Learning the Satisfiability of Pseudo-Boolean Problem with Graph Neural Networks

2Citations
Citations of this article
3Readers
Mendeley users who have this article in their library.
Get full text

Abstract

Graph Neural Network (GNN) has shown great power on many practical tasks in the past few years. It is also considered to be a potential technique in bridging the gap between machine learning and symbolic reasoning. Experimental investigations have also shown that some-Hard constraint satisfaction problems can be well learned by the GNN models. In this paper, a GNN-based classification model to learn the satisfiability of pseudo-Boolean (PB) problem is proposed. After constructing the bipartite graph representation, a two-phase message passing process is executed. Experiments on 0–1 knapsack and weighted independent set problems show that the model can effectively learn the features related to the problem distribution and satisfiability. As a result, competitive prediction accuracy has been achieved with some generalization to larger-scale problems. The studies indicate that GNN has great potential in solving constraint satisfaction problems with numerical coefficients.

Cite

CITATION STYLE

APA

Liu, M., Zhang, F., Huang, P., Niu, S., Ma, F., & Zhang, J. (2020). Learning the Satisfiability of Pseudo-Boolean Problem with Graph Neural Networks. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 12333 LNCS, pp. 885–898). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-58475-7_51

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free