强化学习论文趋势 2023–2026

策略优化、离线 RL 与基于环境的决策学习

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 6,885 篇被归入该主题的论文。

主要结论

每年论文数

各会议论文数

按年份和会议统计

强化学习论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
2023484233297561,070
2024466304356641,190
2025784379405721,640
20267609111,0212932,985

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

全部强化学习论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 2,079 篇，第 14 / 42 页

  1. ICML 2026Prioritize the Process, Not Just the Outcome: Rewarding Latent Thought Trajectories Improves Reasoning in Looped Language Models
  2. ICML 2026Prioritized Model Experience Replay
  3. ICML 2026PRISM: Demystifying Retention and Interaction in Mid-Training
  4. ICML 2026Privileged Information Distillation for Language Models
  5. ICML 2026PRM-PBE: Process Reward Model for Reinforcement Learning in Programming-by-Example
  6. ICML 2026Probabilistic Performance Guarantees for Multi-Task Reinforcement Learning
  7. ICML 2026Probing RLVR Training Instability through the Lens of Objective-Level Hacking
  8. ICML 2026Procedural Generation Of Algorithm Discovery Tasks in Machine Learning
  9. ICML 2026PromptPilot: Game-Theoretic Multi-Agent Prompt Optimization for Segment Anything
  10. ICML 2026PromptRL: Prompt Matters in RL for Flow-Based Image Generation
  11. ICML 2026ProRL: Effective Reinforcement Learning for Proactive Recommendation via Rectified Policy Gradient Estimation
  12. ICML 2026Provable Benefits of RLVR over SFT for Reasoning Models: Learning to Backtrack Efficiently
  13. ICML 2026Provable Sample Efficiency of Curriculum Post-Training for Transformer Reasoning
  14. ICML 2026Provably Convergent Actor-Critic for MARL through Risk-aversion
  15. ICML 2026Provably Efficient Policy-Reward Co-Pretraining for Adversarial Imitation Learning
  16. ICML 2026PRPO: Paragraph-level Policy Optimization for Vision-Language Deepfake Detection
  17. ICML 2026PS-PPO : Prefix-Sampling PPO for Critic-Free RLHF
  18. ICML 2026Pushing Forward Pareto Frontiers of Proactive Agents with Behavioral Agentic Optimization
  19. ICML 2026PyVision-RL: Forging Open Agentic Vision Models via RL
  20. ICML 2026Q-Flow: Stable and Expressive Reinforcement Learning with Flow-based Policy
  21. ICML 2026Q-SAM: Unlocking Sharpness-Aware Minimization for Generalization in Offline Reinforcement Learning
  22. ICML 2026QHyer: Q-conditioned Hybrid Attention-mamba Transformer for Offline Goal-conditioned RL
  23. ICML 2026Quantifying and Optimizing Simplicity via Polynomial Representations
  24. ICML 2026Quantum Robust Inner Minimization for Reinforcement Learning with Quadratic Speed-Up in Query Complexity
  25. ICML 2026QUATRO: Query-Adaptive Trust Region Policy Optimization for LLM Fine-tuning
  26. ICML 2026R-Diverse: Mitigating Diversity Illusion in Self-Play LLM Training
  27. ICML 2026R1-SyntheticVL: Is Synthetic Data from Generative Models Ready for Multimodal Large Language Model?
  28. ICML 2026R2R2: Robust Representation for Intensive Experience Reuse via Redundancy Reduction in Self-Predictive Learning
  29. ICML 2026RAD: Retrieval High-quality Demonstrations to Enhance Decision-making
  30. ICML 2026RAGEN-2: Reasoning Collapse in Agentic RL
  31. ICML 2026RAMAC: Multimodal Risk-Aware Offline Reinforcement Learning and the Role of Behavior Regularization
  32. ICML 2026Randomized Advantage Transformation (RAT): Computing Natural Policy Gradients via Direct Backpropagation
  33. ICML 2026RAST-MoE-RL: A Regime-Aware Spatio-Temporal MoE Framework for Deep Reinforcement Learning in Ride-Hailing
  34. ICML 2026Rate or Fate? RLV$^{\varepsilon}$R: Reinforcement Learning with Verifiable Noisy Rewards
  35. ICML 2026Ratio-Variance Regularized Policy Optimization
  36. ICML 2026Rationality Measurement and Theory for Reinforcement Learning Agents
  37. ICML 2026Real-Time Aligned Reward Model beyond Semantics
  38. ICML 2026REAL: Regression-Aware Reinforcement Learning for LLM-as-a-Judge
  39. ICML 2026Reasoning Cache: Continual Improvement Over Long Horizons via Short-Horizon RL
  40. ICML 2026Reasoning-VLA: An Efficient and Spatial-Guided General Vision-Language-Action Reasoning Model for Autonomous Driving
  41. ICML 2026ReAugment: Targeted Few-Shot Time Series Augmentation via Model Zoo-Guided Reinforcement Learning
  42. ICML 2026Recontextualization Mitigates Specification Gaming Without Modifying the Specification
  43. ICML 2026Recovering Hidden Reward in Diffusion-Based Policies
  44. ICML 2026Recurrent Equivariant Constraint Modulation: Learning Per-Layer Symmetry Relaxation from Data
  45. ICML 2026Recurrent Structural Policy Gradient for Partially Observable Mean Field Games
  46. ICML 2026Refined Analysis of Entropy-Regularized Actor-Critic
  47. ICML 2026Reflect-then-Correct: Rebalancing Task Optimization for Generalizable Meta-Reinforcement Learning via Distributional Value Error Reduction
  48. ICML 2026Reflector: Internalizing Step-wise Reflection against Indirect Jailbreaks
  49. ICML 2026REG: In-Sample RL via Regularizing the Evaluation Gap
  50. ICML 2026Regularized Offline Policy Optimization with Posterior Hybrid Bayesian Belief

在 Claude、Cursor 里直接问这个趋势

接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。

可以直接复制这句话问 AI

用 paperlist 查强化学习在 NeurIPS 2023–2026 的论文趋势，哪一年涨得最快？

AI 会调用的工具：get_topic_stats {"topic": "reinforcement-learning"}

两分钟接入 Paperlist MCP（Claude Code、Cursor、Codex、ChatGPT）

常见问题

2023–2026 年四大会议一共有多少篇强化学习论文？
共 6,885 篇，按年份分别是 2023 年 1,070 篇、2024 年 1,190 篇、2025 年 1,640 篇、2026 年 2,985 篇。
NeurIPS 2026 有多少篇强化学习论文？
NeurIPS 2026 有 760 篇强化学习论文，上一届是 784 篇。
ICLR 2026 有多少篇强化学习论文？
ICLR 2026 有 911 篇强化学习论文，上一届是 379 篇。
ICML 2026 有多少篇强化学习论文？
ICML 2026 有 1,021 篇强化学习论文，上一届是 405 篇。
CVPR 2026 有多少篇强化学习论文？
CVPR 2026 有 293 篇强化学习论文，上一届是 72 篇。
哪个会议的强化学习论文最多？
NeurIPS，2023–2026 年共 2,494 篇。