Research map / Multi-agent systems and games / Cooperative multi-agent RL
Regularized Softmax Deep 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 |
|---|---|---|---|
| add regularization and softmax operator (RES) to QMIX | QMIX · multi-agent particle tasks (PP, PD, W, CC) | return / overestimation bias | Helped |
| add regularizer to QMIX (RE-QMIX vs QMIX) (ablation) | RES-QMIX · predator-prey | mean normalized return | Helped |
| add softmax operator on top of regularizer (ablation) | RES-QMIX · predator-prey | learning efficiency | Helped |
| use proposed softmax approximation instead of direct computation over joint actions (ablation) | RES-QMIX · predator-prey | runtime (hours) and performance: 10.7 -> 5.2 hours | Helped |
| use proposed approximation instead of random sampling of joint actions (ablation) | RES-QMIX · predator-prey | mean normalized return | Helped |
| vary inverse temperature beta (ablation) | RES-QMIX · predator-prey | mean normalized return | No clear effect |
| vary regularization coefficient lambda (too small or too large) (ablation) | RES-QMIX · predator-prey | mean normalized return | Mixed |
| apply RES to Weighted QMIX and QPLEX | Weighted QMIX, QPLEX · multi-agent particle tasks | mean normalized return | Helped |
| apply RES-QMIX instead of QMIX | QMIX · SMAC (2s3z, 3s5z, 2c_vs_64zg, MMM2) | test win rate | Helped |
Closest papers by meaning
- Inverse Factorized Soft Q-Learning for Cooperative Multi-agent Imitation Learning (NeurIPS 2024 · similarity 0.90)
- High-order Interactions Modeling for Interpretable Multi-Agent Q-Learning (NeurIPS 2025 · similarity 0.88)
- Learning Joint Behaviors with Large Variations (AAAI 2025 · similarity 0.88)
- Value-Based Deep Multi-Agent Reinforcement Learning with Dynamic Sparse Training (NeurIPS 2024 · similarity 0.87)
- Revisiting Cooperative Off-Policy Multi-Agent Reinforcement Learning (ICML 2025 · similarity 0.87)
- Maximum Entropy Heterogeneous-Agent Reinforcement Learning (ICLR 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.