Research map / Multi-agent systems and games / Cooperative multi-agent RL
Universally Expressive Communication in Multi-Agent Reinforcement Learning
Matthew Morris, Thomas D Barrett, Arnu Pretorius
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 unique agent IDs to observations | IC3Net · Medium Traffic Junction | success rate: 0.971 -> 0.804 | Hurt |
| add unique agent IDs to observations | MAGIC · Easy Traffic Junction | success rate: 0.634 -> 0.764 | Improved |
| add 0.25 RNI to observations | MAGIC · Easy Traffic Junction | success rate: 0.634 -> 0.787 | Improved |
| add 0.25 RNI to observations | CommNet · Predator-Prey | success rate: 0.88 -> 0.476 | Hurt |
| add 0.25 RNI to observations | MAGIC · Predator-Prey | success rate: 0.892 -> 0.451 | Hurt |
| add 0.25 RNI to observations | IC3Net · Predator-Prey | success rate: 0.952 -> 0.933 | No clear effect |
| use 0.75 RNI instead of baseline | CommNet · Predator-Prey | success rate: 0.88 -> 0.194 | Hurt |
| use 0.75 RNI instead of baseline | CommNet · Medium Traffic Junction | success rate: 0.761 -> 0.046 | Hurt |
| add 0.25 RNI to observations | MAGIC · Box Pushing | ratio cleared: 0.958 -> 0.998 | Improved |
| add unique agent IDs to observations | IC3Net · Box Pushing | ratio cleared: 0.49 -> 0.617 | Improved |
| add 0.25 RNI to observations | IC3Net · Box Pushing | ratio cleared: 0.49 -> 0.676 | Improved |
| add unique agent IDs to observations | IC3Net · Drone Scatter (stochastic eval) | pairwise distance: 9.108 -> 13.3 | Improved |
Closest papers by meaning
- Expressive Multi-Agent Communication via Identity-Aware Learning (AAAI 2024 · similarity 0.92)
- Learning Efficient and Robust Multi-Agent Communication via Graph Information Bottleneck (AAAI 2024 · similarity 0.90)
- G-Designer: Architecting Multi-agent Communication Topologies via Graph Neural Networks (ICML 2025 · similarity 0.90)
- Bridging Training and Execution via Dynamic Directed Graph-Based Communication in Cooperative Multi-Agent Systems (AAAI 2025 · similarity 0.89)
- Learning Multi-Agent Communication from Graph Modeling Perspective (ICLR 2024 · similarity 0.88)
- Multi-Agent Reinforcement Learning with Communication-Constrained Priors (NeurIPS 2025 · 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.
