TacticZero: Learning to Prove Theorems from Scratch with Deep Reinforcement Learning

被引:0
|
作者
Wu, Minchao [1 ]
Norrish, Michael [2 ]
Walder, Christian [1 ]
Dezfouli, Amir [3 ]
机构
[1] Australian Natl Univ, CSIRO, Data61, Canberra, ACT, Australia
[2] Australian Natl Univ, Canberra, ACT, Australia
[3] CSIRO, Data61, Canberra, ACT, Australia
来源
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021) | 2021年
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We propose a novel approach to interactive theorem proving (ITP) using deep reinforcement learning. The proposed framework is able to learn proof search strategies as well as tactic and arguments prediction in an end-to-end manner. We formulate the process of ITP as a Markov decision process (MDP) in which each state represents a set of potential derivation paths. This structure allows us to introduce a search mechanism which enables the agent to efficiently discard (predicted) dead-end derivations and restart from promising alternatives. We implement the framework in the HOL4 theorem prover. Experimental results show that the framework using learned search strategies outperforms existing automated theorem provers (i.e. hammers) available in HOL4 when evaluated on unseen problems. We further elaborate the role of key components of the framework using ablation studies.
引用
收藏
页数:13
相关论文
共 50 条
  • [21] Learning Deep Ship Detector in SAR Images From Scratch
    Deng, Zhipeng
    Sun, Hao
    Zhou, Shilin
    Zhao, Juanping
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2019, 57 (06): : 4021 - 4039
  • [22] Deep Reinforcement Learning for Adaptive Learning Systems
    Li, Xiao
    Xu, Hanchen
    Zhang, Jinming
    Chang, Hua-hua
    JOURNAL OF EDUCATIONAL AND BEHAVIORAL STATISTICS, 2023, 48 (02) : 220 - 243
  • [23] Contrastive Learning Methods for Deep Reinforcement Learning
    Wang, Di
    Hu, Mengqi
    IEEE ACCESS, 2023, 11 : 97107 - 97117
  • [24] Transfer Learning in Deep Reinforcement Learning: A Survey
    Zhu, Zhuangdi
    Lin, Kaixiang
    Jain, Anil K.
    Zhou, Jiayu
    IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2023, 45 (11) : 13344 - 13362
  • [25] Learning an Index Advisor with Deep Reinforcement Learning
    Lai, Sichao
    Wu, Xiaoying
    Wang, Senyang
    Peng, Yuwei
    Peng, Zhiyong
    WEB AND BIG DATA, APWEB-WAIM 2021, PT II, 2021, 12859 : 178 - 185
  • [26] Learning to Walk via Deep Reinforcement Learning
    Haarnoja, Tuomas
    Ha, Sehoon
    Zhou, Aurick
    Tan, Jie
    Tucker, George
    Levine, Sergey
    ROBOTICS: SCIENCE AND SYSTEMS XV, 2019,
  • [27] Learning to Break Rocks With Deep Reinforcement Learning
    Samtani, Pavan
    Leiva, Francisco
    Ruiz-del-Solar, Javier
    IEEE ROBOTICS AND AUTOMATION LETTERS, 2023, 8 (02) : 1077 - 1084
  • [28] The Difficulty of Passive Learning in Deep Reinforcement Learning
    Ostrovski, Georg
    Castro, Pablo Samuel
    Dabney, Will
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021), 2021,
  • [29] Deep learning and reinforcement learning approach on microgrid
    Chandrasekaran, Kumar
    Kandasamy, Prabaakaran
    Ramanathan, Srividhya
    INTERNATIONAL TRANSACTIONS ON ELECTRICAL ENERGY SYSTEMS, 2020, 30 (10):
  • [30] Deep learning, reinforcement learning, and world models
    Matsuo, Yutaka
    LeCun, Yann
    Sahani, Maneesh
    Precup, Doina
    Silver, David
    Sugiyama, Masashi
    Uchibe, Eiji
    Morimoto, Jun
    NEURAL NETWORKS, 2022, 152 : 267 - 275