AITOP

proof solving

共 1 条相关 AI 资讯
8月27日
10:25
10:25官方账号arXiv cs.AI@Fredrik Rømming, Mantas Bakšys, Martin S. Fixman, Sean B. Holden
An automated theorem prover uses graph neural networks for policy learning in connection-tableau construction, solving up to 46% more problems than leanCoP with fewer steps. Trained via imitation learning from existing proofs, the model achieves significant performance improvements on M2k, MPTP2078-bushy, and TPTP v9.2.1 benchmarks.
论文graph neural networkpolicy learningautomated theorem proving
推荐理由：Check out this research on using graph neural networks for theorem proving. It outperforms leanCoP by solving more problems in fewer steps, thanks to policy learning from existing proofs.
原文