Approximate Counting, the Lovasz Local Lemma and Inference in Graphical Models

TL;DR

This paper presents a novel algorithm for approximate counting and sampling in bounded degree systems with higher-order constraints, bridging a significant gap in theoretical bounds and extending to graphical models.

Contribution

The paper introduces a new approach for approximate counting and sampling in complex systems, leveraging a framework to compute marginal probabilities under Lovász Local Lemma conditions.

Findings

Algorithm efficiently counts solutions in logarithmic-width CNF formulas.

Extension to approximate sampling under similar conditions.

Application to sampling from posterior distributions in graphical models.

Abstract

In this paper we introduce a new approach for approximately counting in bounded degree systems with higher-order constraints. Our main result is an algorithm to approximately count the number of solutions to a CNF formula $\Phi$ when the width is logarithmic in the maximum degree. This closes an exponential gap between the known upper and lower bounds. Moreover our algorithm extends straightforwardly to approximate sampling, which shows that under Lov\'asz Local Lemma-like conditions it is not only possible to find a satisfying assignment, it is also possible to generate one approximately uniformly at random from the set of all satisfying assignments. Our approach is a significant departure from earlier techniques in approximate counting, and is based on a framework to bootstrap an oracle for computing marginal probabilities on individual variables. Finally, we give an application of our results to show that it is algorithmically possible to sample from the posterior distribution in an interesting class of graphical models.