数学推理论文趋势 2023–2026

主要结论

• 数学推理论文从 2023 年的 57 篇增长到 2026 年的 577 篇，三年间扩大到约 10 倍。
• NeurIPS 是发数学推理论文最多的会议，2023–2026 年共 469 篇，占四个会议总量的 43%。
• 交叉方向里涨得最快的是强化学习：同时属于两个主题的论文从 2023 年的 3 篇增至 2026 年的 140 篇。
• 在「神经符号」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的可微推理是 2.5 倍，数学推理是 10 倍。

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

• NeurIPS 2026A mathematical theory of balancing relational generalization and memorization
• ICML 2026ASyMOB: Algebraic Symbolic Mathematical Operations Benchmark
• ICLR 2026AutoGPS: Automated Geometry Problem Solving via Multimodal Formalization and Deductive Reasoning
• CVPR 2026Affine Perspective-Three-Point Problem
• NeurIPS 2025Are Language Models Efficient Reasoners? A Perspective from Logic Programming
• ICML 2025Autoformulation of Mathematical Optimization Models Using LLMs
• ICLR 2025Advancing Mathematical Reasoning in Language Models: The Impact of Problem-Solving Data, Data Synthesis Methods, and Training Stages
• CVPR 2025Bridging Viewpoint Gaps: Geometric Reasoning Boosts Semantic Correspondence
• NeurIPS 2024A Careful Examination of Large Language Model Performance on Grade School Arithmetic
• ICML 2024Accelerating Legacy Numerical Solvers by Non-intrusive Gradient-based Meta-solving
• ICLR 2024Don't Trust: Verify -- Grounding LLM Quantitative Reasoning with Autoformalization
• CVPR 2024E-GPS: Explainable Geometry Problem Solving via Top-Down Solver and Bottom-Up Generator
• NeurIPS 2023Evaluating and Improving Tool-Augmented Computation-Intensive Math Reasoning
• ICML 2023Automated Search for Conjectures on Mathematical Constants using Analysis of Integer Sequences
• ICLR 2023Differentiable Mathematical Programming for Object-Centric Representation Learning
• CVPR 2023Abstract Visual Reasoning: An Algebraic Approach for Solving Raven's Progressive Matrices
• NeurIPS 2026A Mechanistic Analysis of Looped Reasoning Language Models
• ICML 2026Beyond Theorem Proving: Formulation, Framework and Benchmark for Formal Problem-Solving
• ICLR 2026CogFlow: Bridging Perception and Reasoning through Knowledge Internalization for Visual Mathematical Problem Solving
• CVPR 2026Hilbert-Geo Solving Solid Geometric Problems by Neural-Symbolic Reasoning

