mlsanityData to 14 Oct 2026

Probabilistic Verification of Recurrent Neural Networks for Single and Multi-Agent Reinforcement Learning

arXiv 2605.14758 ↗First posted 14 May 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

History-dependent policies induced by recurrent neural networks (RNNs) rely on latent hidden state dynamics, making verification in partially observable reinforcement learning (RL) challenging. Existing RNN verification tools typically rely on restrictive modeling assumptions or coarse over-approximations of the hidden state space, which can lead to overly conservative or inconclusive results. We propose RNN Probabilistic Verification (textttRNN-ProVe), a probabilistic framework that textitestimates the likelihood of undesired behaviors in RNN-based policies. textttRNN-ProVe uses policy-driven sampling to approximate the set of hidden states that are feasible under a trained policy, and derives statistical error bounds to produce bounded-error, high-confidence estimates of behavioral violations. Experiments on partially observable single-agent and cooperative multi-agent tasks show that textttRNN-ProVe yields more quantitative, feasibility-aware probabilistic guarantees than existing tools, while scaling to recurrent and multi-agent settings.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 27 of its 41 references that are in the map. First posted on arXiv 14 May 2026.

Builds on

27 of 41 references.

References · 27

  1. 2014On the Properties of Neural Machine Translation: Encoder-Decoder Approachescited by 1,304
  2. 2017Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networkscited by 744
  3. 2015Deep Recurrent Q-Learning for Partially Observable MDPscited by 465
  4. 2021Beta-CROWN: Efficient Bound Propagation with Per-neuron Split Constraints for Complete and Incomplete Neural Network Robustness Verificationcited by 228
  5. 2019Algorithms for Verifying Deep Neural Networkscited by 221
  6. 2020Benchmarking Multi-Agent Deep Reinforcement Learning Algorithms in Cooperative Taskscited by 212
  7. 2020Fast and Complete: Enabling Complete Neural Network Verification with Rapid and Massively Parallel Incomplete Verifierscited by 187
  8. 2019Branch and Bound for Piecewise Linear Neural Network Verificationcited by 138
  9. 2024An Introduction to Centralized Training for Decentralized Execution in Cooperative Multi-Agent Reinforcement Learningcited by 50
  10. 2021Reachability Analysis of Neural Feedback Loopscited by 41
  11. 2019POPQORN: Quantifying Robustness of Recurrent Neural Networkscited by 37
  12. 2023Provably Bounding Neural Network Preimagescited by 29
15 more
  1. 2020Verifying Recurrent Neural Networks using Invariant Inferencecited by 26
  2. 2023The #DNN-Verification Problem: Counting Unsafe Inputs for Deep Neural Networkscited by 19
  3. 2023Enumerating Safe Regions in Deep Neural Networks with Provable Probabilistic Guaranteescited by 19
  4. 2020Macro-Action-Based Deep Multi-Agent Reinforcement Learningcited by 18
  5. 2023Provable Preimage Under-Approximation for Neural Networks (Full Version)cited by 14
  6. 2020DiffRNN: Differential Verification of Recurrent Neural Networkscited by 12
  7. 2023Safe Deep Reinforcement Learning by Verifying Task-Level Propertiescited by 9
  8. 2025RL2Grid: Benchmarking Reinforcement Learning in Power Grid Operationscited by 9
  9. 2023Online Safety Property Collection and Refinement for Safe Deep Reinforcement Learning in Mapless Navigationcited by 7
  10. 2024ModelVerification.jl: a Comprehensive Toolbox for Formally Verifying Deep Neural Networkscited by 7
  11. 2024Improving Policy Optimization via varepsilon-Retraincited by 5
  12. 2024On Stateful Value Factorization in Multi-Agent Reinforcement Learningcited by 5
  13. 2025Advancing Neural Network Verification through Hierarchical Safety Abstract Interpretationcited by 4
  14. 2025Designing Control Barrier Function via Probabilistic Enumeration for Safe Reinforcement Learning Navigationcited by 3
  15. 2025On the Probabilistic Learnability of Compact Neural Network Preimage Boundscited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026A Survey on the Verification of Reinforcement Learning Policies
  2. 2024Probabilistic Model Checking of Stochastic Reinforcement Learning Policies
  3. 2026Belief-State RWKV for Reinforcement Learning under Partial Observability
  4. 2026Why Linear Recurrent Memory Works in Partially Observable Reinforcement Learning
  5. 2018Learning Finite State Representations of Recurrent Policy Networks
  6. 2026Investigating Action Encodings in Recurrent Neural Networks in Reinforcement Learning
  7. 2025Overcoming Non-stationary Dynamics with Evidential Proximal Policy Optimization
  8. 2026Does Learning Recurrent Networks Matter in Partially Observable RL?