Research map / Multi-agent systems and games / Cooperative multi-agent RL
QPLEX: Duplex Dueling Multi-Agent 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 |
|---|---|---|---|
| Use duplex dueling value factorization with exact IGM consistency | Multi-agent Q-learning · StarCraft II, online and offline data collection | performance | Helped |
| Use duplex dueling value factorization | Multi-agent Q-learning · Hard matrix game | joint strategy optimality | Helped |
| Use richer value-function classes | Multi-agent Q-learning · Two-state MMDP | training stability | Helped |
| Increase the neural-network size used to learn importance weights (ablation) | QPLEX · Hard matrix game | convergence speed | Helped |
| Use a more complex importance-weight network (ablation) | QPLEX · Complex domains | learning efficiency | Hurt |
| Add the discrepancy-correction term to the Qatten-like joint value (ablation) | QPLEX · StarCraft II benchmark tasks | performance | Helped |
| Train with offline data without additional online exploration | QPLEX · StarCraft II offline data collection | performance | Helped |
Closest papers by meaning
- Multiagent Q-learning with Sub-Team Coordination (NeurIPS 2022 · similarity 0.90)
- QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning (ICML 2018 · similarity 0.89)
- Towards Understanding Cooperative Multi-Agent Q-Learning with Value Factorization (NeurIPS 2021 · similarity 0.89)
- Q-value Path Decomposition for Deep Multiagent Reinforcement Learning (ICML 2020 · similarity 0.88)
- LOQA: Learning with Opponent Q-Learning Awareness (ICLR 2024 · similarity 0.87)
- ConcaveQ: Non-monotonic Value Function Factorization via Concave Representations in Deep Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.87)
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.