When All Information is Not Created Equal

Following Shannon's landmark paper, classical theoretical framework forcommunication is based on a simplifying assumption that all information isequally important and aims to provide a uniform protection to allinformation. However, this homogeneous view of information is not suitablefor a variety of modern-day communication scenarios such as wireless andsensor networks, video transmission, interactive systems, and controlapplications. For example, an emergency alarm from a sensor network ismore important than other ordinary information it sends; or the coarseresolution of an image is more important than its finer details. For suchheterogeneous information, when providing a uniformly high protectionlevel to all parts of information is infeasible, it is desirable toprovide different protection levels based on the importance of these parts.The main objective of this talk is to extend the classical informationtheory for addressing this heterogeneous nature of information. Manytheoretical tools needed for this are fundamentally different from theconventional homogeneous setting. One key issue is that `bits' are no morea universal measure of information. We develop a general framework forunderstanding the fundamental limits of transmitting such information,calculate such fundamental limits, and provide optimal architectures forachieving these limits. Our analysis shows that even without sacrificingthe data-rate from channel capacity, some crucial parts of information canbe protected with exponential reliability. This research would challengethe notion of only restricting to a set of homogeneous bits as theuniversal interface to the physical layer, which would potentially impactthe design of network architectures.

Bio: Shashi Borade received his Ph.D. in 2008 and M.S. in 2004, both from the Department of Electrical Engineering and Computer Science at MIT. Earlier he received his B.Tech. degree in Electrical Engineering from IIT Bombay in 2002. He was a Hewlett-Packard Fellow at MIT during 2005-2007, a MIT Presidential Fellow during 2002-2003, and was awarded the Institute Silver Medal by IIT Bombay in 2002 for the best academic performance in Electrical Engineering. He has been a quantitative researcher at D. E. Shaw & Co. since finishing his doctoral studies. His research interests include information theory, statistics, and quantitative finance.