Distortion in Secure Communication with Side Information

Abstract

This project analyzes what the expected distortion at the adversary is for various achievability schemes involving side information. The main goal is to examine the average distortion incurred at the adversary with causal disclosure using the achievability scheme presented in Villard and Piantanida. The analysis makes use of a technique called channel encoding. The result is that there exists a threshold for which the adversary's distortion is maximal before the threshold, and after the threshold, the adversary's distortion is minimal. The analysis focuses on finding what this threshold is.