当前位置: X-MOL 学术J. Autom. Reason. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
TacticToe: Learning to Prove with Tactics
Journal of Automated Reasoning ( IF 1.1 ) Pub Date : 2020-08-20 , DOI: 10.1007/s10817-020-09580-x
Thibault Gauthier , Cezary Kaliszyk , Josef Urban , Ramana Kumar , Michael Norrish

We implement an automated tactical prover TacticToe on top of the HOL4 interactive theorem prover. TacticToe learns from human proofs which mathematical technique is suitable in each proof situation. This knowledge is then used in a Monte Carlo tree search algorithm to explore promising tactic-level proof paths. On a single CPU, with a time limit of 60 s, TacticToe proves 66.4% of the 7164 theorems in HOL4’s standard library, whereas E prover with auto-schedule solves 34.5%. The success rate rises to 69.0% by combining the results of TacticToe and E prover.

中文翻译:

TacticToe:学习用战术证明

我们在 HOL4 交互式定理证明器之上实现了一个自动战术证明器 TacticToe。TacticToe 从人类证明中学习哪种数学技术适用于每种证明情况。然后将这些知识用于蒙特卡罗树搜索算法,以探索有希望的战术级证明路径。在单个 CPU 上,时间限制为 60 秒,TacticToe 证明了 HOL4 标准库中 7164 个定理的 66.4%,而具有自动调度功能的 E 证明者解决了 34.5%。通过结合 TacticToe 和 E 证明者的结果,成功率上升到 69.0%。
更新日期:2020-08-20
down
wechat
bug