多智能体 RL论文趋势 2023–2026
博弈与 MARL
统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 2,126 篇被归入该主题的论文。
主要结论
- 多智能体 RL论文从 2023 年的 311 篇增长到 2026 年的 868 篇，三年间扩大到约 2.8 倍。
- NeurIPS 是发多智能体 RL论文最多的会议，2023–2026 年共 809 篇，占四个会议总量的 38%。
- 交叉方向里涨得最快的是多模态大模型：同时属于两个主题的论文从 2023 年的 9 篇增至 2026 年的 70 篇。
- 在「博弈与规划」的 5 个主题里，它的 2026 年论文数排第 1；同组增长最快的经典规划是 3.3 倍，多智能体 RL是 2.8 倍。
各会议论文数
按年份和会议统计
| 年份 | NeurIPS | ICLR | ICML | CVPR | 合计 |
|---|---|---|---|---|---|
| 2023 | 88 | 37 | 80 | 106 | 311 |
| 2024 | 147 | 39 | 82 | 105 | 373 |
| 2025 | 255 | 91 | 115 | 113 | 574 |
| 2026 | 319 | 153 | 224 | 172 | 868 |
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 20262-Step Agent: How a Bayesian Decision Maker Learns from AI-Decision Support
- ICML 2026A Causal Decomposition Approach for Fair Contextual Multi-Armed Bandits
- ICLR 2026Calibrated Information Bottleneck for Trusted Multi-modal Clustering
- CVPR 2026A Sanity Check for Multi-In-Domain Face Forgery Detection in the
- NeurIPS 2025Algorithms and SQ Lower Bounds for Robustly Learning Real-valued Multi-Index Models
- ICML 2025A Classification View on Meta Learning Bandits
- ICLR 2025BEEM: Boosting Performance of Early Exit DNNs using Multi-Exit Classifiers as Experts
- CVPR 2025All-Day Multi-Camera Multi-Target Tracking
- NeurIPS 2024A New Multi-Source Light Detection Benchmark and Semi-Supervised Focal Light Detection
- ICML 2024A Federated Stochastic Multi-level Compositional Minimax Algorithm for Deep AUC Maximization
- ICLR 2024An Intuitive Multi-Frequency Feature Representation for SO(3)-Equivariant Networks
- CVPR 2024DeconfuseTrack: Dealing with Confusion for Multi-Object Tracking
- NeurIPS 2023A Novel Approach for Effective Multi-View Clustering with Information-Theoretic Perspective
- ICML 2023A Framework for Adapting Offline Algorithms to Solve Combinatorial Multi-Armed Bandit Problems with Bandit Feedback
- ICLR 2023An Additive Instance-Wise Approach to Multi-class Model Interpretation
- CVPR 2023A Light Touch Approach to Teaching Transformers Multi-View Geometry
- NeurIPS 2026A Black-Box Reduction from Regret to Multi-Level Coverage
- ICML 2026A Consensus Anchor-guided Hypergraph Framework for Incomplete Multi-view Clustering
- ICLR 2026Confident Block Diagonal Structure-Aware Invariable Graph Completion for Incomplete Multi-view Clustering
- CVPR 2026Anti-Degradation Lifelong Multi-View Clustering
全部多智能体 RL论文
共 868 篇，第 18 / 18 页
- NeurIPS 2026VerifyMAS: Hypothesis Verification for Failure Attribution in LLM Multi-Agent Systems
- NeurIPS 2026VSearcher: Long-Horizon Multimodal Search Agent via Reinforcement Learning
- NeurIPS 2026Wayfinder: Adaptive Resource Routing from Agent Citations
- NeurIPS 2026Welfare, Improvability, and Variance: A Principal-Agent Approach to Optimal Benchmark Item Aggregation
- NeurIPS 2026What Makes Two States the Same? Value-Aware State Matching for Multi-Turn Agent Credit Assignment
- NeurIPS 2026When and Why Does Multi-Agent Debate Fail and Does It Really Underperform?
- NeurIPS 2026When Debate Helps: Proposal Supply and Verification-Aware Readout in Multi-Agent Reasoning
- NeurIPS 2026When Do Multi-Agent Systems Help? An Information Bottleneck Perspective
- NeurIPS 2026When Parallelism Pays Off: Cohesion-Aware Task Partitioning for Multi-Agent Coding
- NeurIPS 2026When Stored Evidence Stops Being Usable: Scale-Conditioned Evaluation of Agent Memory
- NeurIPS 2026When the Most Disruptive Messages Are Safe to Prune in Aggregate: Functional Analysis of Communication Pruning in Two-Agent LLM Systems
- NeurIPS 2026Who Should Evolve? Uncertainty-Aware Role Bottleneck Inference for Multi-Agent LLM Training
- NeurIPS 2026Why Copy Others? Insights into Social Learning from Multi-Agent Reinforcement Learning
- NeurIPS 2026WorldMemArena: Evaluating Multimodal Agent Memory Through Action–World Interaction
- NeurIPS 2026WURI: Watching Unfolding Risk in Agent Interactions
- NeurIPS 2026Youdunit: Single-Call Counterfactual Necessity in Multi-Agent LLM Systems
- NeurIPS 2026Zero-Violation Regret for Cooperative Markov Games with Coupled Instantaneous Hard Constraints
- NeurIPS 2026Zeroth-Order Stackelberg Control in Combinatorial Congestion Games
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 查多智能体 RL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": "multi-agent-rl"}
常见问题
- 2023–2026 年四大会议一共有多少篇多智能体 RL论文？
- 共 2,126 篇，按年份分别是 2023 年 311 篇、2024 年 373 篇、2025 年 574 篇、2026 年 868 篇。
- NeurIPS 2026 有多少篇多智能体 RL论文？
- NeurIPS 2026 有 319 篇多智能体 RL论文，上一届是 255 篇。
- ICLR 2026 有多少篇多智能体 RL论文？
- ICLR 2026 有 153 篇多智能体 RL论文，上一届是 91 篇。
- ICML 2026 有多少篇多智能体 RL论文？
- ICML 2026 有 224 篇多智能体 RL论文，上一届是 115 篇。
- CVPR 2026 有多少篇多智能体 RL论文？
- CVPR 2026 有 172 篇多智能体 RL论文，上一届是 113 篇。
- 哪个会议的多智能体 RL论文最多？
- NeurIPS，2023–2026 年共 809 篇。