FIMO: A Challenge Formal Dataset for Automated Theorem Proving

Chengwu Liu, Jianhao Shen, Huajian Xin, Zhengying Liu, Ye Yuan +7 more
9/8/2023
cs.AI

No official code found. Best implementation: liuchengwucn/FIMO (38 stars, Apache-2.0, last commit 2024-10-29).

Abstract

We present FIMO, an innovative dataset comprising formal mathematical problem statements sourced from the International Mathematical Olympiad (IMO) Shortlisted Problems. Designed to facilitate advanced automated theorem proving at the IMO level, FIMO is currently tailored for the Lean formal language. It comprises 149 formal problem statements, accompanied by both informal problem descriptions and their corresponding LaTeX-based informal proofs. Through initial experiments involving GPT-4, our findings underscore the existing limitations in current methodologies, indicating a substantial journey ahead before achieving satisfactory IMO-level automated theorem proving outcomes.

View on arXivView PDF

Code Implementations(1)

liuchengwucn/FIMOCommunity100%
Apache-2.0
384LeanDec 3, 20231 years agoApache-2.0

Similar Papers

TheoremQA: A Theorem-driven Question Answering dataset

86% match2 repos1612023

Enhancing Formal Theorem Proving: A Comprehensive Dataset for Training AI Models on Coq Code

85% match2024

TheoremLlama: Transforming General-Purpose LLMs into Lean4 Experts

84% match2024

IndiMathBench: Autoformalizing Mathematical Reasoning Problems with a Human Touch

83% match1 repo42025

HorizonMath: Measuring AI Progress Toward Mathematical Discovery with Automatic Verification

83% match2026

A Survey on Deep Learning for Theorem Proving

83% match1 repo2292024

Cite this paper

@article{liu2023fimo,
  title  = {FIMO: A Challenge Formal Dataset for Automated Theorem Proving},
  author = {Chengwu Liu and Jianhao Shen and Huajian Xin and Zhengying Liu and Ye Yuan and Haiming Wang and Wei Ju and Chuanyang Zheng and Yichun Yin and Lin Li and Ming Zhang and Qun Liu},
  year   = {2023},
  eprint = {2309.04295},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2309.04295v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2309.04295.svg)](https://codeofpaper.com/paper/2309.04295)

Discussion