Research map / Multi-agent systems and games / Cooperative multi-agent RL
DFAC Framework: Factorizing the Value Function via Quantile Mixture for Multi-Agent Distributional Q-Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace VDN with its distributional DFAC variant DDN | VDN · SMAC Super Hard maps | median win rate (%): 0.00 -> 83.92; 89.20 -> 94.03; 89.20 -> 97.22; 63.12 -> 91.48; 85.34 -> 95.40 | Helped |
| Replace QMIX with its distributional DFAC variant DMIX | QMIX · SMAC Super Hard maps | median win rate (%): 12.78 -> 49.43; 67.22 -> 91.08; 92.44 -> 95.11; 84.77 -> 85.45; 37.61 -> 90.45 | Helped |
| Replace VDN with its distributional DFAC variant DDN | VDN · SMAC Super Hard maps | averaged score: 15.41 -> 19.40; 19.75 -> 20.94; 19.36 -> 20.90; 18.45 -> 19.71; 19.47 -> 20.00 | Helped |
| Replace QMIX with its distributional DFAC variant DMIX | QMIX · SMAC Super Hard maps | averaged score: 14.37 -> 17.14; 20.16 -> 19.70; 19.42 -> 19.87; 19.41 -> 19.43; 15.07 -> 19.66 | Mixed |
| Use IQN-based distributional learning instead of expected-value IQL | IQL · SMAC Super Hard maps | median win rate (%): 0.00 -> 0.00; 29.83 -> 62.22; 68.92 -> 85.23; 2.27 -> 6.02; 84.87 -> 91.62 | Mixed |
| Use IQN-based distributional learning instead of expected-value IQL | IQL · SMAC Super Hard maps | averaged score: 13.78 -> 14.94; 16.54 -> 17.52; 17.50 -> 19.21; 14.01 -> 14.45; 19.42 -> 19.68 | Helped |
Closest papers by meaning
- Towards Understanding Cooperative Multi-Agent Q-Learning with Value Factorization (NeurIPS 2021 · similarity 0.92)
- ResQ: A Residual Q Function-based Approach for Multi-Agent Reinforcement Learning Value Factorization (NeurIPS 2022 · similarity 0.91)
- PAC: Assisted Value Factorization with Counterfactual Predictions in Multi-Agent Reinforcement Learning (NeurIPS 2022 · similarity 0.89)
- QTRAN: Learning to Factorize with Transformation for Cooperative Multi-Agent Reinforcement Learning (ICML 2019 · similarity 0.89)
- Factored Value Functions for Graph-Based Multi-Agent Reinforcement Learning (ICML 2026 · similarity 0.89)
- ConcaveQ: Non-monotonic Value Function Factorization via Concave Representations in Deep Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.88)
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.