Research map / Multi-agent systems and games / Cooperative multi-agent RL
Learning Individually Inferred 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 |
|---|---|---|---|
| Use I2C's individually inferred communication instead of MADDPG without communication | MADDPG · Cooperative Navigation | Reward: -1.26 -> -0.73 | Helped |
| Add correlation regularization (ablation) | I2C · Cooperative Navigation | Reward: -0.77 -> -0.73 | Helped |
| Use full communication with observed agents instead of no communication (ablation) | MADDPG · Cooperative Navigation | Reward: -1.26 -> -1.08 | Helped |
| Use random communication instead of no communication (ablation) | MADDPG · Cooperative Navigation | Reward: -1.26 -> -1.30 | Hurt |
| Select communication partners using I2C rather than communicating with all observed agents (ablation) | MADDPG · Cooperative Navigation | Reward: -1.08 -> -0.73 | Helped |
| Select communication partners using I2C rather than random communication at nearly the same communication amount (ablation) | MADDPG · Cooperative Navigation | Reward: -1.30 -> -0.73 | Helped |
| Use I2C's individually inferred communication instead of MADDPG without communication | MADDPG · Predator Prey | Mean reward: -0.98 -> -0.83 | Helped |
| Add I2C communication control to TarMAC | TarMAC · Traffic Junction medium | Success rate: 97.60% -> 97.92% | Helped |
| Add I2C communication control to TarMAC | TarMAC · Traffic Junction hard | Success rate: 89.24% -> 92.17% | Helped |
| Reduce TarMAC communication using I2C communication control | TarMAC · Traffic Junction medium | Communication overhead: nearly 34% | Helped |
| Reduce TarMAC communication using I2C communication control | TarMAC · Traffic Junction hard | Communication overhead: nearly 37% | Helped |
| Lower the causal-effect threshold from 90% to 70% (ablation) | I2C · Cooperative Navigation | Final reward: 90% to 70% | Helped |
Closest papers by meaning
- Efficient Communication in Multi-Agent Reinforcement Learning with Implicit Consensus Generation (AAAI 2025 · similarity 0.90)
- Multi-Agent Reinforcement Learning with Communication-Constrained Priors (NeurIPS 2025 · similarity 0.90)
- Learning Multi-Agent Communication from Graph Modeling Perspective (ICLR 2024 · similarity 0.89)
- T2MAC: Targeted and Trusted Multi-Agent Communication through Selective Engagement and Evidence-Driven Integration (AAAI 2024 · similarity 0.89)
- Expressive Multi-Agent Communication via Identity-Aware Learning (AAAI 2024 · similarity 0.88)
- Learning Multi-Agent Communication with Contrastive Learning (ICLR 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.