全部数学推理论文

共 208 篇，第 3 / 5 页

101. ICML 2026Towards Efficient Large Language Reasoning Models via Extreme-Ratio Chain-of-Thought Compression
102. ICML 2026Towards Solving the Gilbert-Pollak Conjecture via Large Language Models
103. ICML 2026Truthfulness Does Not Scale Like Reasoning: Why Polling Fails as a Proxy Verifier
104. ICML 2026Turbo Connection: Reasoning as Information Flow from Higher to Lower Layers
105. ICML 2026Turning Drift into Constraint: Robust Reasoning Alignment in Non-Stationary Multi-Stream Environments
106. ICML 2026Uncovering the Gradient Geometry of Long CoT: A Spectral-guided Approach to Reasoning Distillation
107. ICML 2026Unveiling the Entropy Dynamics of Chain-of-Thought Reasoning
108. ICML 2026Verifying Meta-Awareness via Predictive Rewards in Reasoning Models
109. ICML 2026Wait, Wait, Wait... Why Do Reasoning Models Loop?
110. ICML 2026What Characterizes Effective Reasoning? Revisiting Length, Review, and Structure of CoT
111. ICML 2026What If We Allocate Test-Time Compute Adaptively?
112. ICML 2026What Really Improves Mathematical Reasoning: Structured Reasoning Signals Beyond Pure Code
113. ICML 2026When to Trust the Cheap Check: Weak and Strong Verification for Reasoning
114. ICML 2025A Mathematical Framework for AI-Human Integration in Work
115. ICML 2025A Reasoning-Based Approach to Cryptic Crossword Clue Solving
116. ICML 2025A Sub-Problem Quantum Alternating Operator Ansatz for Correlation Clustering
117. ICML 2025AdaptiveStep: Automatically Dividing Reasoning Step through Model Confidence
118. ICML 2025Attention-Level Speculation
119. ICML 2025Autoformulation of Mathematical Optimization Models Using LLMs
120. ICML 2025Beyond Task-Specific Reasoning: A Unified Conditional Generative Framework for Abstract Visual Reasoning
121. ICML 2025Blink of an eye: a simple theory for feature localization in generative models
122. ICML 2025Can Transformers Reason Logically? A Study in SAT Solving
123. ICML 2025Discrete Neural Algorithmic Reasoning
124. ICML 2025Do NOT Think That Much for 2+3=? On the Overthinking of Long Reasoning Models
125. ICML 2025Fleet of Agents: Coordinated Problem Solving with Large Language Models
126. ICML 2025Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning
127. ICML 2025Free Process Rewards without Process Labels
128. ICML 2025From Language Models over Tokens to Language Models over Characters
129. ICML 2025Graph-constrained Reasoning: Faithful Reasoning on Knowledge Graphs with Large Language Models
130. ICML 2025GSM-$\infty$: How Do your LLMs Behave over Infinitely Increasing Reasoning Complexity and Context Length?
131. ICML 2025HyperTree Planning: Enhancing LLM Reasoning via Hierarchical Thinking
132. ICML 2025Improving Rationality in the Reasoning Process of Language Models through Self-playing Game
133. ICML 2025Leveraging Online Olympiad-Level Math Problems for LLMs Training and Contamination-Resistant Evaluation
134. ICML 2025Lower Bounds for Chain-of-Thought Reasoning in Hard-Attention Transformers
135. ICML 2025MA-LoT: Model-Collaboration Lean-based Long Chain-of-Thought Reasoning enhances Formal Theorem Proving
136. ICML 2025Machine Learning meets Algebraic Combinatorics: A Suite of Datasets Capturing Research-level Conjecturing Ability in Pure Mathematics
137. ICML 2025Machines and Mathematical Mutations: Using GNNs to Characterize Quiver Mutation Classes
138. ICML 2025MARGE: Improving Math Reasoning with Guided Exploration
139. ICML 2025MASS: Mathematical Data Selection via Skill Graphs for Pretraining Large Language Models
140. ICML 2025MATH-Perturb: Benchmarking LLMs' Math Reasoning Abilities against Hard Perturbations
141. ICML 2025MathConstruct: Challenging LLM Reasoning with Constructive Proofs
142. ICML 2025Metastable Dynamics of Chain-of-Thought Reasoning: Provable Benefits of Search, RL and Distillation
143. ICML 2025MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and Efficiency
144. ICML 2025One Example Shown, Many Concepts Known! Counterexample-Driven Conceptual Reasoning in Mathematical LLMs
145. ICML 2025Position: AI's growing due process problem
146. ICML 2025Position: Formal Mathematical Reasoning—A New Frontier in AI
147. ICML 2025Premise-Augmented Reasoning Chains Improve Error Identification in Math reasoning with LLMs
148. ICML 2025Putnam-AXIOM: A Functional & Static Benchmark for Measuring Higher Level Mathematical Reasoning in LLMs
149. ICML 2025Regress, Don't Guess: A Regression-like Loss on Number Tokens for Language Models
150. ICML 2025rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking

常见问题

2023–2026 年四大会议一共有多少篇数学推理论文？

共 1,085 篇，按年份分别是 2023 年 57 篇、2024 年 137 篇、2025 年 314 篇、2026 年 577 篇。

NeurIPS 2026 有多少篇数学推理论文？

NeurIPS 2026 有 208 篇数学推理论文，上一届是 190 篇。

ICLR 2026 有多少篇数学推理论文？

ICLR 2026 有 175 篇数学推理论文，上一届是 61 篇。

ICML 2026 有多少篇数学推理论文？

ICML 2026 有 113 篇数学推理论文，上一届是 45 篇。

CVPR 2026 有多少篇数学推理论文？

CVPR 2026 有 81 篇数学推理论文，上一届是 18 篇。

哪个会议的数学推理论文最多？

NeurIPS，2023–2026 年共 469 篇。