Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Learning to Cooperate with Minimal Observability
Chin-wing Leung, Paolo Turrini, Fernando P. Santos, Mirco Musolesi
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 mutual partner-selection update (selected agent also learns) (ablation) | Local Information Model · Prisoner's Dilemma, 30 agents | mutual cooperation (%): below 20% -> above 40% | Improved |
| add global information with probability pG (ablation) | Local Information Model · Prisoner's Dilemma | mutual cooperation (%): stabilises around 85% at pG = 0.3 | Improved |
| add one random third-party observer per gameplay | Local Information Model · Prisoner's Dilemma, 30 agents | mutual cooperation (%): over 90% of interactions mutual cooperation | Improved |
| add one observer per gameplay | Local Information Model · Prisoner's Dilemma, 30 agents | distinct partners selected per agent: 1.5 -> 4.26 | Improved |
| increase population size to 50 (ablation) | Observer Model · Prisoner's Dilemma, 50 agents | mutual cooperation (%): drops to about 85% | Hurt |
| increase observers from 1 to 2 (ablation) | Observer Model · Prisoner's Dilemma, 50 agents | mutual cooperation (%): 85% -> above 90% | Improved |
| lower exploration rate (0.005) (ablation) | Observer Model · Prisoner's Dilemma, 30 agents | convergence speed: cooperation starts to grow at 25,000 instead of 12,500 | Hurt |
| raise learning and exploration rates (ablation) | Observer Model · Prisoner's Dilemma, 30 agents | mutual cooperation (%): faster convergence, but cooperation kept below 80% | Mixed |
Closest papers by meaning
- Multi-agent cooperation through learning-aware policy gradients (ICLR 2025 · similarity 0.91)
- Training Generalizable Collaborative Agents via Strategic Risk Aversion (NeurIPS 2026 · similarity 0.89)
- Cross-environment Cooperation Enables Zero-shot Multi-agent Coordination (ICML 2025 · similarity 0.88)
- Reciprocal Reward Influence Encourages Cooperation From Self-Interested Agents (NeurIPS 2024 · similarity 0.88)
- Enhancing Cooperative Multi-Agent Reinforcement Learning with State Modelling and Adversarial Exploration (ICML 2025 · similarity 0.87)
- Diversity Is Not All You Need: Training A Robust Cooperative Agent Needs Specialist Partners (NeurIPS 2024 · 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.
