arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.03345math.HOcs.AIcs.CCmath.LO
收藏

人工智能与数学的内在困难性

Artifical intelligence and inherent mathematical difficulty

University of Warwick(华威大学);Université Paris 1 Panthéon-Sorbonne(巴黎第一大学先贤祠-索邦大学)

Walter Dean, Alberto Naibo

AI总结:

本文通过更新可计算性与复杂性理论的限制性论证,并结合自动定理证明、SAT求解器和大语言模型等AI方法,论证数学证明发现本质上困难,AI技术只是蛮力搜索,只能处理低逻辑复杂度命题。

摘要:

本文探讨了人工智能与解决数学中未决问题这一任务之间的关系。我们首先呈现了一个传统论证的更新版本,该论证表明,来自可计算性理论和复杂性理论的限制性结果表明,证明发现本质上是一个困难的问题。然后,我们阐述了人工智能启发方法的若干近期应用——分别涉及自动定理证明、SAT求解器和大语言模型——确实提出了关于数学证明本质的新问题。我们还论证,这些技术所获得的结果并不反驳我们的基本论证。之所以如此,是因为它们是蛮力搜索的体现,因此只能判定逻辑复杂度较低的命题。

英文摘要:

This paper explores the relationship of artificial intelligence to the task of resolving open questions in mathematics. We first present an updated version of a traditional argument that limitative results from computability and complexity theory show that proof discovery is an inherently difficult problem. We then illustrate how several recent applications of artificial intelligence-inspired methods -- respectively involving automated theorem proving, SAT-solvers, and large language models -- do indeed raise novel questions about the nature of mathematical proof. We also argue that the results obtained by such techniques do not tell against our basic argument. This is so because they are embodiments of brute force search and are thus capable of deciding only statements of low logical complexity.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