MetaWorld

MetaWorld is a benchmark that contains a collection of diverse robotic manipulation tasks used to evaluate the performance and robustness of world-model agents in continuous control settings.

Papers using MetaWorld (8)

ARB4WM: An Adversarial Robustness Benchmark for World Models in Continuous Control2026Centralized Reward Agent for Knowledge Sharing and Transfer in Multi-Task Reinforcement Learning2024 · 1 citesData-Efficient Multitask DAgger2025Goal-Oriented Skill Abstraction for Offline Multi-Task Reinforcement Learning2025Continual World: A Robotic Benchmark For Continual Reinforcement Learning2021 · 20 citesProcedural Generalization by Planning with Self-Supervised World Models2021 · 4 citesTask-Agnostic Continual Reinforcement Learning: Gaining Insights and Overcoming Challenges2022 · 2 citesHierarchical Transformers are Efficient Meta-Reinforcement Learners2024