Meta-World+: An Improved, Standardized, RL Benchmark

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: McLean, Reginald, Chatzaroulas, Evangelos, McCutcheon, Luc, Röder, Frank, Yu, Tianhe, He, Zhanpeng, Zentner, K. R., Julian, Ryan, Terry, J K, Woungang, Isaac, Farsad, Nariman, Castro, Pablo Samuel
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918212980965376
author McLean, Reginald
Chatzaroulas, Evangelos
McCutcheon, Luc
Röder, Frank
Yu, Tianhe
He, Zhanpeng
Zentner, K. R.
Julian, Ryan
Terry, J K
Woungang, Isaac
Farsad, Nariman
Castro, Pablo Samuel
author_facet McLean, Reginald
Chatzaroulas, Evangelos
McCutcheon, Luc
Röder, Frank
Yu, Tianhe
He, Zhanpeng
Zentner, K. R.
Julian, Ryan
Terry, J K
Woungang, Isaac
Farsad, Nariman
Castro, Pablo Samuel
contents Meta-World is widely used for evaluating multi-task and meta-reinforcement learning agents, which are challenged to master diverse skills simultaneously. Since its introduction however, there have been numerous undocumented changes which inhibit a fair comparison of algorithms. This work strives to disambiguate these results from the literature, while also leveraging the past versions of Meta-World to provide insights into multi-task and meta-reinforcement learning benchmark design. Through this process we release a new open-source version of Meta-World (https://github.com/Farama-Foundation/Metaworld/) that has full reproducibility of past results, is more technically ergonomic, and gives users more control over the tasks that are included in a task set.
format Preprint
id arxiv_https___arxiv_org_abs_2505_11289
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Meta-World+: An Improved, Standardized, RL Benchmark
McLean, Reginald
Chatzaroulas, Evangelos
McCutcheon, Luc
Röder, Frank
Yu, Tianhe
He, Zhanpeng
Zentner, K. R.
Julian, Ryan
Terry, J K
Woungang, Isaac
Farsad, Nariman
Castro, Pablo Samuel
Artificial Intelligence
Machine Learning
Meta-World is widely used for evaluating multi-task and meta-reinforcement learning agents, which are challenged to master diverse skills simultaneously. Since its introduction however, there have been numerous undocumented changes which inhibit a fair comparison of algorithms. This work strives to disambiguate these results from the literature, while also leveraging the past versions of Meta-World to provide insights into multi-task and meta-reinforcement learning benchmark design. Through this process we release a new open-source version of Meta-World (https://github.com/Farama-Foundation/Metaworld/) that has full reproducibility of past results, is more technically ergonomic, and gives users more control over the tasks that are included in a task set.
title Meta-World+: An Improved, Standardized, RL Benchmark
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2505.11289