Witsenhausen's counterexample and its links with multimedia security problems

Abstract

Witsenhausen's counterexample was proposed more than four decades ago in order to show that affine control strategies are not optimal for systems with non-classical information patterns. Finding the optimal solution to Witsenhausen's problem however remains an open problem. Recently, the stochastic control community has re-discovered Costa's Dirty Paper result as a potential solution to Witsenhausen's problem. In this paper the similarities and differences between Witsenhausen's scenario and multimedia security problems are reviewed, and the historical evolution of the solutions to Witsenhausen's problem compared with those proposed for watermarking detection. © 2012 Springer-Verlag.