BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Joint Source-Channel Coding with Fading Channel and Side Informati
 on - Dr. Deniz Gunduz\, Imperial College London
DTSTART:20140130T150000Z
DTEND:20140130T160000Z
UID:TALK49102@talks.cam.ac.uk
CONTACT:Prof. Ramji Venkataramanan
DESCRIPTION:In a point-to-point communication system we can separate sourc
 e and channel coding without the loss of end-to-end optimality. This funda
 mental result in information theory suggests that sources can be compresse
 d independent of the channel parameters\, and similarly\, channel coding c
 an be optimized independent of the source characteristics\, leading to a m
 odular system design. However\, the optimality of separate design breaks d
 own when communicating delay-limited sources\, such as video\, over slow f
 ading wireless channels\, due to the non-ergodicity of the channel within 
 the delay deadline of the source samples. \n\nIn this talk we will study o
 ne such source-channel communication scenario\, which considers\, in addit
 ion to a time-varying communication channel\, the presence of time-varying
  correlated side information at the receiver. While the optimality of sepa
 ration holds for some side information distributions when the communicatio
 n channel is static\, in general\, separation is suboptimal in this scenar
 io. Surprisingly\, uncoded transmission is shown to be optimal for a class
  of side information distributions. To the best of our knowledge\, this is
  the first example in which uncoded transmission achieves the optimal perf
 ormance thanks to the time-varying nature of the system\, while it is subo
 ptimal in the static version of the problem. We then consider a hybrid dig
 ital-analog transmission scheme\, and show its asymptotic optimality in th
 e high SNR regime for a wider class of system parameters\, for which pure 
 digital or pure analog schemes perform far below the optimal performance.\
 n\n*BIO*: Bio\nDeniz Gunduz received his M.S. and Ph.D. degrees in electri
 cal engineering from Polytechnic Institute of New York University\, Brookl
 yn\, NY in 2004 and 2007\, respectively. Currently he is a Lecturer in the
  Electrical and Electronic Engineering Department of Imperial College Lond
 on\, UK. He was a research associate at CTTC in Barcelona\, Spain from Nov
 ember 2009 until September 2012. He also held a visiting researcher positi
 on at Princeton University from November 2009 until November 2011. Previou
 sly he was a consulting assistant professor at the Department of Electrica
 l Engineering\, Stanford University\, and a postdoctoral Research Associat
 e at the Department of Electrical Engineering\, Princeton University.\nHe 
 is an Associate Editor of the IEEE Transactions on Communications\, and se
 rved as a guest editor for the EURASIP Journal on Wireless Communications 
 and Networking\, Special Issue on Recent Advances in Optimization Techniqu
 es in Wireless Communication Networks. He is currently serving as a co-cha
 ir of the IEEE Information Theory Society Student Committee. He was a co-c
 hair of the Network Theory Symposium at the 2013 IEEE Global Conference on
  Signal and Information Processing (GlobalSIP)\, and also served as a co-c
 hair of the 2012 IEEE European School of Information Theory (ESIT). His re
 search interests lie in the areas of communication theory and information 
 theory with special emphasis on joint source-channel coding\, multi-user n
 etworks\, energy efficient communications and security.
LOCATION:BN3-05 (Sigproc Meeting Room)\, 3rd Floor of CUED
END:VEVENT
END:VCALENDAR