Presenter: Mingqi Yang
Abstract: Reasoning about the robustness of probabilistic programs has become an increasingly important topic. One prominent line of work is formal analysis of probabilistic program sensitivity. Probabilistic program sensitivity, in this context, refers to bounding the distance between the output distributions of two runs of a probabilistic program with similar inputs. The goal is to provide formal upper bounds on such distances, thereby quantifying how much randomness and nondeterminism may amplify small input differences.