Lean Refactor: Multi-Objective Controllable Proof Optimization via Agentic Strategy Search
The article discusses a new framework called Lean Refactor designed for optimizing Lean proofs. This framework addresses challenges such as proof length, compilation costs, and version compatibility. Experiments demonstrate significant improvements in token-level compression and compilation time compared to previous methods.
- ▪Lean Refactor is a retrieval-augmented framework for multi-objective proof optimization.
- ▪It tackles issues related to proof length, compilation costs, and compatibility across Lean versions.
- ▪Experiments show over 70% token-level compression and up to 60% reduction in compilation time.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Logic in Computer Science arXiv:2605.20244 (cs) [Submitted on 18 May 2026] Title:Lean Refactor: Multi-Objective Controllable Proof Optimization via Agentic Strategy Search Authors:Jialin Lu, Soonho Kong, Rodrigo Stehling, Kaiyu Yang, Zhangyang Wang, Weiran Sun, Wuyang Chen View a PDF of the paper titled Lean Refactor: Multi-Objective Controllable Proof Optimization via Agentic Strategy Search, by Jialin Lu and 6 other authors View PDF HTML (experimental) Abstract:We present Lean Refactor, a plug-and-play retrieval-augmented agentic framework for multi-objective, controllable, and version-robust refactoring of Lean proofs.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
