Variational Inference on the Boolean Hypercube with the Quantum Entropy

Eliot Beyler, Francis Bach

TL;DR: Recent developments in the approximation of Kullback-Leibler divergence provide an opportunity to improve variational inference.

Abstract: In this paper, we derive variational inference upper-bounds on the log-partition function of a pairwize Markov random fields on the Boolean hypercube, based on quantum relaxations of the Kullback-Leibler divergence. We then propose an efficient algorithm to compute these bounds based on primal-dual optimization. An improvement of these bounds through the use of "hierarchies", similar to sum-of-squares (SoS) hierarchies is proposed, and we present a greedy algorithm to select among these relaxations. We carry extensive numerical experiments and compare with state-of-the-art methods for this inference problem.

Submission Number: 394