Cargando…

Peano: learning formal mathematical reasoning

General mathematical reasoning is computationally undecidable, but humans routinely solve new problems. Moreover, discoveries developed over centuries are taught to subsequent generations quickly. What structure enables this, and how might that inform automated mathematical reasoning? We posit that...

Descripción completa

Detalles Bibliográficos
Autores principales: Poesia, Gabriel, Goodman, Noah D.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: The Royal Society 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10239677/
https://www.ncbi.nlm.nih.gov/pubmed/37271179
http://dx.doi.org/10.1098/rsta.2022.0044
_version_ 1785053541224677376
author Poesia, Gabriel
Goodman, Noah D.
author_facet Poesia, Gabriel
Goodman, Noah D.
author_sort Poesia, Gabriel
collection PubMed
description General mathematical reasoning is computationally undecidable, but humans routinely solve new problems. Moreover, discoveries developed over centuries are taught to subsequent generations quickly. What structure enables this, and how might that inform automated mathematical reasoning? We posit that central to both puzzles is the structure of procedural abstractions underlying mathematics. We explore this idea in a case study on five sections of beginning algebra on the Khan Academy platform. To define a computational foundation, we introduce Peano, a theorem-proving environment where the set of valid actions at any point is finite. We use Peano to formalize introductory algebra problems and axioms, obtaining well-defined search problems. We observe existing reinforcement learning methods for symbolic reasoning to be insufficient to solve harder problems. Adding the ability to induce reusable abstractions (‘tactics’) from its own solutions allows an agent to make steady progress, solving all problems. Furthermore, these abstractions induce an order to the problems, seen at random during training. The recovered order has significant agreement with the expert-designed Khan Academy curriculum, and second-generation agents trained on the recovered curriculum learn significantly faster. These results illustrate the synergistic role of abstractions and curricula in the cultural transmission of mathematics. This article is part of a discussion meeting issue ‘Cognitive artificial intelligence’.
format Online
Article
Text
id pubmed-10239677
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher The Royal Society
record_format MEDLINE/PubMed
spelling pubmed-102396772023-06-05 Peano: learning formal mathematical reasoning Poesia, Gabriel Goodman, Noah D. Philos Trans A Math Phys Eng Sci Articles General mathematical reasoning is computationally undecidable, but humans routinely solve new problems. Moreover, discoveries developed over centuries are taught to subsequent generations quickly. What structure enables this, and how might that inform automated mathematical reasoning? We posit that central to both puzzles is the structure of procedural abstractions underlying mathematics. We explore this idea in a case study on five sections of beginning algebra on the Khan Academy platform. To define a computational foundation, we introduce Peano, a theorem-proving environment where the set of valid actions at any point is finite. We use Peano to formalize introductory algebra problems and axioms, obtaining well-defined search problems. We observe existing reinforcement learning methods for symbolic reasoning to be insufficient to solve harder problems. Adding the ability to induce reusable abstractions (‘tactics’) from its own solutions allows an agent to make steady progress, solving all problems. Furthermore, these abstractions induce an order to the problems, seen at random during training. The recovered order has significant agreement with the expert-designed Khan Academy curriculum, and second-generation agents trained on the recovered curriculum learn significantly faster. These results illustrate the synergistic role of abstractions and curricula in the cultural transmission of mathematics. This article is part of a discussion meeting issue ‘Cognitive artificial intelligence’. The Royal Society 2023-07-24 2023-06-05 /pmc/articles/PMC10239677/ /pubmed/37271179 http://dx.doi.org/10.1098/rsta.2022.0044 Text en © 2023 The Authors. https://creativecommons.org/licenses/by/4.0/Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/ (https://creativecommons.org/licenses/by/4.0/) , which permits unrestricted use, provided the original author and source are credited.
spellingShingle Articles
Poesia, Gabriel
Goodman, Noah D.
Peano: learning formal mathematical reasoning
title Peano: learning formal mathematical reasoning
title_full Peano: learning formal mathematical reasoning
title_fullStr Peano: learning formal mathematical reasoning
title_full_unstemmed Peano: learning formal mathematical reasoning
title_short Peano: learning formal mathematical reasoning
title_sort peano: learning formal mathematical reasoning
topic Articles
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10239677/
https://www.ncbi.nlm.nih.gov/pubmed/37271179
http://dx.doi.org/10.1098/rsta.2022.0044
work_keys_str_mv AT poesiagabriel peanolearningformalmathematicalreasoning
AT goodmannoahd peanolearningformalmathematicalreasoning