Title:  Covertness and Learning in Information-Theoretic Security

Committee: 

Dr. Bloch, Advisor        

Dr. Barry, Chair

Dr. Romberg

Abstract:

The objective of the proposed research is to study covert communications and the possibility of learning in the context of information-theoretic security. We first establish results pertaining to the refined asymptotics of covert communications under three different metrics to measure covertness of a communication: Kullback-Leibler divergence and variational distance between the probability distributions induced without and with communication, as well as the optimal probability of missed detection at a fixed probability of false alarm. We also study the fundamental limits of covert secret key generation, in which two parties attempt to generate a secret key from correlated observations and communications over a public channel without being detected, in the classical and quantum regime. In the context of learning, we propose a new model, in which the transmitter can sense a noisy signal correlated with the adversary’s actions. We prove that the achievable rates, defined precisely for the problem, are arbitrarily close to those obtained with hindsight, had the transmitter known the actions ahead of time.