[Colloquium] JOB TALK TODAY: Risi Kondor, CalTech

[Katie Casey]

DEPARTMENT OF COMPUTER SCIENCE

UNIVERSITY OF CHICAGO

Date: Friday, January 28, 2011
Time: 12:00 p.m.
Place: Eckhart Hall 110

----------------------------------------------

Speaker:		Risi Kondor

From:		California Institute of Technology

Web page:	http://www.its.caltech.edu/~risi/index.html

Title: 		Non-commutative harmonic analysis in machine learning

Abstract:  	Non-commutative harmonic analysis allows one to generalize the notion of Fourier transformation to rotations, permutations, or indeed, any compact group of transformations acting on some underlying space. We have found that this theory and the corresponding fast Fourier transforms have a whole range of natural applications in machine learning and other areas of computer science. I will give an overview of these developments, touching on invariant features for computer vision, compact representations of uncertainty in multi-object tracking, graph kernels, and strategies for solving hard optimization problems.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20110128/d0d63348/attachment.htm