Safety-Gymnasium: A Unified Safe Reinforcement Learning Benchmark

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ji, Jiaming, Zhang, Borong, Zhou, Jiayi, Pan, Xuehai, Huang, Weidong, Sun, Ruiyang, Geng, Yiran, Zhong, Yifan, Dai, Juntao, Yang, Yaodong
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910633585278976
author Ji, Jiaming
Zhang, Borong
Zhou, Jiayi
Pan, Xuehai
Huang, Weidong
Sun, Ruiyang
Geng, Yiran
Zhong, Yifan
Dai, Juntao
Yang, Yaodong
author_facet Ji, Jiaming
Zhang, Borong
Zhou, Jiayi
Pan, Xuehai
Huang, Weidong
Sun, Ruiyang
Geng, Yiran
Zhong, Yifan
Dai, Juntao
Yang, Yaodong
contents Artificial intelligence (AI) systems possess significant potential to drive societal progress. However, their deployment often faces obstacles due to substantial safety concerns. Safe reinforcement learning (SafeRL) emerges as a solution to optimize policies while simultaneously adhering to multiple constraints, thereby addressing the challenge of integrating reinforcement learning in safety-critical scenarios. In this paper, we present an environment suite called Safety-Gymnasium, which encompasses safety-critical tasks in both single and multi-agent scenarios, accepting vector and vision-only input. Additionally, we offer a library of algorithms named Safe Policy Optimization (SafePO), comprising 16 state-of-the-art SafeRL algorithms. This comprehensive library can serve as a validation tool for the research community. By introducing this benchmark, we aim to facilitate the evaluation and comparison of safety performance, thus fostering the development of reinforcement learning for safer, more reliable, and responsible real-world applications. The website of this project can be accessed at https://sites.google.com/view/safety-gymnasium.
format Preprint
id arxiv_https___arxiv_org_abs_2310_12567
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Safety-Gymnasium: A Unified Safe Reinforcement Learning Benchmark
Ji, Jiaming
Zhang, Borong
Zhou, Jiayi
Pan, Xuehai
Huang, Weidong
Sun, Ruiyang
Geng, Yiran
Zhong, Yifan
Dai, Juntao
Yang, Yaodong
Artificial Intelligence
Machine Learning
Artificial intelligence (AI) systems possess significant potential to drive societal progress. However, their deployment often faces obstacles due to substantial safety concerns. Safe reinforcement learning (SafeRL) emerges as a solution to optimize policies while simultaneously adhering to multiple constraints, thereby addressing the challenge of integrating reinforcement learning in safety-critical scenarios. In this paper, we present an environment suite called Safety-Gymnasium, which encompasses safety-critical tasks in both single and multi-agent scenarios, accepting vector and vision-only input. Additionally, we offer a library of algorithms named Safe Policy Optimization (SafePO), comprising 16 state-of-the-art SafeRL algorithms. This comprehensive library can serve as a validation tool for the research community. By introducing this benchmark, we aim to facilitate the evaluation and comparison of safety performance, thus fostering the development of reinforcement learning for safer, more reliable, and responsible real-world applications. The website of this project can be accessed at https://sites.google.com/view/safety-gymnasium.
title Safety-Gymnasium: A Unified Safe Reinforcement Learning Benchmark
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2310.12567