Research map / Large language models / Agents, code and math
HOList: An Environment for Machine Learning of Higher Order Logic Theorem Proving
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 WaveNet argument selection to ASM_MESON_TAC | ASM_MESON_TAC · complex validation | proof success: $6.1$% -> $9.2$% | Helped |
| Use a deeper WaveNet with four blocks of five layers (ablation) | WaveNet · complex validation | proof success: $31.72$% -> $32.65$% | Helped |
| Increase the number of filters in the WaveNet (ablation) | WaveNet · complex validation | proof success: $31.72$% -> $27.60$% | Hurt |
| Train on loop-generated proof output without using that model to guide proof search (ablation) | Loop · complex validation | proof success: $36.3$% -> $36.8$% | Helped |
| Condition argument ranking on the selected tactic (ablation) | Loop · complex validation | proof success: $36.3$% -> $38.9$% | Helped |
| Train the loop to prove internal proof states as separate targets (ablation) | Reinforcement learning loop · training set | theorems proved: $5679$ ($55.7$%) -> $1988$ ($19.5$%) | Hurt |
| Apply the Loop model to Flyspeck theorems | Loop · flyspeck sample | proof success: 37.0\% | Helped |
Closest papers by meaning
- HolStep: A Machine Learning Dataset for Higher-order Logic Theorem Proving (ICLR 2017 · similarity 0.93)
- ProofAug: Efficient Neural Theorem Proving via Fine-grained Proof Structure Analysis (ICML 2025 · similarity 0.87)
- APOLLO: Automated LLM and Lean Collaboration for Advanced Formal Reasoning (NeurIPS 2025 · similarity 0.87)
- LEGO-Prover: Neural Theorem Proving with Growing Libraries (ICLR 2024 · similarity 0.87)
- Reviving DSP for Advanced Theorem Proving in the Era of Reasoning Models (NeurIPS 2025 · similarity 0.87)
- Scaling up Multi-Turn Off-Policy RL and Multi-Agent Tree Search for LLM Step-Provers (ICML 2026 · similarity 0.87)
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.