mlsanityData to 7 Oct 2026

Sampling-Based Safe Reinforcement Learning

arXiv 2605.19469 ↗First posted 19 May 2026cs.LG cs.AI cs.ROcited by 0

Not on the map. Why?

Abstract (arXiv)

Safe exploration remains a fundamental challenge in reinforcement learning (RL), limiting the deployment of RL agents in the real world. We propose Sampling-Based Safe Reinforcement Learning (SBSRL), a model-based RL algorithm that maintains safety throughout the learning process by enforcing constraints jointly across a finite set of dynamics samples. This formulation approximates an intractable worst-case optimization over uncertain dynamics and enables practical safety guarantees in continuous domains. We further introduce an exploration strategy based on constraining epistemic uncertainty, eliminating the need for explicit exploration bonuses. Under regularity conditions, we derive high-probability guarantees of safety throughout learning and a finite-time sample complexity bound for recovering a near-optimal policy. Empirically, SBSRL achieves safe and efficient exploration both in simulation and in real robotic hardware, and readily extends to practical deep-ensemble implementations that scale to high-dimensional continuous control problems.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 45 of its 69 references that are in the map. First posted on arXiv 19 May 2026.

Builds on

45 of 69 references.

References · 45

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,888
  3. 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,401
  4. 2018DeepMind Control Suitecited by 935
  5. 2017Constrained Policy Optimizationcited by 928
  6. 2019When to Trust Your Model: Model-Based Policy Optimizationcited by 753
  7. 2018Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Modelscited by 748
  8. 2017On Kernelized Multi-armed Banditscited by 348
  9. 2017Safe Model-based Reinforcement Learning with Stability Guaranteescited by 347
  10. 2021Safe Learning in Robotics: From Learning-Based Control to Safe Reinforcement Learningcited by 337
  11. 2024Deep Reinforcement Learning for Robotics: A Survey of Real-World Successescited by 153
  12. 2020Provably Efficient Safe Exploration via Primal-Dual Policy Optimizationcited by 131
33 more
  1. 2018Learning-based Model Predictive Control for Safe Explorationcited by 116
  2. 2020Safe Reinforcement Learning in Constrained Markov Decision Processescited by 109
  3. 2017Cautious Model Predictive Control using Gaussian Process Regressioncited by 91
  4. 2018A predictive safety filter for learning-based control of constrained nonlinear dynamical systemscited by 91
  5. 2020Sample-efficient Cross-Entropy Method for Real-time Planningcited by 90
  6. 2020Information Theoretic Regret Bounds for Online Nonlinear Controlcited by 87
  7. 2020Efficient Model-Based Reinforcement Learning through Optimistic Policy Search and Planningcited by 83
  8. 2020An empirical investigation of the challenges of real-world reinforcement learningcited by 75
  9. 2016Safe Exploration in Finite Markov Decision Processes with Gaussian Processescited by 72
  10. 2022Saute RL: Almost Surely Safe Reinforcement Learning Using State Augmentationcited by 63
  11. 2022Constrained Policy Optimization via Bayesian World Modelscited by 55
  12. 2023SafeDreamer: Safe Reinforcement Learning with World Modelscited by 50
  13. 2016On Explore-Then-Commit Strategiescited by 46
  14. 2021Safe Reinforcement Learning Using Advantage-Based Interventioncited by 45
  15. 2022Near-Optimal Sample Complexity Bounds for Constrained MDPscited by 36
  16. 2021DOPE: Doubly Optimistic and Pessimistic Exploration for Safe Reinforcement Learningcited by 34
  17. 2023Optimistic Active Exploration of Dynamical Systemscited by 32
  18. 2023Safe Exploration in Reinforcement Learning: A Generalized Formulation and Algorithmscited by 27
  19. 2022Convergence and sample complexity of natural policy gradient primal-dual methods for constrained MDPscited by 26
  20. 2024Truly No-Regret Learning in Constrained MDPscited by 23
  21. 2024Transductive Active Learning: Theory and Applicationscited by 22
  22. 2024ActSafe: Active Exploration with Safety Constraints for Reinforcement Learningcited by 22
  23. 2021A Simple and Efficient Sampling-based Algorithm for General Reachability Analysiscited by 12
  24. 2020Prediction with Approximated Gaussian Process Dynamical Modelscited by 8
  25. 2024Safe Guaranteed Exploration for Non-linear Systemscited by 8
  26. 2025SPiDR: A Simple Approach for Zero-Shot Safety in Sim-to-Real Transfercited by 8
  27. 2025SOMBRL: Scalable and Optimistic Model-Based RLcited by 7
  28. 2024Towards safe and tractable Gaussian process-based MPC: Efficient sampling within a sequential quadratic programming frameworkcited by 6
  29. 2026Safe Exploration via Policy Priorscited by 6
  30. 2022Safe Reinforcement Learning via Confidence-Based Filterscited by 5
  31. 2025Finite-Sample-Based Reachability for Safe Control with Gaussian Process Dynamicscited by 5
  32. 2025Sampling from Gaussian Processes: A Tutorial and Applications in Global Sensitivity Analysis and Optimizationcited by 3
  33. 2025Safe and Near-Optimal Control with Online Dynamics Learningcited by 1

Built on by

Not cited yet.