Research map / Multi-agent systems and games / Human-AI and goal-conditioned
TEAMSTER: Model-Based Reinforcement Learning for Ad Hoc Teamwork (Abstract Reprint)
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes TEAMSTER, a model-based RL approach for ad hoc teamwork that learns separate models of the environment and of teammates' behavior, and compares it to PLASTIC across four domains.
Problem
Ad hoc teamwork, where an agent must cooperate with previously unknown teammates without prior coordination.
Approach
Learn the environment model and the teammate behavior model separately within a model-based RL framework, rather than relying on model-free approaches such as PLASTIC.
Key results
- TEAMSTER is more flexible than PLASTIC across four domains from the multi-agent systems literature, according to the abstract (the abstract is truncated, so specific numbers are not available).
Experimental setup
Benchmarks: four multi-agent systems domains (unspecified in abstract)
Baselines: PLASTIC
Builds on: PLASTIC
Closest papers by meaning
- N-agent Ad Hoc Teamwork (NeurIPS 2024 · similarity 0.92)
- Ad Hoc Teamwork via Offline Goal-Based Decision Transformers (ICML 2025 · similarity 0.89)
- Open Ad Hoc Teamwork with Cooperative Game Theory (ICML 2024 · similarity 0.89)
- ROTATE: Regret-driven Open-ended Training for Ad Hoc Teamwork (NeurIPS 2026 · similarity 0.88)
- FightLadder: A Benchmark for Competitive Multi-Agent Reinforcement Learning (ICML 2024 · similarity 0.87)
- COSAC: Counterfactual Credit Assignment in Sequential Cooperative Teams (NeurIPS 2026 · 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.