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
- 强化学习论文从 2023 年的 1,070 篇增长到 2026 年的 2,985 篇，三年间扩大到约 2.8 倍。
- NeurIPS 是发强化学习论文最多的会议，2023–2026 年共 2,494 篇，占四个会议总量的 36%。
- 交叉方向里涨得最快的是推理模型：同时属于两个主题的论文从 2023 年的 9 篇增至 2026 年的 437 篇。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 484 | 233 | 297 | 56 | 1,070 |
| 2024 | 466 | 304 | 356 | 64 | 1,190 |
| 2025 | 784 | 379 | 405 | 72 | 1,640 |
| 2026 | 760 | 911 | 1,021 | 293 | 2,985 |
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
- NeurIPS 2026\texttt{FEROM}: Frontier Endogenous Reveal-Order Marginal Policy Optimization for Masked Diffusion LMs
- ICML 2026$\text{DT}^\text{2}$: Decision-Targeted Digital Twins
- ICLR 2026$\nabla$-Reasoner: LLM Reasoning via Test-Time Gradient Descent in Textual Space
- CVPR 2026A Causal Marriage between VLM and IRM from Understanding to
- NeurIPS 2025$\texttt{G1}$: Teaching LLMs to Reason on Graphs with Reinforcement Learning
- ICML 2025A Computationally Efficient Algorithm for Infinite-Horizon Average-Reward Linear MDPs
- ICLR 2025$\phi$-Update: A Class of Policy Update Methods with Policy Convergence Guarantee
- CVPR 2025All-Optical Nonlinear Diffractive Deep Network for Ultrafast Image Denoising
- NeurIPS 2024A Critical Evaluation of AI Feedback for Aligning Large Language Models
- ICML 2024${\rm E}(3)$-Equivariant Actor-Critic Methods for Cooperative Multi-Agent Reinforcement Learning
- ICLR 2024A Black-box Approach for Non-stationary Multi-agent Reinforcement Learning
- CVPR 2024DMR: Decomposed Multi-Modality Representations for Frames and Events Fusion in Visual Reinforcement Learning
- NeurIPS 2023“Why Not Looking backward?” A Robust Two-Step Method to Automatically Terminate Bayesian Optimization
- ICML 2023A Connection between One-Step RL and Critic Regularization in Reinforcement Learning
- ICLR 2023$O(T^{-1})$ Convergence of Optimistic-Follow-the-Regularized-Leader in Two-Player Zero-Sum Markov Games
- CVPR 2023"Seeing" Electric Network Frequency From Events
- NeurIPS 2026$\varphi$TD: Distributional Reinforcement Learning using Characteristic Functions
- ICML 2026$\textit{S}$-SPPO: Semantic-Calibrated Self-Play Preference Optimization
- ICLR 2026$AutoDrive\text{-}P^3$: Unified Chain of Perception–Prediction–Planning Thought via Reinforcement Fine-Tuning
- CVPR 2026AdaptVision Efficient Vision-Language Models via Adaptive Visual Acquisition
全部强化学习论文
共 6,885 篇，第 22 / 138 页
- ICLR 2026Stackelberg Learning from Human Feedback: Preference Optimization as a Sequential Game
- ICLR 2026STAIRS-Former: Spatio-Temporal Attention with Interleaved Recursive Structure TransFormer for Offline Mulit-task Multi-agent Reinforcement Learning
- ICLR 2026STAR: Similarity-guided Teacher-Assisted Refinement for Super-Tiny Function Calling Models
- ICLR 2026Statistical Guarantees for Offline Domain Randomization
- ICLR 2026StepORLM: A Self-Evolving Framework With Generative Process Supervision For Operations Research Language Models
- ICLR 2026Stop Unnecessary Reflection: Training LRMs for Efficient Reasoning with Adaptive Reflection and Length Coordinated Penalty
- ICLR 2026Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement Learning
- ICLR 2026Stronger-MAS: Multi-Agent Reinforcement Learning for Collaborative LLMs
- ICLR 2026Structured Reasoning for LLMs: A Unified Framework for Efficiency and Explainability
- ICLR 2026STVG-R1: Incentivizing Instance-Level Reasoning and Grounding in Videos via Reinforcement Learning
- ICLR 2026Supervised Reinforcement Learning: From Expert Trajectories to Step-wise Reasoning
