Multi-terminal communication strategies based on foundationally new approaches and code designs

Information Theory (IT) has served as a template to evolve efficient communication strategies and thereby shaped the architectural framework of today's communication Systems. While the code designs and tools developed by Shannon and enhanced over the years have yielded optimal performance limits for single-terminal systems - systems with a single Tx and single Rx, these code designs are not proven to be information theoretically optimal for several multi-terminal systems such as the broadcast and intereference channels. In fact, recent results indicate the IID random coding approach that has been long pursued in IT are sub-optimal. In this talk, I present two new frameworks for designing more efficient coding schemes. The first is based on algebraic structured codes and have proven to yield strictly larger rates for both broadcast and interference channels with more than two receivers. Next, I will present a new approach that leverages a fundamental undelying tension between law of large numbers and correlation to design a fixed block-length new coding scheme for communicating correlated sources over MAC. The broad goal is communicate to the audience the new and recent developments in network information theory over the last decade or so.