Communication by Regression
Mon., April 21, 2014
3:00 - 4:00 PM
521 Cory Hall
Abstract
For the additive white Gaussian noise (AWGN) channel, with average codeword power constraint, codes based on the statistical high-dimensional regression framework are developed. The codewords are linear combinations of subsets of vectors from a given dictionary, with the possible messages indexed by the choice of subset. An adaptive successive decoder is developed, with which communication is shown to be reliable, with error probability nearly exponentially small, for all rates below the Shannon capacity. I will also discuss extensions of this work to the problem of lossy compression with the squared-error distortion criterion.
Bio
Antony Joseph completed his Ph.D. in Statistics in June 2012 from Yale University under the supervision of Andrew Barron. His Ph.D. research was on developing information-theoretically optimal codes based on the statistical high-dimensional regression framework. Currently, he is a joint post-doctoral scholar with Bin Yu (Dept. of Statistics & EECS, Berkeley) and Erwin Frise (Dept. of Genome Dynamics, LBNL). His research interests includes high-dimensional statistics, community detection in networks, and application of statistical methods to systems biology.
UC Berkeley Networking
Varun Jog and Ka Kit Lam Last Modification Date: Sunday, January 26, 2014