Distinguished Lecturer Series

Graphical Models, Kernel Methods and Statistical Learning Algorithms

Michael Jordan

University of California, Berkeley

Friday, April 29, 2005
10:30am - SENSQ 5317

Refreshments at 10:00am

Abstract

In this lecture we will discuss two broad classes of statistical learning algorithms: "probabilistic graphical models," a formalism that exploits the conjoined talents of graph theory and probability theory to build complex models out of simpler pieces, and "kernel methods," a formalism based on computationally-efficient representations of generalized notions of similarity in Hilbert spaces. We will talk about the mathematical underpinnings for recent developments, which in both cases lie in convex analysis and convex optimization theory. We will also discuss a number of applications: (1) gene-finding with phylogenetic models, (2) haplotype phasing, (3) functional annotation of proteins, (4) finding bugs in computer programs, (5) learning hierarchical topic models of document collections, and (6) automatic image annotation.

Biography of Speaker

Michael Jordan is Professor in the Department of Electrical Engineering and Computer Science and the Department of Statistics at the University of California at Berkeley. He received his Masters from Arizona State University, and earned his PhD from the University of California, San Diego. He was a professor at the Massachusetts Institute of Technology for eleven years. He has published over 200 research papers on topics in computer science, statistics, electrical engineering, biology and cognitive science. His research in recent years has focused on probabilistic graphical models, on kernel machines, and on applications of statistical machine learning to problems in bioinformatics, information retrieval, and signal processing.