Abstract
In this work we study how to learn good algorithms for selecting reasoning steps in theorem proving. We explore this in the connection tableau calculus implemented by leanCoP where the partial tableau provides a clean and compact notion of a state to which a limited number of inferences can be applied. We start by incorporating a state-of-the-art learning algorithm — a graph neural network (GNN) – into the plCoP theorem prover. Then we use it to observe the system’s behavior in a reinforcement learning setting, i.e., when learning inference guidance from successful Monte-Carlo tree searches on many problems. Despite its better pattern matching capability, the GNN initially performs worse than a simpler previously used learning algorithm. We observe that the simpler algorithm is less confident, i.e., its recommendations have higher entropy. This leads us to explore how the entropy of the inference selection implemented via the neural network influences the proof search. This is related to research in human decision-making under uncertainty, and in particular the probability matching theory. Our main result shows that a proper entropy regularization, i.e., training the GNN not to be overconfident, greatly improves plCoP ’s performance on a large mathematical corpus.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Assuming Disjunctive Normal Form.
- 2.
In this sense, theorem proving can be considered as a meta learning task.
- 3.
A discount factor of 0.99 is applied to positive rewards to favor shorter proofs.
- 4.
This is motivated by the experiments with the ENIGMA-GNN system [18], where 8–10 layers produce better results than 5 layers.
- 5.
The new extensions described here and the experimental configuration files are publicly available at plCoP ’s repository: https://github.com/zsoltzombori/plcop.
- 6.
X and G stand for the probability distributions predicted by XGBoost and GNN, respectively.
References
Alama, J., Heskes, T., Kühlwein, D., Tsivtsivadze, E., Urban, J.: Premise selection for mathematics by corpus analysis and kernel methods. J. Autom. Reason. 52(2), 191–213 (2014)
Alemi, A.A., Chollet, F., Een, N., Irving, G., Szegedy, C., Urban, J.: DeepMath - deep sequence models for premise selection. In: Proceedings of the 30th International Conference on Neural Information Processing Systems, NIPS 2016, USA, pp. 2243–2251. Curran Associates Inc. (2016)
Anthony, T., Tian, Z., Barber, D.: Thinking fast and slow with deep learning and tree search. CoRR, abs/1705.08439 (2017)
Bansal, K., Loos, S.M., Rabe, M.N., Szegedy, C., Wilcox, S.: HOList: an environment for machine learning of higher order logic theorem proving. In: Chaudhuri, K., Salakhutdinov, R. (eds.) International Conference on Machine Learning, ICML 2019. Proceedings of Machine Learning Research, vol. 97, pp. 454–463. PMLR (2019)
Bibel, W.: A vision for automated deduction rooted in the connection method. In: Schmidt, R.A., Nalon, C. (eds.) TABLEAUX 2017. LNCS (LNAI), vol. 10501, pp. 3–21. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-66902-1_1
Blaauwbroek, L., Urban, J., Geuvers, H.: Tactic learning and proving for the Coq proof assistant. In: Albert, E., Kovács, L. (eds.) LPAR 2020: 23rd International Conference on Logic for Programming, Artificial Intelligence and Reasoning. EPiC Series in Computing, vol. 73, pp. 138–150. EasyChair (2020)
Browne, C., et al.: A survey of Monte Carlo tree search methods. IEEE Trans. Comput. Intell. AI Games 4, 1–43 (2012)
Chen, T., Guestrin, C.: XGBoost: a scalable tree boosting system. In: Proceedings of the 22Nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2016, New York, NY, USA, pp. 785–794. ACM (2016)
Chvalovský, K., Jakubův, J., Suda, M., Urban, J.: ENIGMA-NG: efficient neural and gradient-boosted inference guidance for E. In: Fontaine, P. (ed.) CADE 2019. LNCS (LNAI), vol. 11716, pp. 197–215. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-29436-6_12
The Coq Proof Assistant. http://coq.inria.fr
Crouse, M., et al.: A deep reinforcement learning approach to first-order logic theorem proving. Artificial Intelligence (2019). arXiv
Färber, M., Kaliszyk, C., Urban, J.: Machine learning guidance for connection tableaux. J. Autom. Reason. 65(2), 287–320 (2021)
