Awesome Paper List
We collect most of the existing MARL algorithms based on the multi-agent environment they choose to conduct on, with tag to annotate the sub-topic.
[B] Basic [S] Information Sharing [RG] Behavior/Role Grouping [I] Imitation [G] Graph [E] Exploration [R] Robust [P] Reward Shaping [F] Offline [T] Tree Search [MT] Multi-task
MPE
Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments [B][2017]
Learning attentional communication for multi-agent cooperation [S][2018]
learning when to communicate at scale in multiagent cooperative and competitive tasks [S][2018]
Qtran: Learning to factorize with transformation for cooperative multi-agent reinforcement learning [B][2019]
Robust multi-agent reinforcement learning via minimax deep deterministic policy gradient [R][2019]
Learning Individually Inferred Communication for Multi-Agent Cooperation [S][2020]
Multi-Agent Game Abstraction via Graph Attention Neural Network [G+S][2020]
Promoting Coordination through Policy Regularization in Multi-Agent Deep Reinforcement Learning [E][2020]
Robust Multi-Agent Reinforcement Learning with Model Uncertainty [R][2020]
Shared Experience Actor-Critic for Multi-Agent Reinforcement Learning [B][2020]
Cooperative Exploration for Multi-Agent Deep Reinforcement Learning [E][2021]
Multiagent Adversarial Collaborative Learning via Mean-Field Theory [R][2021]
The Surprising Effectiveness of PPO in Cooperative, Multi-Agent Games [B][2021]
Variational Automatic Curriculum Learning for Sparse-Reward Cooperative Multi-Agent Problems [2021]
ToM2C: Target-oriented Multi-agent Communication and Cooperation with Theory of Mind [2021]
SPD: Synergy Pattern Diversifying Oriented Unsupervised Multi-agent Reinforcement Learning [2022]
Distributional Reward Estimation for Effective Multi-Agent Deep Reinforcement Learning [2022]
SMAC
Value-Decomposition Networks For Cooperative Multi-Agent Learning [B][2017]
Multi-Agent Common Knowledge Reinforcement Learning [RG+S][2018]
QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning [B][2018]
Efficient Communication in Multi-Agent Reinforcement Learning via Variance Based Control [S][2019]
Exploration with Unreliable Intrinsic Reward in Multi-Agent Reinforcement Learning [P+E][2019]
Learning nearly decomposable value functions via communication minimization [S][2019]
Liir: Learning individual intrinsic reward in multi-agent reinforcement learning [P][2019]
Counterfactual Multi-Agent Reinforcement Learning with Graph Convolution Communication [S+G][2020]
Deep implicit coordination graphs for multi-agent reinforcement learning [G][2020]
DOP: Off-policy multi-agent decomposed policy gradients [B][2020]
From few to more Large-scale dynamic multiagent curriculum learning [MT][2020]
Learning structured communication for multi-agent reinforcement learning [S+G][2020]
Learning efficient multi-agent communication: An information bottleneck approach [S][2020]
On the robustness of cooperative multi-agent reinforcement learning [R][2020]
Qatten: A general framework for cooperative multiagent reinforcement learning [B][2020]
Revisiting parameter sharing in multi-agent deep reinforcement learning [RG][2020]
ROMA: Multi-Agent Reinforcement Learning with Emergent Roles [RG][2020]
Towards Understanding Cooperative Multi-Agent Q-Learning with Value Factorization [B][2021]
Contrasting centralized and decentralized critics in multi-agent reinforcement learning [B][2021]
Learning in nonzero-sum stochastic games with potentials [B][2021]
Natural emergence of heterogeneous strategies in artificially intelligent competitive teams [S+G][2021]
Rode: Learning roles to decompose multi-agent tasks [RG][2021]
SMIX(λ): Enhancing Centralized Value Functions for Cooperative Multiagent Reinforcement Learning [B][2021]
Tesseract: Tensorised Actors for Multi-Agent Reinforcement Learning [B][2021]
The Surprising Effectiveness of PPO in Cooperative, Multi-Agent Games [B][2021]
UPDeT: Universal Multi-agent Reinforcement Learning via Policy Decoupling with Transformers [MT][2021]
Randomized Entity-wise Factorization for Multi-Agent Reinforcement Learning [MT][2021]
Cooperative Multi-Agent Transfer Learning with Level-Adaptive Credit Assignment [MT][2021]
Uneven: Universal value exploration for multi-agent reinforcement learning [B][2021]
