GraSS: Combining Graph Neural Networks with Expert Knowledge for SAT Solver Selection

• URL: http://arxiv.org/abs/2405.11024v1
• Date: Fri, 17 May 2024 18:00:50 GMT
• Title: GraSS: Combining Graph Neural Networks with Expert Knowledge for SAT Solver Selection
• Authors: Zhanguang Zhang, Didier Chetelat, Joseph Cotnareanu, Amur Ghose, Wenyi Xiao, Hui-Ling Zhen, Yingxue Zhang, Jianye Hao, Mark Coates, Mingxuan Yuan,
• Abstract summary: We present GraSS, a novel approach for automatic SAT solver selection based on tripartite graph representations of instances. We enrich the graph representation with domain-specific decisions, such as novel node feature design. We demonstrate that this combination of raw representations and domain-specific choices leads to improvements in runtime.
• Score: 45.222591894755105
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Boolean satisfiability (SAT) problems are routinely solved by SAT solvers in real-life applications, yet solving time can vary drastically between solvers for the same instance. This has motivated research into machine learning models that can predict, for a given SAT instance, which solver to select among several options. Existing SAT solver selection methods all rely on some hand-picked instance features, which are costly to compute and ignore the structural information in SAT graphs. In this paper we present GraSS, a novel approach for automatic SAT solver selection based on tripartite graph representations of instances and a heterogeneous graph neural network (GNN) model. While GNNs have been previously adopted in other SAT-related tasks, they do not incorporate any domain-specific knowledge and ignore the runtime variation introduced by different clause orders. We enrich the graph representation with domain-specific decisions, such as novel node feature design, positional encodings for clauses in the graph, a GNN architecture tailored to our tripartite graphs and a runtime-sensitive loss function. Through extensive experiments, we demonstrate that this combination of raw representations and domain-specific choices leads to improvements in runtime for a pool of seven state-of-the-art solvers on both an industrial circuit design benchmark, and on instances from the 20-year Anniversary Track of the 2022 SAT Competition.

Related papers

• Machine Learning for SAT: Restricted Heuristics and New Graph Representations [0.8870188183999854]
  SAT is a fundamental NP-complete problem with many applications, including automated scheduling. To solve large instances, SAT solvers have to rely on Booleans, e.g., choosing a branching variable in DPLL and CDCL solvers. We suggest a strategy of making a few initial steps with a trained ML model and then releasing control to classical runtimes.
  arXiv  Detail & Related papers  (2023-07-18T10:46:28Z)
• Addressing Variable Dependency in GNN-based SAT Solving [19.38746341365531]
  We propose AsymSAT, a GNN-based architecture which integrates recurrent neural networks to generate dependent predictions for variable assignments. Experiment results show that dependent variable prediction extends the solving capability of the GNN-based method as it improves the number of solved SAT instances on large test sets.
  arXiv  Detail & Related papers  (2023-04-18T05:33:33Z)
• W2SAT: Learning to generate SAT instances from Weighted Literal Incidence Graphs [13.173307471333619]
  W2SAT is a framework to generate SAT formulas by learning intrinsic structures and properties from given real-world/industrial instances. We introduce a novel SAT representation called Weighted Literal Incidence Graph (WLIG), which exhibits strong representation ability and generalizability. Decoding from WLIG into SAT problems is then modeled as finding overlapping cliques with a novel hill-climbing optimization method.
  arXiv  Detail & Related papers  (2023-02-01T06:30:41Z)
• Estimating the hardness of SAT encodings for Logical Equivalence Checking of Boolean circuits [58.83758257568434]
  We show that the hardness of SAT encodings for LEC instances can be estimated textitw.r.t some SAT partitioning. The paper proposes several methods for constructing partitionings, which, when used in practice, allow one to estimate the hardness of SAT encodings for LEC with good accuracy.
  arXiv  Detail & Related papers  (2022-10-04T09:19:13Z)
• DeepSAT: An EDA-Driven Learning Framework for SAT [9.111341161918375]
  We present DeepSAT, a novel end-to-end learning framework for the Boolean satisfiability (SAT) problem. DeepSAT achieves significant accuracy improvements over state-of-the-art learning-based SAT solutions.
  arXiv  Detail & Related papers  (2022-05-27T03:20:42Z)
• Machine Learning Methods in Solving the Boolean Satisfiability Problem [72.21206588430645]
  The paper reviews the recent literature on solving the Boolean satisfiability problem (SAT) with machine learning techniques. We examine the evolving ML-SAT solvers from naive classifiers with handcrafted features to the emerging end-to-end SAT solvers such as NeuroSAT.
  arXiv  Detail & Related papers  (2022-03-02T05:14:12Z)
• Transformer-based Machine Learning for Fast SAT Solvers and Logic Synthesis [63.53283025435107]
  CNF-based SAT and MaxSAT solvers are central to logic synthesis and verification systems. In this work, we propose a one-shot model derived from the Transformer architecture to solve the MaxSAT problem.
  arXiv  Detail & Related papers  (2021-07-15T04:47:35Z)
• Learning on Attribute-Missing Graphs [66.76561524848304]
  There is a graph where attributes of only partial nodes could be available and those of the others might be entirely missing. Existing graph learning methods including the popular GNN cannot provide satisfied learning performance. We develop a novel distribution matching based GNN called structure-attribute transformer (SAT) for attribute-missing graphs.
  arXiv  Detail & Related papers  (2020-11-03T11:09:52Z)
• Evaluating Logical Generalization in Graph Neural Networks [59.70452462833374]
  We study the task of logical generalization using graph neural networks (GNNs) Our benchmark suite, GraphLog, requires that learning algorithms perform rule induction in different synthetic logics. We find that the ability for models to generalize and adapt is strongly determined by the diversity of the logical rules they encounter during training.
  arXiv  Detail & Related papers  (2020-03-14T05:45:55Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.