Research map / Large language models / Agents, code and math
EvolProver: Advancing Automated theorem proving by Evolving Formalized Problems via Symmetry and Difficulty
Yuchen Tian, Ruiyuan Huang, WANG XUANWU, Jing Ma, Zengfeng Huang, Ziyang Luo, Hongzhan Lin, Da Zheng, Lun Du
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add EvolDomain and EvolDifficulty augmentation (SFT) (ablation) | EvolProver-Base · FormalMATH-Lite | pass@32: 44.71% -> 50.35% | Improved |
| add EvolDomain and EvolDifficulty augmentation (SFT) (ablation) | EvolProver-Base · MiniF2F-Test | pass@32: 52.05% -> 65.16% | Improved |
| add EvolDomain and EvolDifficulty augmentation (SFT) (ablation) | EvolProver-Base · Ineq-Comp Transformed | pass@32: 14.89% -> 29.19% | Improved |
| add EvolAST on top of EvolDomain and EvolDifficulty (SFT) (ablation) | EvolProver-Ablation-SFT · FormalMATH-Lite | pass@32: 50.35% -> 51.53% | Improved |
| add EvolAST on top of EvolDomain and EvolDifficulty (SFT) (ablation) | EvolProver-Ablation-SFT · MiniF2F-Test | pass@32: 65.16% -> 66.39% | Improved |
| add EvolAST (SFT+RL) (ablation) | EvolProver-Ablation-RL · FormalMATH-Lite | pass@32: 51.98% -> 53.96% | Improved |
| add EvolAST (SFT+RL) (ablation) | EvolProver-Ablation-RL · MiniF2F-Test | pass@32: 68.22% -> 69.80% | Improved |
| add EvolAST (SFT+RL) (ablation) | EvolProver-Ablation-RL · Ineq-Comp pass ratio | ratio (%): 65.62% -> 65.17% | No clear effect |
| add RL stage after SFT (ablation) | EvolProver · FormalMATH-Lite | pass@32: 51.53% -> 53.96% | Improved |
| evolve formal statements directly instead of evolving NL problems then formalizing (ablation) | EvolDomain and EvolDifficulty · verified candidates from 400 seeds | number verified | Improved |
| apply EvolDomain to broaden domain distribution (ablation) | EvolProver · FormalMATH-Lite Calculus | problems proved: 0 -> 3 | Improved |
Closest papers by meaning
- Alchemy: Amplifying Theorem-Proving Capability Through Symbolic Mutation (ICLR 2025 · similarity 0.90)
- APOLLO: Automated LLM and Lean Collaboration for Advanced Formal Reasoning (NeurIPS 2025 · similarity 0.90)
- MPS-Prover: Advancing Stepwise Theorem Proving by Multi-Perspective Search and Data Curation (NeurIPS 2025 · similarity 0.89)
- MUSTARD: Mastering Uniform Synthesis of Theorem and Proof Data (ICLR 2024 · similarity 0.89)
- ProofAug: Efficient Neural Theorem Proving via Fine-grained Proof Structure Analysis (ICML 2025 · similarity 0.89)
- ImProver: Agent-Based Automated Proof Optimization (ICLR 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
