paper-with-me

홈 › Papers

Formalizing Numerical Analysis: An Agent Pipeline and Quality Audit Beyond Kernel Acceptance

2026-06-12 · Theodore Meek, Siyuan Ge, Di Qiu Xiang, Simon Chess, Vasily Ilin arxiv

Recent work has demonstrated that coding agents can formalize entire advanced mathematics textbooks in Lean 4, yet existing efforts concentrate on branches of mathematics already well-represented in mathlib and measure success solely through kernel acceptance. We address both limitations by applying a coding agent to formalize Numerical Methods for Ordinary Differential Equations, a textbook in numerical analysis that is largely absent from mathlib, stressing the agent's capacity to develop new theory from scratch. We further introduce a systematic, reproducible three-dimensional framework for evaluating the quality of agent-produced formalizations beyond compilation: semantic correctness, Mathlib reuse, and cross-file reuse via LLM-as-judge methods. Applying this framework to our own formalization and to the released outputs of RepoProver and M2F, we uncover recurring unfaithful formalization patterns, including incomplete multi-part statements, added weakening hypotheses, and parameter restrictions, that kernel acceptance entirely obscures. Our results suggest that compilation-based metrics substantially overstate formalization quality, and we provide a reproducible audit methodology to support more rigorous evaluation of future autoformalization systems.

📄 PDF Abstract BibTeX arXiv:2606.14000

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

GTA: Generating Long-Horizon Tasks for Web Agents at Scale

2026-05-28 · Tenghao Huang, Kung-Hsiang Huang, Prafulla Kumar Choubey, Yilun Zhou 외 arxiv

Web agents, which couple language models with browsing and tool-use capabilities, show promise as open web assistants. Yet progress is increasingly limited by the lack of scalable, process-level supervision. Existing ben…

Beyond the Library: An Agentic Framework for Autoformalizing Research Mathematics

2026-06-30 · Arshia Soltani Moakhar, Iman Gholami, Max Springer, Mahdi JafariRaviz 외 arxiv

While Large Language Models (LLMs) have demonstrated exceptional capabilities in mathematical reasoning, they frequently produce subtle errors that evade human detection. Formal mathematical languages like Lean 4 offer m…

Mathematical Reasoning

Formalizing Mathematics at Scale

2026-05-28 · Ahmad Rammal, Niket Patel, Fabian Gloeckle, Amaury Hayat 외 arxiv

We present AutoformBot, a multi-agent system for building an Autoformalized Textbook Library At Scale (Atlas) in Lean 4. AutoformBot orchestrates thousands of LLM agents, equipped with formal verification tools, dependen…

IndiMathBench: Autoformalizing Mathematical Reasoning Problems with a Human Touch

2025-11-30 · Param Biyani, Shashank Kirtania, Yasharth Bajpai, Sumit Gulwani 외 arxiv

Reliable autoformalization remains challenging even in the era of large language models (LLMs). The scarcity of high-quality training data is a major bottleneck. Expert annotation requires substantial time and deep exper…

Mathematical Reasoning

Grading the Grader: Lessons from Evaluating an Agentic Data Analysis System

2026-06-23 · Tian Zheng, Kai-Tai Hsu arxiv

Agentic data analysis systems produce rich outputs, including code, numerical results, and verbal diagnostics. This makes them more challenging to evaluate than single-turn LLM responses. It is therefore necessary to dis…