The Capacity of Causal Adversarial Channels

Explore a one-hour lecture by Prof. Sidharth Jaggi from the University of Bristol's School of Mathematics, delving into the capacity characterization of discrete-time arbitrarily varying channels. Learn about the complex interplay between encoders and decoders without common randomness, cost-constrained inputs and states, and deterministic transition matrices in channel communication. Discover innovative strategies including stochastic encoding with list decoding and the two-phase "babble-and-push" approach for channel symmetrization. Understand how these findings extend previous research on specific channel models to broader discrete alphabets and models. Drawing from his extensive background in electrical engineering and information theory, Prof. Jaggi, who leads the CAN-DO-IT research team, brings expertise from his journey through IIT Bombay, CalTech, MIT, Chinese University of Hong Kong, and the University of Bristol to explain these advanced concepts in network information theory, coding theory, and algorithms.