Overcooked

The 'Overcooked' dataset/benchmark is an environment used to evaluate zero-shot coordination performance among multi-agent systems, focusing on the ability of agents to adapt to unseen teammates.

Papers using Overcooked (16)

HDDLGym: A Tool for Studying Multi-Agent Hierarchical Problems Defined in HDDL with OpenAI Gym2025 · 1 citesZero Shot Coordination for Sparse Reward Tasks with Diverse Reward Shapings2026Theory of Mind Guided Strategy Adaptation for Zero-Shot Coordination2026Probing Dec-POMDP Reasoning in Cooperative MARL2026Synergizing Code Coverage and Gameplay Intent: Coverage-Aware Game Playtesting with LLM-Guided Reinforcement Learning2025Modulation of temporal decision-making in a deep reinforcement learning agent under the dual-task paradigm2025Zero-Shot Coordination in Ad Hoc Teams with Generalized Policy Improvement and Difference Rewards2025NiceWebRL: a Python library for human subject experiments with reinforcement learning environments2025Fixing Incomplete Value Function Decomposition for Multi-Agent Reinforcement Learning2025Verco: Learning Coordinated Verbal Communication for Multi-agent Reinforcement Learning2024 · 1 citesMulti-Agent Transfer Learning via Temporal Contrastive Learning2024 · 1 citesME-IGM: Individual-Global-Max in Maximum Entropy Multi-Agent Reinforcement Learning2024 · 1 citesLearning to Cooperate with Humans using Generative Agents2024 · 1 citesBCR-DRL: Behavior- and Context-aware Reward for Deep Reinforcement Learning in Human-AI Coordination2024Role Play: Learning Adaptive Role-Specific Strategies in Multi-Agent Interactions2024Learning to Assist Humans without Inferring Rewards2024