Research map / Multi-agent systems and games / Cooperative multi-agent RL
Intrinsic-reward exploration for MARL
Drives multi-agent exploration with individual and collective intrinsic rewards and curiosity.
34 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2019–2026) · an approach in Cooperative multi-agent RL, keyword cluster “reinforcement marl · communication · marl algorithms”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
12 papers in 2023–24 and 11 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 8.5% to 4.8%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2019 · Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning (ICML 2019 · 254 citations)
- 2020 · Promoting Coordination through Policy Regularization in Multi-Agent Deep Reinforcement Learning (NeurIPS 2020 · 13 citations)
- 2021 · Episodic Multi-agent Reinforcement Learning with Curiosity-driven Exploration (NeurIPS 2021 · 41 citations)
- 2022 · LIGS: Learnable Intrinsic-Reward Generation Selection for Multi-Agent Learning (ICLR 2022 · 3 citations)
- 2023 · Subspace-Aware Exploration for Sparse-Reward Multi-Agent Tasks (AAAI 2023 · 15 citations)
- 2024 · FoX: Formation-Aware Exploration in Multi-Agent Reinforcement Learning (AAAI 2024 · 12 citations)
- 2025 · AIR: Unifying Individual and Collective Exploration in Cooperative Multi-Agent Reinforcement Learning (AAAI 2025 · 0 citations)
- 2026 · The Horizon Threshold in Cooperative Multi-Agent Reward-Free Exploration (NeurIPS 2026 · 0 citations)
Earliest papers here
- Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning (ICML 2019 · 254 citations)
- LIIR: Learning Individual Intrinsic Reward in Multi-Agent Reinforcement Learning (NeurIPS 2019 · 75 citations)
- Promoting Coordination through Policy Regularization in Multi-Agent Deep Reinforcement Learning (NeurIPS 2020 · 13 citations)
What has been tried
27 papers here report 289 measured changes (a component added, removed or swapped, compared with a baseline): helped 59%, no clear effect 9%, hurt 28%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add formation-based exploration reward | GRF Academy_3_vs_1_with_keeper · performance | Helped | FoX: Formation-Aware Exploration in Multi-Agent Reinforcement Learning (AAAI 2024 · 12 citations) | |
| add summed local novelty from other agents to approximate global novelty | GridWorld Pass, SecretRoom, MultiRoom · mean reward | Helped | Settling Decentralized Multi-Agent Coordinated Exploration by Novelty Sharing (AAAI 2024 · 6 citations) | |
| Remove the return-to-go prompt component | SMAC MMM2, 6h_vs_8z, and corridor · performance | Hurt | Imagine, Initialize, and Explore: An Effective Exploration Method in Multi-Agent Reinforcement Learning (AAAI 2024 · 2 citations) | |
| replace policy entropy incentive with conditional CS policy divergence | MA-Mujoco and Bi-DexHands · episodic reward | Helped | Measuring Mutual Policy Divergence for Multi-Agent Sequential Exploration (NeurIPS 2024 · 1 citation) | |
| add individual exploration only (constrain temperature to be positive) | SMAC MMM2 and 6h_vs_8z · performance | Hurt | AIR: Unifying Individual and Collective Exploration in Cooperative Multi-Agent Reinforcement Learning (AAAI 2025 · 0 citations) | |
| Remove the 1-Lipschitz constraint | SMAC · performance | Hurt | Learning Joint Behaviors with Large Variations (AAAI 2025 · 0 citations) | |
| Remove KL regularization | LBF task (45th time step) · agent-ID classification accuracy on t-SNE embeddings | 57.5% -> 99.3% | Hurt | Enhancing Cooperative Multi-Agent Reinforcement Learning with State Modelling and Adversarial Exploration (ICML 2025 · 0 citations) |
| remove exploration loss | VMAS Flocki* · mean episodic reward | 1.06 (±0.12) -> 0.82 (±0.15) | Hurt | Wonder Wins Ways: Curiosity-Driven Exploration through Multi-Agent Contextual Calibration (NeurIPS 2025 · 0 citations) |
| replace individual contributions with a global contribution scaffold | GRF 3_vs_1_with_keeper · winning rate | over 60% -> 40% | Hurt | Individual Contributions as Intrinsic Exploration Scaffolds for Multi-agent Reinforcement Learning (ICML 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 8 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- SMACv2 · 3 of 8 papers
- SMAC · 2 of 8 papers
- LBF · 2 of 8 papers
Baselines
- QMIX · 3 of 8 papers
- MAVEN · 2 of 8 papers
- EMC · 2 of 8 papers
- EOI · 2 of 8 papers
- MAPPO · 2 of 8 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 23 papers; from OpenAlex and the papers' affiliations).
- Peking University · 4 papers
- Chinese Academy of Sciences · 3 papers
- University of Chinese Academy of Sciences · 2 papers
- Xi’an Jiaotong University · 2 papers
- Duke University · 2 papers
- University of Science and Technology of China · 2 papers
Use it
Download BibTeX of 30 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Cooperative MARL algorithms (149 papers)
- MARL benchmarks and grouping (121 papers)
- Learned inter-agent communication (58 papers)
Elsewhere in the topic
- Decentralized actor-critic MARL (77 papers)
- Offline and networked MARL (65 papers)
- Coordination graphs (50 papers)
- Value factorization (39 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).