Research map / Multi-agent systems and games / Games and equilibria
MAESTRO: Open-Ended Environment Design for Multi-Agent Reinforcement Learning
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 a joint regret-based curriculum over environment–co-player pairs | MAESTRO · LaserTag, OOD human-designed levels | round-robin return | Helped |
| Use a joint regret-based curriculum over environment–co-player pairs | MAESTRO · MultiCarRacing, F1 tracks | robustness | Helped |
| Train on randomized environments with a joint autocurriculum instead of only the target environment | MAESTRO versus specialist agents · LaserTag and MultiCarRacing target environments | cross-play performance | Helped |
| Select co-players uniformly at random instead of prioritizing them by per-environment regret (ablation) | MAESTRO-R · LaserTag and MultiCarRacing | sample-efficiency and robustness | Hurt |
| Select co-players using PFSP win-rate prioritization instead of per-environment regret (ablation) | MAESTRO-P · LaserTag and MultiCarRacing | sample-efficiency and robustness | Hurt |
Closest papers by meaning
- Improving Regret Approximation for Unsupervised Dynamic Environment Generation (NeurIPS 2025 · similarity 0.87)
- Foundations of Multi-Agent Learning in Dynamic Environments: Where Reinforcement Learning Meets Strategic Decision-Making (AAAI 2025 · similarity 0.87)
- Unsupervised Training Sequence Design: Efficient and Generalizable Agent Training (AAAI 2024 · similarity 0.87)
- Cross-environment Cooperation Enables Zero-shot Multi-agent Coordination (ICML 2025 · similarity 0.87)
- No Regrets: Investigating and Improving Regret Approximations for Curriculum Discovery (NeurIPS 2024 · similarity 0.87)
- Adversarial Environment Design via Regret-Guided Diffusion Models (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.