Optimal Real-Time Coding of Markov Sources

Add to your list(s) Download to your calendar using vCal

• Dr Tamas Linder, Queens University, Kingston, Canada
• Thursday 10 May 2018, 11:00-12:00
• Room 611 (Gabor Seminar Room), Level 6, EEE Department.

If you have a question about this talk, please contact Joan P O'Brien.

Abstract: The traditional information theory of data compression allows asymptotically large coding delays, but in many applications the real-time operation is essential and even moderate delays may not be tolerated. In this talk we focus on optimal real-time (zero-delay) coding of Markov sources. After a review of some fundamental classical structural results, we give a stochastic control formulation of the optimal coding problem for a general class of real valued Markov sources. The resulting controlled Markov process facilitates the study of the existence and structure of optimal codes, but it poses subtle technical challenges due to its abstract state and action spaces. For the technically simpler case of irreducible and aperiodic discrete Markov sources we demonstrate the optimality of deterministic and stationary Markov coding policies and quantify the finite block length (time horizon) performance of such codes. Finally, we review what is known regarding the connection of the optimal performance of zero delay codes and the so-called causal information-theoretic rate distortion function.

Bio: Tamas Linder received the M.Sc. degree from the Technical University of Budapest, Hungary, in 1988, and the Ph.D degree from the Hungarian Academy of Sciences in 1992, both in electrical engineering. He was a post-doctoral researcher at the University of Hawaii in 1992 and a Visiting Fulbright Scholar at the University of Illinois at Urbana-Champaign in 1993-1994. From 1994 to 1998 he was a faculty member at the Technical University of Budapest. From 1996 to 1998 he was also a visiting researcher in the Department of Electrical and Computer Engineering, University of California, San Diego. In 1998 he joined Queen’s University where he is now a Professor of Mathematics and Engineering in the Department of Mathematics and Statistics. His research interests include communications and information theory, source coding and vector quantization, machine learning, statistical pattern recognition, and tochastic control. Dr. Linder is an elected Fellow of the IEEE .

This talk is part of the Featured talks series.

This talk is included in these lists:

• Andrea Picciau's list
• COMMSP & CP list
• COMMSP Seminar
• Featured talks
• Room 611 (Gabor Seminar Room), Level 6, EEE Department
• SAP Talks

Note that ex-directory lists are not shown.