English

Process-Driven Autoformalization in Lean 4

Computation and Language 2024-10-15 v2 Machine Learning Logic in Computer Science

Abstract

Autoformalization, the conversion of natural language mathematics into formal languages, offers significant potential for advancing mathematical reasoning. However, existing efforts are limited to formal languages with substantial online corpora and struggle to keep pace with rapidly evolving languages like Lean 4. To bridge this gap, we propose a new benchmark \textbf{Form}alization for \textbf{L}ean~\textbf{4} (\textbf{\name}) designed to evaluate the autoformalization capabilities of large language models (LLMs). This benchmark encompasses a comprehensive assessment of questions, answers, formal statements, and proofs. Additionally, we introduce a \textbf{P}rocess-\textbf{S}upervised \textbf{V}erifier (\textbf{PSV}) model that leverages the precise feedback from Lean 4 compilers to enhance autoformalization. Our experiments demonstrate that the PSV method improves autoformalization, enabling higher accuracy using less filtered training data. Furthermore, when fine-tuned with data containing detailed process information, PSV can leverage the data more effectively, leading to more significant improvements in autoformalization for Lean 4. Our dataset and code are available at \url{https://github.com/rookie-joe/PDA}.

Keywords

Cite

@article{arxiv.2406.01940,
  title  = {Process-Driven Autoformalization in Lean 4},
  author = {Jianqiao Lu and Yingjia Wan and Zhengying Liu and Yinya Huang and Jing Xiong and Chengwu Liu and Jianhao Shen and Hui Jin and Jipeng Zhang and Haiming Wang and Zhicheng Yang and Jing Tang and Zhijiang Guo},
  journal= {arXiv preprint arXiv:2406.01940},
  year   = {2024}
}

Comments

32 pages, 1 figures, 15 tables

R2 v1 2026-06-28T16:52:20.079Z