mlsanityData to 7 Oct 2026

Nightmare Dreamer: Dreaming About Unsafe States And Planning Ahead

arXiv 2601.04686 ↗First posted 8 Jan 2026cs.LG cs.ROcited by 0

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 14 of its 24 references that are in the map. First posted on arXiv 8 Jan 2026.

Builds on

14 of 24 references.

References · 14

  1. 2022Generative Adversarial Networkscited by 5,603
  2. 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 1,916
  3. 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 1,598
  4. 2018Learning Latent Dynamics for Planning from Pixelscited by 1,221
  5. 2015Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)cited by 1,061
  6. 2020Mastering Atari with Discrete World Modelscited by 990
  7. 2017Constrained Policy Optimizationcited by 928
  8. 2020Fast Global Convergence of Natural Policy Gradient Methods with Entropy Regularizationcited by 170
  9. 2023Safety-Gymnasium: A Unified Safe Reinforcement Learning Benchmarkcited by 135
  10. 2020CRPO: A New Approach for Safe Reinforcement Learning with Convergence Guaranteecited by 128
  11. 2019Benchmarking Batch Deep Reinforcement Learning Algorithmscited by 108
  12. 2022Constrained Policy Optimization via Bayesian World Modelscited by 55
2 more
  1. 2022Safe Reinforcement Learning From Pixels Using a Stochastic Latent Representationcited by 11
  2. 2023DexCatch: Learning to Catch Arbitrary Objects with Dexterous Handscited by 8

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2023SafeDreamer: Safe Reinforcement Learning with World Models
  2. 2025SafeVLA: Towards Safety Alignment of Vision-Language-Action Model via Constrained Learning
  3. 2022DayDreamer: World Models for Physical Robot Learning
  4. 2026SafeDojo: Safe Reinforcement Learning for VLA via Interactive World Model
  5. 2025Safe Planning and Policy Optimization via World Model Learning
  6. 2026SafeVLA-Bench: A Benchmark for the Success-Safety Gap in Vision-Language-Action Models
  7. 2024FOSP: Fine-tuning Offline Safe Policy through World Models
  8. 2020Dreaming: Model-based Reinforcement Learning by Latent Imagination without Reconstruction