Gauthier, T., Kaliszyk, C., Urban, J.: TacticToe: learning to reason with HOL4 tactics. In: Eiter, T., Sands, D. (eds.) LPAR-21. 21st International Conference on Logic for Programming, Artificial Intelligence and Reasoning. EPiC Series in Computing, vol. 46, pp. 125–143. EasyChair (2017)
Gauthier, T., Kaliszyk, C., Urban, J., Kumar, R., Norrish, M.: TacticToe: learning to prove with tactics. J. Autom. Reason. 65(2), 257–286 (2021)
Gordon, M.J.C., Melham, T.F. (eds.): Introduction to HOL: A Theorem Proving Environment for Higher Order Logic. Cambridge University Press, Cambridge (1993)
Harrison, J.: HOL light: a tutorial introduction. In: Srivas, M., Camilleri, A. (eds.) FMCAD 1996. LNCS, vol. 1166, pp. 265–269. Springer, Heidelberg (1996). https://doi.org/10.1007/BFb0031814
Huang, D., Dhariwal, P., Song, D., Sutskever, I.: Gamepad: a learning environment for theorem proving. In: 7th International Conference on Learning Representations, ICLR 2019. OpenReview.net (2019)
Jakubův, J., Chvalovský, K., Olšák, M., Piotrowski, B., Suda, M., Urban, J.: ENIGMA anonymous: symbol-independent inference guiding machine (system description). In: Peltier, N., Sofronie-Stokkermans, V. (eds.) IJCAR 2020. LNCS (LNAI), vol. 12167, pp. 448–463. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-51054-1_29
Jakubův, J., Urban, J.: ENIGMA: efficient learning-based inference guiding machine. In: Geuvers, H., England, M., Hasan, O., Rabe, F., Teschke, O. (eds.) CICM 2017. LNCS (LNAI), vol. 10383, pp. 292–302. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-62075-6_20
Jakubuv, J., Urban, J.: Hammering Mizar by learning clause guidance. In: Harrison, J., O’Leary, J., Tolmach, A. (eds.) 10th International Conference on Interactive Theorem Proving, ITP 2019, Portland, OR, USA, 9–12 September 2019. LIPIcs, vol. 141, pp. 34:1–34:8. Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2019)
Kaliszyk, C., Urban, J.: M2K dataset. https://github.com/JUrban/deepmath/blob/master/M2k_list
Kaliszyk, C., Urban, J.: Mizar40 dataset. https://github.com/JUrban/deepmath
Kaliszyk, C., Urban, J.: FEMaLeCoP: fairly efficient machine learning connection prover. In: Davis, M., Fehnker, A., McIver, A., Voronkov, A. (eds.) LPAR 2015. LNCS, vol. 9450, pp. 88–96. Springer, Heidelberg (2015). https://doi.org/10.1007/978-3-662-48899-7_7
Kaliszyk, C., Urban, J.: MizAR 40 for Mizar 40. J. Autom. Reason. 55(3), 245–256 (2015). https://doi.org/10.1007/s10817-015-9330-8
Kaliszyk, C., Urban, J., Michalewski, H., Olsák, M.: Reinforcement learning of theorem proving. In: NeurIPS, pp. 8836–8847 (2018)
Kaliszyk, C., Urban, J., Vyskočil, J.: Efficient semantic features for automated reasoning over large theories. In: Yang, Q., Wooldridge, M. (eds.) IJCAI 2015, pp. 3084–3090. AAAI Press (2015)
Kocsis, L., Szepesvári, C.: Bandit based Monte-Carlo planning. In: Fürnkranz, J., Scheffer, T., Spiliopoulou, M. (eds.) ECML 2006. LNCS (LNAI), vol. 4212, pp. 282–293. Springer, Heidelberg (2006). https://doi.org/10.1007/11871842_29
Kovács, L., Voronkov, A.: First-order theorem proving and Vampire. In: Sharygina, N., Veith, H. (eds.) CAV 2013. LNCS, vol. 8044, pp. 1–35. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-39799-8_1
Loos, S.M., Irving, G., Szegedy, C., Kaliszyk, C.: Deep network guided proof search. In: 21st International Conference on Logic for Programming, Artificial Intelligence, and Reasoning (LPAR) (2017)
Mohamed, O.A., Muñoz, C., Tahar, S. (eds.): TPHOLs 2008. LNCS, vol. 5170. Springer, Heidelberg (2008). https://doi.org/10.1007/978-3-540-71067-7
Nagashima, Y., He, Y.: PaMpeR: proof method recommendation system for Isabelle/HOL. In: Huchard, M., Kästner, C., Fraser, G. (eds.) Proceedings of the 33rd ACM/IEEE International Conference on Automated Software Engineering, ASE 2018, Montpellier, France, 3–7 September 2018, pp. 362–372. ACM (2018)
Olsák, M., Kaliszyk, C., Urban, J.: Property invariant embedding for automated reasoning. In: Giacomo, G.D., et al. (eds.) ECAI 2020–24th European Conference on Artificial Intelligence, Santiago de Compostela, Spain, 29 August-8 September 2020 - Including 10th Conference on Prestigious Applications of Artificial Intelligence (PAIS 2020). Frontiers in Artificial Intelligence and Applications, vol. 325, pp. 1395–1402. IOS Press (2020)
