ECE Colloquium - October 24


Dr. Sundeep Rangan, New York University - Polytechnic Institute

DateTime: 
Wednesday, October 24, 2012 - 10:00am - 12:00pm
Location: 

CoRE Building Room 701 Boardroom

Abstract

Sparsity has emerged as a powerful modeling concept in a wide range of signal processing and learning problems: Many high-dimensional objects admit a sparse representation in a suitable transform basis and the recent field of compressed sensing has provided efficient methods to exploit this sparse structure in estimation, dimensionality reduction and feature extraction. However, sparsity is generally not the only aspect of many practical inverse problems. This talk presents a powerful new class of algorithms called generalized approximate message passing (GAMP) that significantly extends the compressed sensing framework. The GAMP algorithm use Gaussian approximations of loopy belief propagation to reduce estimation problems on complex interconnected systems to smaller, local problems associated with the individual system components. For compressed sensing estimation, the GAMP method can incorporate arbitrary priors and nonlinearities, perform joint estimation of latent variables, and can exploit large classes of complex statistical relationships between variables. Moreover, for certain inverse problems with large random transforms, the GAMP method admits remarkably precise asymptotic performance characterizations with testable conditions for optimality -- even for problems that are non-convex and nonlinear. Applications are demonstrated in challenging identification problems in neuroscience and in optimization problems in wireless interference coordination. Joint work with Alyson Fletcher (UCSC), Phil Schniter (Ohio State), Ulugbek Kamilov (EPFL), Vivek K Goyal (MIT) and Lav Varshney (IBM).

Biography

Dr. Sundeep Rangan received the B.A.Sc. at the University of Waterloo, Canada and the M.Sc. and Ph.D. at the University of California, Berkeley, all in Electrical Engineering. He has held postdoctoral appointments at the University of Michigan, Ann Arbor and Bell Labs. In 2000, he co-founded (with four others) Flarion Technologies, a spin off of Bell Labs, that developed Flash OFDM, one of the first cellular OFDM data systems, and precursor to 4G cellular standards such as LTE and WiMax. Flarion grew to over 150 employees with trials worldwide. In 2006, Flarion was acquired by Qualcomm Technologies where Dr. Rangan was a Director of Engineering involved in OFDM infrastructure products. He joined the ECE department at NYU-Poly in 2010. His research interests are in wireless communications, signal processing, information theory and control theory.