Finite-Length Optimization of Iterative Coding Systems Via Scaling

SPEAKER: Prof. Ruediger Urbanke, Swiss Federal Institute of TechnologyABSTRACT: The infinite-length performance of iterative coding systems is by now relatively well understood. Much less is known about the practical important case of finite-length codes. I will argue that a method called "finite-length scaling", which is based on relating the code performance of various lengths, is a useful tool for the analysis and optimization of real systems. This method has the advantage that it is in principle applicable in a fairly general setting.So far, however, the method has been worked out explicitly only for the case of transmission over the binary erasure channel and we are just about to understand the case of transmission over the binary symmetric channel under Gallager A decoding.I will describe what remains to be done and what we can expect to see as final result.[This is joint work with Abdelaziz Amraoui, Jeremie Ezri, and Andrea Montanari.]Bio: Ruediger L. Urbanke received his Diplomingenieur from the Vienna Institute of Technology in 1990 and his M.S. and Ph.D. in Electrical Engineering from Washington University in 1992 and 1995, respectively. From 1995-1999 he held a position at the Mathematics of Communications Department at Bell Labs. Since November 1999, he has been on the Faculty of the Swiss Federal Institute of Technology in Lausanne. He is a recipient of a Fulbright Scholarship and a co-recipient of the IEEE Transactions on Information Theory 2002 Best Paper Award. Since October 2000 he has been an Associate Editor of the IEEE Transactions on Information Theory.Host: Prof. Giuseppe Caire, caire@usc.edu

Location: Vivian Hall of Engineering (VHE) - 217

Audiences: Everyone Is Invited

Contact: Mayumi Thrasher