Research map / Multi-agent systems and games / Cooperative multi-agent RL
Concentration Network for Reinforcement Learning of Large-Scale Multi-Agent Systems
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 motivation-driven ranking, pruning, and aggregation | Conc · DCA 50vs50-S | win rate: 90% vs below 30% | Helped |
| remove the survival-time motivation index (ablation) | Conc · DCA | performance | Hurt |
| increase d_c (ablation) | Conc · DCA | final episode reward | No clear effect |
| add a self-attention layer (ablation) | Conc · DCA 50vs50 | performance | Helped |
| add a self-attention layer (ablation) | Conc · DCA 100vs100 | win rate | Hurt |
| use Single-ConcNet instead of Dual-ConcNet (ablation) | Conc · DCA | performance | No clear effect |
| use past observations with a FIFO memory pool (ablation) | Conc · DCA with 80% interference | performance: 80% | Helped |
| replace ConcNet aggregation with top-d_c soft attention weights (ablation) | Conc · DCA 50vs50 | win rate | Hurt |
| increase observation interference | attention-based methods · DCA 50vs50-L | win rate | Helped |
| train on 100vs100 and test with equal team sizes | Conc · DCA 130vs130 | win rate: about 50% | Helped |
| use an independent critic head to estimate state value (ablation) | Single-ConcNet · DCA | model performance | No clear effect |
| remove the survival-time motivation index (ablation) | Conc · DCA with stronger observation interference | performance | Hurt |
Closest papers by meaning
- Mixture-of-Experts Meets In-Context Reinforcement Learning (NeurIPS 2025 · similarity 0.85)
- HGAP: Boosting Permutation Invariant and Permutation Equivariant in Multi-Agent Reinforcement Learning via Graph Attention Network (ICML 2024 · similarity 0.84)
- Mastering Massive Multi-Task Reinforcement Learning via Mixture-of-Expert Decision Transformer (ICML 2025 · similarity 0.84)
- HYGMA: Hypergraph Coordination Networks with Dynamic Grouping for Multi-Agent Reinforcement Learning (ICML 2025 · similarity 0.84)
- Mixtures of Experts Unlock Parameter Scaling for Deep RL (ICML 2024 · similarity 0.84)
- Efficient Reinforcement Learning by Discovering Neural Pathways (NeurIPS 2024 · similarity 0.84)
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.