University of California, Irvine
Wednesday, June 2nd, 2010
Abstract: Recently, interference alignment has been shown to be an extremely effective interference management technique in wireless networks which potentially yields rates much higher than conventional interference management schemes. In this talk, we will explore the principles and applications of alignment in three parts. The first part of the talk will provide an overview of the current state of the art in interference alignment literature. In particular coding techniques that are fundamental to alignment will be explained through simple examples. We will argue that the coding techniques of joint coding, symbol extensions and asymmetric complex signaling, which are not required for point-to-point (interference-free) networks, are however useful in aligning interference in wireless interference networks. While a significant body of recent literature on interference networks has focused on exploring the benefits of these techniques (viz. joint coding, symbol extensions and lattice coding), a parallel body of work has been to find regimes in interference networks where these (relatively sophisticated) interference management techniques are not required. These regimes, also called the “noisy interference regimes” find conditions in interference networks where random, circularly symmetric Gaussian coding and treating interference as noise achieves capacity. In the second part of the talk, we will explore the fundamental idea of alignment and present an interesting connection between these two seemingly distinct bodies of work, viz. interference alignment and noisy interference regimes. In the third part of the talk, we will present a network coding based framework with applications to distributed storage systems, data dissemination and harvesting, and distributed sensing and estimation. We will tailor this framework to the particular application of exact regeneration of failed nodes in distributed storage systems, and present a solution for the same. A key insight of our solution is that exact regeneration and functional regeneration are asymptotically equally efficient, in terms of repair bandwidth of failed nodes, in distributed storage systems.
Viveck R. Cadambe received his B.S. and M.S. degrees in Electrical engineering from the Indian Institute of Technology Madras, Chennai, India in 2006. He is currently working toward my Ph.D. degree at the University of California, Irvine. His research interests include multiuser information theory and wireless networks. Mr. Cadambe is a recipient of the 2009 IEEE Information Theory Society Paper Award and the UCI Electrical Engineering and Computer Science Department Best Paper Award for 2008-09. He also received the University of California, Irvine CPCC graduate fellowship for the year 2007-08.
Host: Alex Dimakis, dimakis [at] usc.edu
Back to CommNetS Seminar Page