mlsanityData to 7 Oct 2026

Provably Safe Reinforcement Learning for Stochastic Reach-Avoid Problems with Entropy Regularization

arXiv 2601.08646 ↗First posted 13 Jan 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

We consider the problem of learning the optimal policy for Markov decision processes with safety constraints. We formulate the problem in a reach-avoid setup. Our goal is to design online reinforcement learning algorithms that ensure safety constraints with arbitrarily high probability during the learning phase. To this end, we first propose an algorithm based on the optimism in the face of uncertainty (OFU) principle. Based on the first algorithm, we propose our main algorithm, which utilizes entropy regularization. We investigate the finite-sample analysis of both algorithms and derive their regret bounds. We demonstrate that the inclusion of entropy regularization improves the regret and drastically controls the episode-to-episode variability that is inherent in OFU-based safe RL algorithms.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 15 of its 36 references that are in the map. First posted on arXiv 13 Jan 2026.

Builds on

15 of 36 references.

References · 15

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,635
  3. 2015Trust Region Policy Optimizationcited by 2,503
  4. 2017Constrained Policy Optimizationcited by 928
  5. 2017Safe Reinforcement Learning via Shieldingcited by 427
  6. 2017Unifying PAC and Regret: Uniform PAC Bounds for Episodic Reinforcement Learningcited by 175
  7. 2017A unified view of entropy-regularized Markov decision processescited by 174
  8. 2020Exploration-Exploitation in Constrained MDPscited by 146
  9. 2019Convergent Policy Optimization for Safe Reinforcement Learningcited by 55
  10. 2019Tight Regret Bounds for Model-Based Reinforcement Learning with Greedy Policiescited by 38
  11. 2021DOPE: Doubly Optimistic and Pessimistic Exploration for Safe Reinforcement Learningcited by 34
  12. 2023Provably Efficient Model-Free Algorithms for Non-stationary CMDPscited by 21
3 more
  1. 2020Learning in Markov Decision Processes under Constraintscited by 20
  2. 2019Smoothing Policies and Safe Policy Gradientscited by 18
  3. 2024Safe Reinforcement Learning for Constrained Markov Decision Processes with Stochastic Stopping Timecited by 5

Built on by

Not cited yet.