Research map / Multi-agent systems and games / Cooperative multi-agent RL
Learning Attentional Communication for Multi-Agent Cooperation
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 ATOC's attention-selected, dynamic communication instead of no communication (ablation) | DDPG / ATOC · Cooperative Navigation (N=50, L=50) | mean reward: -0.14 -> -0.04 | Helped |
| add ATOC's attention-selected, dynamic communication instead of no communication (ablation) | DDPG / ATOC · Cooperative Navigation (N=50, L=50) | collisions: 32 -> 13 | Helped |
| add ATOC's attention-selected, dynamic communication instead of no communication (ablation) | DDPG / ATOC · Cooperative Navigation (N=50, L=50) | occupied landmarks: 22% -> 92% | Helped |
| replace full communication with attention-selected, dynamic communication | CommNet · Cooperative Navigation (N=50, L=50) | mean reward: -0.60 -> -0.04 | Helped |
| replace full communication with attention-selected, dynamic communication | BiCNet · Cooperative Navigation (N=100, L=100) | collisions: 91 -> 28 | Helped |
| add communication to ATOC (ablation) | ATOC · Cooperative Navigation (N=50, L=50) | mean reward: -0.22 -> -0.04 | Helped |
| add communication to ATOC (ablation) | ATOC · Cooperative Navigation (N=50, L=50) | collisions: 47 -> 13 | Helped |
| add communication to ATOC (ablation) | ATOC · Cooperative Navigation (N=50, L=50) | occupied landmarks: 40% -> 92% | Helped |
| add communication to ATOC (ablation) | ATOC · Cooperative Pushball | normalized mean reward: 0.86 -> 0.95 | Helped |
| replace no communication with ATOC's dynamic communication (ablation) | DDPG / ATOC · Cooperative Pushball | normalized mean reward: 0.50 -> 0.95 | Helped |
| replace full communication with attention-selected communication | CommNet · Cooperative Pushball | normalized mean reward: 0.71 -> 0.95 | Helped |
| use attention-selected communication rather than full connectivity | ATOC · Predator-Prey | episodes to stabilization: 2500 episodes -> 1000 episodes | Helped |
Closest papers by meaning
- Learning Multi-Agent Communication from Graph Modeling Perspective (ICLR 2024 · similarity 0.90)
- Efficient Communication in Multi-Agent Reinforcement Learning with Implicit Consensus Generation (AAAI 2025 · similarity 0.88)
- Multi-Agent Reinforcement Learning with Communication-Constrained Priors (NeurIPS 2025 · similarity 0.88)
- T2MAC: Targeted and Trusted Multi-Agent Communication through Selective Engagement and Evidence-Driven Integration (AAAI 2024 · similarity 0.87)
- Learning Efficient and Robust Multi-Agent Communication via Graph Information Bottleneck (AAAI 2024 · similarity 0.86)
- Learning Multi-Agent Communication with Contrastive Learning (ICLR 2024 · similarity 0.86)
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.