Research map / Multi-agent systems and games / Cooperative multi-agent RL
Learning Nearly Decomposable Value Functions Via Communication Minimization
Tonghan Wang*, Jianhao Wang*, Chongyi Zheng, Chongjie Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Cut all messages (ablation) | NDQ · Independent search | average team reward per episode: 96.0 -> 96.0 | No clear effect |
| Cut all messages (ablation) | TarMAC + QMIX · Independent search | average team reward per episode: 96.0 -> 96.0 | No clear effect |
| Cut 80% of messages using learned message distributions (ablation) | NDQ · StarCraft II micromanagement | performance | No clear effect |
| Cut 80% of messages by removing the smallest attention weights (ablation) | QMIX + TarMAC · StarCraft II micromanagement | performance | Hurt |
| Cut all messages (ablation) | NDQ · StarCraft II micromanagement | win rate | Hurt |
| Increase the communication penalty to β=1.0 (ablation) | NDQ · Sensor | communication | Mixed |
| Decrease the communication penalty to β=10^-5 (ablation) | NDQ · Sensor | communication | Hurt |
| Use β=10^-3 for message regularization (ablation) | NDQ · Sensor | communication: 80% | Improved |
| Add learned communication to value function factorization | NDQ · StarCraft II micromanagement | performance | Improved |
| Add information-theoretic message constraints rather than attention-only communication | NDQ vs QMIX + TarMAC · StarCraft II micromanagement | performance | Improved |
| Add communication to QMIX using TarMAC | QMIX + TarMAC · StarCraft II micromanagement | performance | Hurt |
| Add minimal communication learned by NDQ | NDQ · Hallway | win rate: 100% | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 2 on most shared results.
- TarMAC — TarMAC: Targeted Multi-Agent Communication (ICML 2019) · better on 0 of 1 shared result
- QMIX — QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning (ICML 2018) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Towards Understanding Cooperative Multi-Agent Q-Learning with Value Factorization (NeurIPS 2021 · similarity 0.90)
- QTRAN: Learning to Factorize with Transformation for Cooperative Multi-Agent Reinforcement Learning (ICML 2019 · similarity 0.89)
- ConcaveQ: Non-monotonic Value Function Factorization via Concave Representations in Deep Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.88)
- Multi-Agent Reinforcement Learning with Communication-Constrained Priors (NeurIPS 2025 · similarity 0.88)
- High-order Interactions Modeling for Interpretable Multi-Agent Q-Learning (NeurIPS 2025 · similarity 0.86)
- RGMComm: Return Gap Minimization via Discrete Communications in Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
