I will present new results concerning the approximation of the BV-norm by nonlocal, nonconvex, functionals. The original motivation comes from Image Processing. Numerous problems remain open. The talk is based on a joint work with H.-M. Nguyen.
Polar codes have recently emerged as a new class of low-complexity codes achieving Shannon capacity. This talk introduces polar codes with emphasis on the probabilistic phenomenon underlying the code construction. New results and connections to randomness extraction for structured sources are discussed.
I will survey some of the basic approaches to derandomizing Probabilistic Logspace computations, including the "classical" Nisan, Impagliazzo-Nisan-Widgerson and Reingold-Raz generators, the Saks-Zhou algorithm and some more recent approaches. We'll see why each falls short of complete derandomization, BPL=L, hopefully motivating further work on this basic problem.