Research map / Multi-agent systems and games / Cooperative multi-agent RL
IEC: When Information-Driven Exploration Meets Spectral Consensus via Primal–Dual Reward Regularization in Decentralized Multi-Agent RL
Xuefeng Du, Jiajun Wu, Yuduo Zheng, Fengqi Li
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 |
|---|---|---|---|
| add state-coverage novelty term (ablation) | IEC · GridWorld PassRoom | win rate: 42.8 -> 96.5 | Improved |
| add spectral consensus constraint (ablation) | IEC · SMAC 8m | win rate: 58.2 -> 90.5 | Improved |
| add spectral consensus constraint (ablation) | IEC · Overcooked Narrow | win rate: 56.2 -> 82.5 | Improved |
| add information-driven term (ablation) | IEC · Overcooked Narrow | win rate: 71.8 -> 82.5 | Improved |
| add information-driven term (ablation) | IEC · GridWorld PassRoom | win rate: 92.5 -> 96.5 | Improved |
| use adaptive primal-dual multiplier instead of fixed weight (ablation) | IEC · GridWorld PassRoom | win rate: 89.8 -> 96.5 | Improved |
| use adaptive primal-dual multiplier instead of fixed weight (ablation) | IEC · Overcooked Narrow | win rate: 77.5 -> 82.5 | Improved |
| use IEC instead of MACE (ablation) | Dec-MARL · GridWorld MultiRoom | win rate: 85.2 -> 95.2 | Improved |
| use IEC instead of MACE | Dec-MARL · Overcooked Large | win rate: 70.5 -> 90.2 | Improved |
| use IEC instead of MACE | Dec-MARL · GridWorld PassRoom | win rate: 98.8 -> 97.2 | No clear effect |
| use IEC instead of MACE | Dec-MARL · Overcooked Narrow | win rate: 82.2 -> 81.5 | Hurt |
| reduce sampling budget to N=8 | IEC · SMAC 8m | win rate: 87.5% WR vs. 84.5% for MACE | Improved |
Closest papers by meaning
- Settling Decentralized Multi-Agent Coordinated Exploration by Novelty Sharing (AAAI 2024 · similarity 0.90)
- Influence-Based Multi-Agent Exploration (ICLR 2020 · similarity 0.88)
- Randomized Exploration in Cooperative Multi-Agent Reinforcement Learning (NeurIPS 2024 · similarity 0.87)
- Intrinsic Action Tendency Consistency for Cooperative Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.87)
- Bridging Training and Execution via Dynamic Directed Graph-Based Communication in Cooperative Multi-Agent Systems (AAAI 2025 · similarity 0.87)
- Efficient Communication in Multi-Agent Reinforcement Learning with Implicit Consensus Generation (AAAI 2025 · 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.