- ICLR 2026Supporting High-Stakes Decision Making Through Interactive Preference Elicitation in the Latent Space
- ICLR 2026SUSD: Structured Unsupervised Skill Discovery through State Factorization
- ICLR 2026Swap-guided Preference Learning for Personalized Reinforcement Learning from Human Feedback
- ICLR 2026SWE-RM: Execution-free Feedback for Software Engineering Agents
- ICLR 2026Synergizing Understanding and Generation with Interleaved Analyzing-Drafting Thinking
- ICLR 2026Synthesizing High-Quality Visual Question Answering from Medical Documents with Generator-Verifier LMMs
- ICLR 2026Synthesizing Multimodal Verifiable Game Data to Boost VLMs' General Reasoning
- ICLR 2026Tackling Heavy-Tailed Q-Value Bias in Offline-to-Online Reinforcement Learning with Laplace-Robust Modeling
- ICLR 2026Talking Points: Describing and Localizing Pixels
- ICLR 2026Taming Imperfect Process Verifiers: A Sampling Perspective on Backtracking
- ICLR 2026Task Tokens: A Flexible Approach to Adapting Behavior Foundation Models
- ICLR 2026Task-Agnostic Amortized Multi-Objective Optimization
- ICLR 2026TaTToo: Tool-Grounded Thinking PRM for Test-Time Scaling in Tabular Reasoning
- ICLR 2026TD-JEPA: Latent-predictive Representations for Zero-Shot Reinforcement Learning
- ICLR 2026Temperature as a Meta-Policy: Adaptive Temperature in LLM Reinforcement Learning
- ICLR 2026TEMPFLOW-GRPO: WHEN TIMING MATTERS FOR GRPO IN FLOW MODELS
- ICLR 2026Test-Time Alignment for Large Language Models via Textual Model Predictive Control
- ICLR 2026Text2Grad: Reinforcement Learning from Natural Language Feedback
- ICLR 2026The Alignment Auditor: A Bayesian Framework for Verifying and Refining LLM Objectives
- ICLR 2026The Alignment Waltz: Jointly Training Agents to Collaborate for Safety
- ICLR 2026The Art of Scaling Reinforcement Learning Compute for LLMs
- ICLR 2026The Choice of Divergence: A Neglected Key to Mitigating Diversity Collapse in Reinforcement Learning with Verifiable Reward
- ICLR 2026The Coverage Principle: How Pre-Training Enables Post-Training
- ICLR 2026The Markovian Thinker
- ICLR 2026The Overthinking Predicament: When Reasoning Hurts Ranking
- ICLR 2026The Rank and Gradient Lost in Non-stationarity: Sample Weight Decay for Mitigating Plasticity Loss in Reinforcement Learning
- ICLR 2026The Sample Complexity of Online Reinforcement Learning: A Multi-model Perspective
- ICLR 2026The State of Reinforcement Finetuning for Transformer-based Generative Agents
- ICLR 2026Thinking-Free Policy Initialization Makes Distilled Reasoning Models More Effective and Efficient Reasoners
- ICLR 2026Thompson Sampling via Fine-Tuning of LLMs
- ICLR 2026THOR: Tool-Integrated Hierarchical Optimization via RL for Mathematical Reasoning
- ICLR 2026Thyme: Think Beyond Images
- ICLR 2026TikZilla: Scaling Text-to-TikZ with High-Quality Data and Reinforcement Learning
- ICLR 2026Time Is a Feature: Exploiting Temporal Dynamics in Diffusion Language Models
- ICLR 2026TimeSearch-R: Adaptive Temporal Search for Long-Form Video Understanding via Self-Verification Reinforcement Learning
- ICLR 2026TimeSeg: An Information-Theoretic Segment-Wise Explainer for Time-Series Predictions
- ICLR 2026Tina: Tiny Reasoning Models via LoRA
- ICLR 2026TIPO: Text to Image with Text Pre-sampling for Prompt Optimization
- ICLR 2026TIPS: Turn-level Information-Potential Reward Shaping for Search-Augmented LLMs
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"}
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。