HOList: An Environment for Machine Learning of Higher-Order Theorem Proving
Abstract
We present an environment, benchmark, and deep learning driven automated theorem prover for higher-order logic. Higher-order interactive theorem provers enable the formalization of arbitrary mathematical theories and thereby present an interesting, open-ended challenge for deep learning. We provide an open-source framework based on the HOL Light theorem prover that can be used as a reinforcement learning environment. HOL Light comes with a broad coverage of basic mathematical theorems on calculus and the formal proof of the Kepler conjecture, from which we derive a challenging benchmark for automated reasoning. We also present a deep reinforcement learning driven automated theorem prover, DeepHOL, with strong initial results on this benchmark.
Keywords
Cite
@article{arxiv.1904.03241,
title = {HOList: An Environment for Machine Learning of Higher-Order Theorem Proving},
author = {Kshitij Bansal and Sarah M. Loos and Markus N. Rabe and Christian Szegedy and Stewart Wilcox},
journal= {arXiv preprint arXiv:1904.03241},
year = {2019}
}
Comments
Accepted at ICML 2019