Otten, J.: leanCoP 2.0 and ileanCoP 1.2: high performance lean theorem proving in classical and intuitionistic logic (system descriptions). In: Armando, A., Baumgartner, P., Dowek, G. (eds.) IJCAR 2008. LNCS (LNAI), vol. 5195, pp. 283–291. Springer, Heidelberg (2008). https://doi.org/10.1007/978-3-540-71070-7_23
Otten, J., Bibel, W.: leanCoP: lean connection-based theorem proving. J. Symb. Comput. 36, 139–161 (2003)
Rawson, M., Reger, G.: Automated theorem proving, fast and slow. EasyChair Preprint no. 4433. EasyChair (2020)
Ross, S., Gordon, G., Bagnell, D.: A reduction of imitation learning and structured prediction to no-regret online learning. In: Gordon, G., Dunson, D., Dudík, M. (eds.) Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics. Proceedings of Machine Learning Research, Fort Lauderdale, FL, USA, 11–13 April 2011, vol. 15, pp. 627–635. PMLR (2011)
Scarselli, F., Gori, M., Tsoi, A.C., Hagenbuchner, M., Monfardini, G.: The graph neural network model. Trans. Neur. Netw. 20(1), 61–80 (2009)
Schulz, S.: E - a Brainiac theorem prover. AI Commun. 15(2–3), 111–126 (2002)
Schulz, S.: System description: E 1.8. In: McMillan, K., Middeldorp, A., Voronkov, A. (eds.) LPAR 2013. LNCS, vol. 8312, pp. 735–743. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-45221-5_49
Silver, D., et al.: Mastering the game of go without human knowledge. Nature 550, 354 (2017)
Slind, K., Norrish, M.: A brief overview of HOL4. In: Mohamed et al. [30], pp. 28–32
Suda, M.: New techniques that improve Enigma-style clause selection guidance. In: International Conference on Automated Deduction, CADE 2021 (2021)
Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction, 2nd edn. The MIT Press, Cambridge (2018)
Urban, J.: ERC project AI4Reason final scientific report (2021). http://grid01.ciirc.cvut.cz/~mptp/ai4reason/PR_CORE_SCIENTIFIC_4.pdf
Urban, J., Sutcliffe, G., Pudlák, P., Vyskočil, J.: MaLARea SG1 - machine learner for automated reasoning with semantic guidance. In: IJCAR, pp. 441–456 (2008)
Urban, J., Vyskočil, J., Štěpánek, P.: MaLeCoP machine learning connection prover. In: Brünnler, K., Metcalfe, G. (eds.) TABLEAUX 2011. LNCS (LNAI), vol. 6793, pp. 263–277. Springer, Heidelberg (2011). https://doi.org/10.1007/978-3-642-22119-4_21
Vulkan, N.: An economist’s perspective on probability matching. J. Econ. Surv. 14(1), 101–118 (2000)
Wenzel, M., Paulson, L.C., Nipkow, T.: The Isabelle framework. In: Mohamed et al. [30], pp. 33–38
Williams, R.J., Peng, J.: Function optimization using connectionist reinforcement learning algorithms. Connect. Sci. 3(3), 241–268 (1991)
Yang, K., Deng, J.: Learning to prove theorems via interacting with proof assistants. In: Chaudhuri, K., Salakhutdinov, R. (eds.) Proceedings of the 36th International Conference on Machine Learning, ICML 2019, Long Beach, California, USA, 9–15 June 2019. Proceedings of Machine Learning Research, vol. 97, pp. 6984–6994. PMLR (2019)
Zombori, Z., Urban, J., Brown, C.E.: Prolog technology reinforcement learning prover. In: Peltier, N., Sofronie-Stokkermans, V. (eds.) IJCAR 2020. LNCS (LNAI), vol. 12167, pp. 489–507. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-51054-1_33
Acknowledgments
ZZ was supported by the European Union, co-financed by the European Social Fund (EFOP-3.6.3-VEKOP-16-2017-00002), the Hungarian National Excellence Grant 2018-1.2.1-NKP-00008 and by the Hungarian Ministry of Innovation and Technology NRDI Office within the framework of the Artificial Intelligence National Laboratory Program. JU was funded by the AI4REASON ERC Consolidator grant nr. 649043 and the European Regional Development Fund under the Czech project AI&Reasoning CZ.02.1.01/0.0/0.0/15_003/0000466. MO was supported by the ERC starting grant no. 714034 SMART. We thank the TABLEAUX’21 reviewers for their thoughtful reviews and comments.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Zombori, Z., Urban, J., Olšák, M. (2021). The Role of Entropy in Guiding a Connection Prover. In: Das, A., Negri, S. (eds) Automated Reasoning with Analytic Tableaux and Related Methods. TABLEAUX 2021. Lecture Notes in Computer Science(), vol 12842. Springer, Cham. https://doi.org/10.1007/978-3-030-86059-2_13
Download citation
DOI: https://doi.org/10.1007/978-3-030-86059-2_13
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-86058-5
Online ISBN: 978-3-030-86059-2
eBook Packages: Computer ScienceComputer Science (R0)