arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2205.04423cs.AIcs.LG
收藏

用于命题模型计数的图神经网络

Graph Neural Networks for Propositional Model Counting

University of Pisa(比萨大学);University of Trieste(的里雅斯特大学)

Gaia Saveri, Luca Bortolussi

摘要:

图神经网络(GNN)最近被用来解决几个逻辑推理任务。然而,计数问题,如命题模型计数(#SAT)仍然主要是传统的解决方案。在这里,我们通过提出一个基于Kuch等人的GNN框架的置信传播(BP)架构来解决这个差距,用自我关注的GNN扩展,并训练近似解决#SAT问题。我们进行了彻底的实验研究,结果表明,我们的模型在一小组随机布尔公式上训练,能够有效地扩展到更大的问题规模,具有与最先进的近似求解器相当或更好的性能。此外,我们表明,它可以有效地微调,以提供良好的推广结果不同的公式分布,如那些来自SAT编码的组合问题。

英文摘要:

Graph Neural Networks (GNNs) have been recently leveraged to solve several logical reasoning tasks. Nevertheless, counting problems such as propositional model counting (#SAT) are still mostly approached with traditional solvers. Here we tackle this gap by presenting an architecture based on the GNN framework for belief propagation (BP) of Kuch et al., extended with self-attentive GNN and trained to approximately solve the #SAT problem. We ran a thorough experimental investigation, showing that our model, trained on a small set of random Boolean formulae, is able to scale effectively to much larger problem sizes, with comparable or better performances of state of the art approximate solvers. Moreover, we show that it can be efficiently fine-tuned to provide good generalization results on different formulae distributions, such as those coming from SAT-encoded combinatorial problems.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