Saved in:
Bibliographic Details
Main Authors: Oseni, Oluwatosin, Wang, Shengjie, Zhu, Jun, Corah, Micah
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.04686
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909984477937664
author Oseni, Oluwatosin
Wang, Shengjie
Zhu, Jun
Corah, Micah
author_facet Oseni, Oluwatosin
Wang, Shengjie
Zhu, Jun
Corah, Micah
contents Reinforcement Learning (RL) has shown remarkable success in real-world applications, particularly in robotics control. However, RL adoption remains limited due to insufficient safety guarantees. We introduce Nightmare Dreamer, a model-based Safe RL algorithm that addresses safety concerns by leveraging a learned world model to predict potential safety violations and plan actions accordingly. Nightmare Dreamer achieves nearly zero safety violations while maximizing rewards. Nightmare Dreamer outperforms model-free baselines on Safety Gymnasium tasks using only image observations, achieving nearly a 20x improvement in efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2601_04686
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Nightmare Dreamer: Dreaming About Unsafe States And Planning Ahead
Oseni, Oluwatosin
Wang, Shengjie
Zhu, Jun
Corah, Micah
Machine Learning
Robotics
Reinforcement Learning (RL) has shown remarkable success in real-world applications, particularly in robotics control. However, RL adoption remains limited due to insufficient safety guarantees. We introduce Nightmare Dreamer, a model-based Safe RL algorithm that addresses safety concerns by leveraging a learned world model to predict potential safety violations and plan actions accordingly. Nightmare Dreamer achieves nearly zero safety violations while maximizing rewards. Nightmare Dreamer outperforms model-free baselines on Safety Gymnasium tasks using only image observations, achieving nearly a 20x improvement in efficiency.
title Nightmare Dreamer: Dreaming About Unsafe States And Planning Ahead
topic Machine Learning
Robotics
url https://arxiv.org/abs/2601.04686