Broad Area Colloquium For AI-Geometry-Graphics-Robotics-Vision

From Bits to Information: Theory and Applications of Learning Machines

Tomaso Poggio
Center for Biological and Computational Learning
AI Laboratory and McGovern Institute for Brain Research

Monday, Dec 4, 2000, 4:15PM
TCseq200, Lecture Hall A


Learning is becoming the central problem in trying to understand intelligence and in trying to develop intelligent machines. I will outline some of our recent efforts in developing machines that learn. I will sketch our work on statistical learning theory and recent theoretical results on the problem of classification and function approximation that connect regularization theory and Support Vector Machines. Our main application focus is classification (and regression) in various domains -- such as sound, text, video and bioinformatics. In particular, I will describe the recent evolution of a trainable object detection system for classifying objects -- such as faces and people and cars -- in complex cluttered images. Finally, I will speculate on the implications of this research for how the brain works and review some recent data which provide a glimpse of how visual cortex learns to identify and classify 3D objects.

About the Speaker

Professor Tomaso Poggio is the Uncas and Helen Whitaker Professor in the Department of Brain and Cognitive Sciences at MIT, where he directs research in machine learning, computer vision and neuroscience at the MIT Artificial Intelligence Laboratory and at the MIT Center for Biological and Computational Learning.

Known for his work on the visual system of the fly, nonlinear systems theory, stereo vision and for introducing regularization theory in computer vision, Professor Poggio and his group have spent the past ten years developing a theory of networks for learning in the framework of multivariate function approximation and applying them to several different domains from signal processing to multimedia database search, from finance to bioinformatics, from computer graphics to computer vision for object detection and recognition. He has also contributed to models of the brain and most recently to models of IT cortex for object recognition that are supported in the meantime by physiological data.

Co-Director, since 1992, of CBCL, the Center for Biological and Computational Learning at MIT, Professor Poggio is the author of numerous papers in areas ranging from psychophysics and biophysics to information processing in man and machine, artificial intelligence, machine vision and learning. Serving on the editorial boards of a number of leading interdisciplinary journals, Professor Poggio is a member of the American Academy of Arts and Sciences, a Fellow of the American Association for Artificial Intelligence, and an Honorary Associate of the Neuroscience Research Program at Rockefeller University. Professor Poggio received his doctorate in theoretical physics from the University of Genoa in 1970, had a tenured research position at the Max Planck Institute from 1971 to 1981 when he became Professor at MIT. In the years since then, he has received a number of distinguished international awards in the scientific community. A former Corporate Fellow of Thinking Machines Corporation, he has been involved in several companies in the areas of computer graphics, computer vision, computer networks, financial engineering and bioinformatics.
Back to the Colloquium Page
Last modified: Thu Oct 19 18:18:43 PST 2000