on the Governing Board of the International Neural Network Society (INNS), and he has served on editorial boards of IEEE Transactions on Neural Networks, the Neural Networks Journal, the Natural Computing Journal and the Neural Processing Letters. He was Director of NATO Advanced Study Institute (ASI) From Statistics to Neural Networks: Theory and Pattern Recognition Applications held in France, in 1993. He presented numerous tutorials on neural network and statistical methods for learning from data. He was elected as Fellow of IEEE for ‘contributions and leadership in statistical learning and neural networks’.

SPEAKER CONTACT INFORMATION:

Professor Vladimir Cherkassky
Department of Electrical and Computer Engineering

University of Minnesota
200 Union Street SE
Minneapolis, MN 55455 USA
Phone: (612) 625-9597
Fax: (612) 625-4583
E-mail: cherk001@umn.edu 
Website: http://www.ece.umn.edu/~cherkass/

Related Links

PROFESSOR CHERKASSKY’S RECENT PUBLICATIONS:

V. Cherkassky and F. Mulier, Learning from Data: Concepts, Theory and Methods, Wiley Interscience, 1998.

V. Cherkassky , J.H. Friedman and H. Wechsler (Eds.), From Statistics To Neural Networks. Theory and Pattern Recognition Applications, NATO ASI Series F, v.136, Springer-Verlag, 1994.

V. Cherkassky, X. Shao, F. Mulier and V. Vapnik, Model selection for regression using VC generalization bounds, IEEE Trans on Neural Networks, 10,5, 1999, 1075-1089

X. Shao, V. Cherkassky and W. Li, Measuring the VC-dimension using optimized experimental design, Neural Computation, MIT Press, 2000, 12, 8, 1969-1986

Cherkassky and X. Shao, Signal estimation and denoising using VC-theory, Neural Networks, Pergamon, 14, 2001, 37-52

Cherkassky, Model complexity control and statistical learning theory, Natural Computing, Kluwer,1,2002, 109-133.