# Probabilistic Guarantees for Safe Deep Reinforcement Learning

arXiv: https://arxiv.org/abs/2005.07073 · HTML: https://mlsanity.com/paper/2005.07073/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 May 2020.

## Built on by (most cited first, 15 of 15)

- [How to Certify Machine Learning Based Safety-critical Systems? A Systematic Literature Review](https://mlsanity.com/paper/2107.12045/) (2021)
- [Distillation of RL Policies with Formal Guarantees via Variational Abstraction of Markov Decision Processes (Technical Report)](https://mlsanity.com/paper/2112.09655/) (2021)
- [Certification of Iterative Predictions in Bayesian Neural Networks](https://mlsanity.com/paper/2105.10134/) (2021)
- [Verified Probabilistic Policies for Deep Reinforcement Learning](https://mlsanity.com/paper/2201.03698/) (2022)
- [Unifying Qualitative and Quantitative Safety Verification of DNN-Controlled Systems](https://mlsanity.com/paper/2404.01769/) (2024)
- [Probabilistic Reach-Avoid for Bayesian Neural Networks](https://mlsanity.com/paper/2310.01951/) (2023)
- [Probabilistic Model Checking of Stochastic Reinforcement Learning Policies](https://mlsanity.com/paper/2403.18725/) (2024)
- [Cooperation for Scalable Supervision of Autonomy in Mixed Traffic](https://mlsanity.com/paper/2112.07569/) (2021)
- [Learning on Abstract Domains: A New Approach for Verifiable Guarantee in Reinforcement Learning](https://mlsanity.com/paper/2106.06931/) (2021)
- [Taming Reachability Analysis of DNN-Controlled Systems via Abstraction-Based Training](https://mlsanity.com/paper/2211.11127/) (2022)
- [Wasserstein Auto-encoded MDPs: Formal Verification of Efficiently Distilled RL Policies with Many-sided Guarantees](https://mlsanity.com/paper/2303.12558/) (2023)
- [COOL-MC: Verifying and Explaining RL Policies for Platelet Inventory Management](https://mlsanity.com/paper/2603.02396/) (2026)
- [COOL-MC: Verifying and Explaining RL Policies for Multi-bridge Network Maintenance](https://mlsanity.com/paper/2603.07546/) (2026)
- [Scenario Generation for Risk-Aware Reinforcement Learning with Probably Approximately Safe Guarantees](https://mlsanity.com/paper/2606.04812/) (2026)
- [Probabilistic Model Checking of Autoregressive Neural Sequence Models](https://mlsanity.com/paper/2609.00838/) (2026)

## Abstract (arXiv)

Deep reinforcement learning has been successfully applied to many control tasks, but the application of such agents in safety-critical scenarios has been limited due to safety concerns. Rigorous testing of these controllers is challenging, particularly when they operate in probabilistic environments due to, for example, hardware faults or noisy sensors. We propose MOSAIC, an algorithm for measuring the safety of deep reinforcement learning agents in stochastic settings. Our approach is based on the iterative construction of a formal abstraction of a controller's execution in an environment, and leverages probabilistic model checking of Markov decision processes to produce probabilistic guarantees on safe behaviour over a finite time horizon. It produces bounds on the probability of safe operation of the controller for different initial configurations and identifies regions where correct behaviour can be guaranteed. We implement and evaluate our approach on agents trained for several benchmark control problems.