Pith. sign in

REVIEW 5 cited by

Counterfactual Multi-Agent Policy Gradients

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1705.08926 v3 pith:4IRCROHV submitted 2017-05-24 cs.AI cs.MA

classification cs.AIcs.MA
keywords multi-agentcomacounterfactualagentsdecentralisedusesactor-criticbaseline
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Cooperative multi-agent systems can be naturally used to model many real world problems, such as network packet routing and the coordination of autonomous vehicles. There is a great need for new reinforcement learning methods that can efficiently learn decentralised policies for such systems. To this end, we propose a new multi-agent actor-critic method called counterfactual multi-agent (COMA) policy gradients. COMA uses a centralised critic to estimate the Q-function and decentralised actors to optimise the agents' policies. In addition, to address the challenges of multi-agent credit assignment, it uses a counterfactual baseline that marginalises out a single agent's action, while keeping the other agents' actions fixed. COMA also uses a critic representation that allows the counterfactual baseline to be computed efficiently in a single forward pass. We evaluate COMA in the testbed of StarCraft unit micromanagement, using a decentralised variant with significant partial observability. COMA significantly improves average performance over other multi-agent actor-critic methods in this setting, and the best performing agents are competitive with state-of-the-art centralised controllers that get access to the full state.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 478 citations worldwide. Full citation record

  1. MARS-RA: Rank Aggregation for Credit Assignment via Multimodal Comparisons in Embodied Multi-Agent Cooperation

    cs.AI 2026-07 conditional novelty 6.0 of 10

    Credit assignment via LMM pairwise comparisons plus Bradley–Terry rank aggregation and potential-based shaping improves cooperative MARL under sparse rewards and dynamic agent counts.

  2. Training Large Language Models for Self-Explanation Faithfulness

    cs.LG 2026-07 conditional novelty 6.0 of 10

    RL fine-tuning with a counterfactual mention/influence reward raises LLM self-explanation faithfulness (Phi-CCT) from near zero to ~0.66 in-distribution for two 8B models, with partial transfer to held-out tasks.

  3. Low-Rank Agent-Specific Adaptation (LoRASA) for Multi-Agent Policy Learning

    cs.MA 2025-02 conditional novelty 6.0 of 10

    Per-agent low-rank adapters on a shared backbone let multi-agent policies specialize at a fraction of the memory cost of separate networks, with competitive benchmark performance.

  4. Self-Supervised Goal-Reaching Results in Multi-Agent Cooperation and Exploration

    cs.LG 2025-09 conditional novelty 5.0 of 10

    Self-supervised multi-agent goal-reaching, where each agent independently learns a contrastive critic of its own observations, achieves cooperation and exploration in sparse-reward MARL tasks where standard baselines fail.

  5. Multi-Agent Reinforcement Learning in Cybersecurity: From Fundamentals to Applications

    cs.MA 2025-05 conditional novelty 3.0 of 10

    A narrative survey of multi-agent reinforcement learning for cyber defense, reviewing game-theoretic models, cyber gyms, and applications, concluding MARL is promising but faces scalability and simulation-to-real tran...

Pith tools