Stochastic Systems Group
		Home Group Members Publications Research Alumni Mailing List Demos Mission Statement Programs Seminars Library

Classical theoretical framework for communication trivializes the meaning of information. It assumes that all information is equally important and aims to provide a uniform protection to the entire information. However, some information is more important than other in many communication scenarios, such as wireless and sensor networks, internet, interactive systems, and control applications. In such scenarios, providing a uniform protection to all the information may be either wasteful or infeasible. Instead, it is more efficient to protect a crucial part of information better than the rest.

For a variety of such scenarios where information is an heterogeneous entity,

• We develop a general framework for understanding fundamental limits of communication. We also introduce some new notions of heterogeneous information, which combine hypothesis testing ideas with communication.
• Calculate such fundamental limits and provide optimal strategies for achieving them. Even for data-rates at channel capacity, it is shown that a crucial part of information can be protected with exponential reliability.

In addition to information theory, this investigation provides many new problems in network optimization, coding theory and signal processing. Finally some other approaches to this problem are discussed briefly, which show how ideas from graphical models and differential geometry can help in information theory.

Problems with this site should be emailed to jonesb@mit.edu