arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2007.02559cs.LOcs.AIcs.LG
收藏

使用粘合变量预测增强SAT求解器

Enhancing SAT solvers with glue variable predictions

Amazon Web Services(亚马逊云科技);University of Pittsburgh(匹兹堡大学)

Jesse Michael Han

摘要:

现代SAT求解器通常在一定规模下运行,这使得为每个分支决策查询神经网络变得不切实际。由Selsam和Bjorner提出的NeuroCore提供了一个概念验证,即神经网络仍然可以通过定期重新聚焦基于分数的分支启发式来加速SAT求解器。然而,这项工作受到了一些限制:他们修改后的求解器需要GPU加速,进一步的消融表明它们并不比SATCOMP 2018基准测试中的随机基线更好,并且他们的非sat核心训练目标需要昂贵的数据管道,只能标记相对容易的无法满足的问题。我们解决了所有这些限制,使用一个更简单的网络架构,允许CPU推理甚至具有数百万个子句的大型工业问题,并训练预测{ \em胶水变量}--一个更容易生成标记数据的目标,也可以制定为强化学习任务。我们通过修改最先进的SAT求解器CaDiCaL,通过监督学习提高其在SATCOMP 2018和SATRACE 2019上的性能,以及通过强化学习提高其在SHA-1原像攻击数据集上的性能,来证明我们方法的有效性。

英文摘要:

Modern SAT solvers routinely operate at scales that make it impractical to query a neural network for every branching decision. NeuroCore, proposed by Selsam and Bjorner, offered a proof-of-concept that neural networks can still accelerate SAT solvers by only periodically refocusing a score-based branching heuristic. However, that work suffered from several limitations: their modified solvers require GPU acceleration, further ablations showed that they were no better than a random baseline on the SATCOMP 2018 benchmark, and their training target of unsat cores required an expensive data pipeline which only labels relatively easy unsatisfiable problems. We address all these limitations, using a simpler network architecture allowing CPU inference for even large industrial problems with millions of clauses, and training instead to predict {\em glue variables}---a target for which it is easier to generate labelled data, and which can also be formulated as a reinforcement learning task. We demonstrate the effectiveness of our approach by modifying the state-of-the-art SAT solver CaDiCaL, improving its performance on SATCOMP 2018 and SATRACE 2019 with supervised learning and its performance on a dataset of SHA-1 preimage attacks with reinforcement learning.

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。

↑