RMIX: Learning Risk-Sensitive Policies for Cooperative Reinforcement Learning Agents [2021]
Policy Regularization via Noisy Advantage Values for Cooperative Multi-agent Actor-Critic methods [2021]
ALMA: Hierarchical Learning for Composite Multi-Agent Tasks [2022]
Rethinking Individual Global Max in Cooperative Multi-Agent Reinforcement Learning [2022]
Surprise Minimizing Multi-Agent Learning with Energy-based Models [2022]
SHAQ: Incorporating Shapley Value Theory into Multi-Agent Q-Learning [2022]
Self-Organized Group for Cooperative Multi-agent Reinforcement Learning [2022]
Efficient Multi-agent Communication via Self-supervised Information Aggregation [2022]
Episodic Multi-agent Reinforcement Learning with Curiosity-Driven Exploration [2022]
CTDS: Centralized Teacher with Decentralized Student for Multi-Agent Reinforcement Learning [2022]
MAMuJoCo
FACMAC: Factored Multi-Agent Centralised Policy Gradients [B][2020]
Trust Region Policy Optimisation in Multi-Agent Reinforcement Learning [B][2021]
A Game-Theoretic Approach to Multi-Agent Trust Region Optimization [2021]
Settling the Variance of Multi-Agent Policy Gradients [2021]
Graph-Assisted Predictive State Representations for Multi-Agent Partially Observable Systems [2022]
Google Research Football
Adaptive Inner-reward Shaping in Sparse Reward Games [P][2020]
Factored action spaces in deep reinforcement learning [B][2021]
Semantic Tracklets An Object-Centric Representation for Visual Multi-Agent Reinforcement Learning [B][2021]
TiKick: Towards Playing Multi-agent Football Full Games from Single-agent Demonstrations [F][2021]
Celebrating Diversity in Shared Multi-Agent Reinforcement Learning [2021]
Mingling Foresight with Imagination: Model-Based Cooperative Multi-Agent Reinforcement Learning [2022]
Pommerman
Using Monte Carlo Tree Search as a Demonstrator within Asynchronous Deep RL [I+T][2018]
Accelerating Training in Pommerman with Imitation and Reinforcement Learning [I][2019]
Agent Modeling as Auxiliary Task for Deep Reinforcement Learning [S][2019]
Backplay: man muss immer umkehren [I][2019]
Terminal Prediction as an Auxiliary Task for Deep Reinforcement Learning [B][2019]
Adversarial Soft Advantage Fitting Imitation Learning without Policy Optimization [B][2020]
Evolutionary Reinforcement Learning for Sample-Efficient Multiagent Coordination [B][2020]
LBF & RWARE
Shared Experience Actor-Critic for Multi-Agent Reinforcement Learning [B][2020]
Benchmarking Multi-Agent Deep Reinforcement Learning Algorithms in Cooperative Tasks [B][2021]
Learning Altruistic Behaviors in Reinforcement Learning without External Rewards [B][2021]
Scaling Multi-Agent Reinforcement Learning with Selective Parameter Sharing [RG][2021]
LIGS: Learnable Intrinsic-Reward Generation Selection for Multi-Agent Learning [2022]
MetaDrive
Hanabi
Bayesian Action Decoder for Deep Multi-Agent Reinforcement Learning [B][2019]
Re-determinizing MCTS in Hanabi [S+T][2019]
Joint Policy Search for Multi-agent Collaboration with Imperfect Information [T][20209]
Off-Belief Learning [B][2021]
The Surprising Effectiveness of PPO in Cooperative Multi-Agent Games [B][2021]
2021 Trajectory Diversity for Zero-Shot Coordination [B][2021]
MAgent
Other Tasks
Learning Fair Policies in Decentralized Cooperative Multi-Agent Reinforcement Learning [2020]
Contrasting Centralized and Decentralized Critics in Multi-Agent Reinforcement Learning [2021]
Learning to Ground Multi-Agent Communication with Autoencoders [2021]
Latent Variable Sequential Set Transformers For Joint Multi-Agent Motion Prediction [2021]
Learning to Share in Multi-Agent Reinforcement Learning [2021]
Resilient Multi-Agent Reinforcement Learning with Adversarial Value Decomposition [2021]
Multi-Agent Reinforcement Learning for Active Voltage Control on Power Distribution Networks [2021]
Multi-Agent Reinforcement Learning in Stochastic Networked Systems [2021]
Mirror Learning: A Unifying Framework of Policy Optimisation [2022]
E-MAPP: Efficient Multi-Agent Reinforcement Learning with Parallel Program Guidance [2022]
Shield Decentralization for Safe Multi-Agent Reinforcement Learning [2022]
Provably Efficient Offline Multi-agent Reinforcement Learning via Strategy-wise Bonus [2022]
Asynchronous Actor-Critic for Multi-Agent Reinforcement Learning [2022]
Near-Optimal Multi-Agent Learning for Safe Coverage Control [2022]