Abstract

The search for coding schemes which achieve the capacity of a given point to point communication link has been going on for the last six decades. Major breakthroughs were achieved in the past decade, and a significant contribution is the Polar codes that was recently introduced by Arikan. These codes achieve the capacity for a large class of channels using low complexity encoding and decoding algorithms. The complexity of these algorithms scales as O(N log N) where "N" is the blocklength of the code. These are the first known practical schemes that achieve the capacity for a large class of channels. These codes have interesting relationships to some existing coding schemes, both algebraic and iterative.
In the first talk, we start with a brief review of some existing coding schemes. We will then discuss the construction of polar codes and compare with existing schemes. We futher show that polar codes are also optimal for lossy source coding. We show that they achieve the optimal rate-distortion trade-off with a low-complexity (O(N log N)) successive cancellation algorithm. Generalizations and applications to important problems in multi-terminal source-coding will also be discussed.
In the second talk, we will discuss the details of the polar code constructions and the decoding algorithms for the various problems we mentioned in the first talk. In the end we will discuss some open problems and future research directions.

Speaker Bio

Satish Babu Korada received the B.Tech in EE from the IIT Delhi in 2004 and his PhD in the Communcations Theory Laboratory of EPFL, Lausanne, Switzerland with under Rudiger Urbanke and Nicolas Macris. He is starting at Stanford University in Aug 2009.
The broad area of his research is in coding and information theory. More specifically, he is interested in the analysis of large graphical models in communications. Modern, "iterative," source and channel coding techniques fall naturally into this framework; but also large scale multi-user problems (like CDMA), or polarization codes can be formulated in this context. His work centers around the analysis of such systems, both of optimal as well as of practical (but suboptimal) schemes. Tools from statistical mechanics (such as the interpolation method, correlation inequalities, or the replica method) are proving very useful in this context. Also probabilistic methods and graph theory play an important role in his work.
He received the Best Student Paper award at the International Symposium on Information Theory 2008 in Toronto for the paper "Exchange of Limits: Why Iterative Decoding Works". He also received the Best Student Paper award at the ISIT 2009 in Seoul for the paper "Polar Codes: Characterization of Exponent, Bounds, and Constructions."