mlsanityData to 7 Oct 2026

Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning

arXiv 1910.10897 ↗First posted 24 Oct 2019cs.LG cs.AI cs.ROcited by 1,168

Not on the map. Why?

Cited by 165 in Offline RL and imitation learning from data, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Meta-reinforcement learning algorithms can enable robots to acquire new skills much more quickly, by leveraging prior experience to learn how to learn. However, much of the current research on meta-reinforcement learning focuses on task distributions that are very narrow. For example, a commonly used meta-reinforcement learning benchmark uses different running velocities for a simulated robot as different tasks. When policies are meta-trained on such narrow task distributions, they cannot possibly generalize to more quickly acquire entirely new tasks. Therefore, if the aim of these methods is to enable faster acquisition of entirely new behaviors, we must evaluate them on task distributions that are sufficiently broad to enable generalization to new behaviors. In this paper, we propose an open-source simulated benchmark for meta-reinforcement learning and multi-task learning consisting of 50 distinct robotic manipulation tasks. Our aim is to make it possible to develop algorithms that generalize to accelerate the acquisition of entirely new, held-out tasks. We evaluate 7 state-of-the-art meta-reinforcement learning and multi-task learning algorithms on these tasks. Surprisingly, while each task and its variations (e.g., with different object positions) can be learned with reasonable success, these algorithms struggle to learn with multiple tasks at the same time, even with as few as ten distinct training tasks. Our analysis and open-source environments pave the way for future research in multi-task learning and meta-learning that can enable meaningful generalization, thereby unlocking the full potential of these methods.

Key facts. Cited by 1,168 papers in the map (data through 7 Oct 2026). First posted on arXiv 24 Oct 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2020: 4; 2021: 101; 2022: 131; 2023: 184; 2024: 250; 2025: 257; 2026: 241.

1,168 papers.

Sort:

By area:

  1. 2022RT-1: Robotics Transformer for Real-World Control at Scalecited by 2,007
  2. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,338
  3. 2022A Generalist Agentcited by 802
  4. 2021What Matters in Learning from Offline Human Demonstrations for Robot Manipulationcited by 699
  5. 2022R3M: A Universal Visual Representation for Robot Manipulationcited by 678
  6. 2023BridgeData V2: A Dataset for Robot Learning at Scalecited by 634
  7. 2021CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Taskscited by 564
  8. 2022Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulationcited by 545
  9. 2025SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Roboticscited by 507
  10. 2024Evaluating Real-World Robot Manipulation Policies in Simulationcited by 484
  11. 2023TD-MPC2: Scalable, Robust World Models for Continuous Controlcited by 482
  12. 2025RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulationcited by 477
38 more
  1. 2021Conflict-Averse Gradient Descent for Multi-task Learningcited by 441
  2. 2022VIMA: General Robot Manipulation with Multimodal Promptscited by 403
  3. 2022VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Trainingcited by 384
  4. 2022Temporal Difference Learning for Model Predictive Controlcited by 375
  5. 2021COMBO: Conservative Offline Model-Based Policy Optimizationcited by 365
  6. 2022RLPrompt: Optimizing Discrete Text Prompts with Reinforcement Learningcited by 361
  7. 2024TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulationcited by 331
  8. 2023ManiSkill2: A Unified Benchmark for Generalizable Manipulation Skillscited by 313
  9. 2023MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrationscited by 308
  10. 2021PEBBLE: Feedback-Efficient Interactive Reinforcement Learning via Relabeling Experience and Unsupervised Pre-trainingcited by 298
  11. 2021MetaDrive: Composing Diverse Driving Scenarios for Generalizable Reinforcement Learningcited by 281
  12. 2024Video Prediction Policy: A Generalist Robot Policy with Predictive Visual Representationscited by 279
  13. 2023EmbodiedGPT: Vision-Language Pre-Training via Embodied Chain of Thoughtcited by 276
  14. 2024A Survey on Vision-Language-Action Models for Embodied AIcited by 274
  15. 2022Masked Visual Pre-training for Motor Controlcited by 246
  16. 2022Contrastive Learning as Goal-Conditioned Reinforcement Learningcited by 241
  17. 2023Orbit: A Unified Simulation Framework for Interactive Robot Learning Environmentscited by 239
  18. 2021MT-Opt: Continuous Multi-Task Robotic Reinforcement Learning at Scalecited by 226
  19. 2023Habitat 3.0: A Co-Habitat for Humans, Avatars and Robotscited by 220
  20. 2023LVLM-eHub: A Comprehensive Evaluation Benchmark for Large Vision-Language Modelscited by 217
  21. 2024OGBench: Benchmarking Offline Goal-Conditioned RLcited by 216
  22. 2023Where are we in the search for an Artificial Visual Cortex for Embodied Intelligence?cited by 215
  23. 2022Multi-Game Decision Transformerscited by 212
  24. 2022Multi-Task Learning as a Bargaining Gamecited by 209
  25. 2023RoboGen: Towards Unleashing Infinite Data for Automated Robot Learning via Generative Simulationcited by 206
  26. 2021ManiSkill: Generalizable Manipulation Skill Benchmark with Large-Scale Demonstrationscited by 193
  27. 2021iGibson 2.0: Object-Centric Simulation for Robot Learning of Everyday Household Taskscited by 191
  28. 2023LIV: Language-Image Representations and Rewards for Robotic Controlcited by 180
  29. 2023Scaling Robot Learning with Semantically Imagined Experiencecited by 175
  30. 2023RH20T: A Comprehensive Robotic Dataset for Learning Diverse Skills in One-Shotcited by 173
  31. 2021Multi-Task Reinforcement Learning with Context-based Representationscited by 169
  32. 2021Maximum Entropy RL (Provably) Solves Some Robust RL Problemscited by 169
  33. 2022Masked World Models for Visual Controlcited by 169
  34. 2023RoboAgent: Generalization and Efficiency in Robot Manipulation via Semantic Augmentations and Action Chunkingcited by 167
  35. 2020OPAL: Offline Primitive Discovery for Accelerating Offline Reinforcement Learningcited by 164
  36. 2021BEHAVIOR: Benchmark for Everyday Household Activities in Virtual, Interactive, and Ecological Environmentscited by 161
  37. 2023Learning to Act from Actionless Videos through Dense Correspondencescited by 159
  38. 2023Text2Reward: Reward Shaping with Language Models for Reinforcement Learningcited by 153

50 most cited of 1,168 All 1,168 (JSON)