Research map / Multi-agent systems and games / Cooperative multi-agent RL
UneVEn: Universal Value Exploration for Multi-Agent Reinforcement 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 |
|---|---|---|---|
| adding UneVEn related-task action selection (greedy or uniform GPI) instead of target-only GPI (ablation) | MAUSFs exploration · Predator-prey, higher penalty p in {0.012, 0.016} | success on target task | Helped |
| adding GPI policy at action selection (vs NOGPI) (ablation) | UneVEn action selection · Predator-prey ablation | success | Helped |
| adding a penalty p>0 on miscoordinated captures (higher RO) to the target task | VDN and QMIX · Predator-prey p in {0.008,0.012,0.016} | success | Hurt |
| adding UneVEn greedy/uniform GPI exploration on the target task | VDN/QMIX-style target task with RO · Predator-prey p in {0.004..0.016} | success | Helped |
| uniform (vs greedy) related-task selection at high penalty (ablation) | UneVEn-Uniform-GPI · Predator-prey p=0.016 | variance across seeds | Hurt |
| training on p=0.004 and zero-shot testing with NOGPI policy | UneVEn-Greedy/Uniform-GPI · Zero-shot predator-prey p in {0.2,0.5} | success | Helped |
| SMAC with ally penalty p=1.0 (RO-inducing reward) instead of default | VDN, QMIX, QTRAN, WQMIX, QPLEX · SMAC p=1.0 | win rate | Hurt |
| adding UneVEn greedy GPI exploration | SMAC maps with p=1.0 · SMAC p=1.0 | win rate | Helped |
| adding UneVEn action selection on SMAC low penalty p=0.5 | UneVEn vs VDN/QMIX · SMAC p=0.5 | win rate | No clear effect |
| adding the penalty p=0.004 on predator-prey (low RO) | VDN · Predator-prey p=0.004 | success | No clear effect |
Closest papers by meaning
- MAVEN: Multi-Agent Variational Exploration (NeurIPS 2019 · similarity 0.89)
- Revisiting Some Common Practices in Cooperative Multi-Agent Reinforcement Learning (ICML 2022 · similarity 0.89)
- Learning Joint Behaviors with Large Variations (AAAI 2025 · similarity 0.88)
- QTRAN: Learning to Factorize with Transformation for Cooperative Multi-Agent Reinforcement Learning (ICML 2019 · similarity 0.88)
- Locality Matters: A Scalable Value Decomposition Approach for Cooperative Multi-Agent Reinforcement Learning (AAAI 2022 · similarity 0.88)
- 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.