Title: Information Leakage in Index Coding

Abstract: We study the information leakage to a guessing adversary in index coding with a general message distribution. Under both vanishing-error and zero-error decoding assumptions, we develop lower and upper bounds on the optimal leakage rate, which are based on the broadcast rate of the subproblem induced by the set of messages the adversary tries to guess. When the messages are independent and uniformly distributed, the lower and upper bounds match, establishing an equivalence between the two rates.