Propositional satisfiability (SAT) is an NP-complete problem that impacts many research fields, such as planning, verification, and security. Mainstream modern SAT solvers are based on the Conflict-Driven Clause Learning (CDCL) algorithm. Recent work aimed to enhance CDCL SAT solvers using Graph Neural Networks (GNNs). However, so far this approach either has not made solving more effective, or required substantial GPU resources for frequent online model inferences. Aiming to make GNN improvements practical, this paper proposes an approach called NeuroBack, which builds on two insights: (1) predicting phases (i.e., values) of variables appearing in the majority (or even all) of the satisfying assignments are essential for CDCL SAT solving, and (2) it is sufficient to query the neural model only once for the predictions before the SAT solving starts. Once trained, the offline model inference allows NeuroBack to execute exclusively on the CPU, removing its reliance on GPU resources. To train NeuroBack, a new dataset called DataBack containing 120,286 data samples is created. NeuroBack is implemented as an enhancement to a state-of-the-art SAT solver called Kissat. As a result, it allowed Kissat to solve up to 5.2% and 7.4% more problems on two recent SAT competition problem sets, SATCOMP-2022 and SATCOMP-2023, respectively. NeuroBack therefore shows how machine learning can be harnessed to improve SAT solving in an effective and practical manner.

PDF

In Proceedings of the International Conference on Learning Representations, 2024. (also arXiv:2110.14053).

@inproceedings{wang:iclr24, title={NeuroComb: Improving SAT Solving with Graph Neural Networks}, author={Wenxi Wang and Yang Hu and Mohit Tiwari and Sarfraz Khurshid and Kenneth McMillan and Risto Miikkulainen}, booktitle={Proceedings of the International Conference on Learning Representations}, month={ }, note={(also arXiv:2110.14053)}, url="http://nn.cs.utexas.edu/?wang:iclr24", year={2024} }