How can we trust the correctness of a learned model on a particular input of interest? Model accuracy is typically measured $on\ average$ over a distribution of inputs, giving no guarantee for any fixed input. This paper proposes a theoretically-founded solution to this problem: to train $Self$-$Proving\ models$ that prove ... more >>>

In this paper we introduce a new type of probabilistic search algorithm, which we call the
{\it Bellagio} algorithm: a probabilistic algorithm which is guaranteed to run in expected polynomial time,
and to produce a correct and {\it unique} solution with high probability.
We argue the applicability of such algorithms ... more >>>