mlsanityData to 14 Oct 2026

Safe Meta-Reinforcement Learning via Information Space Reachability

arXiv 2609.15915 ↗First posted 14 Sep 2026cs.LG cs.SY eess.SYcited by 0

On the map: Deep reinforcement learning methods → Safe and constrained RL

Abstract (arXiv)

Meta-reinforcement learning (meta-RL) enables agents to adapt to unseen tasks with limited experience. Despite its promise, the application of meta-RL in real-world tasks is hindered by safety requirements, which have been underexplored in prior work. In this paper, we propose a safe meta-RL framework that explicitly accounts for safety during adaptation. Our key insight is to reason about safety in the information space, which captures both the physical state and the agent's belief over the underlying task. Within this space, we introduce a safety value function that measures the probability of the agent avoiding unsafe regions indefinitely. We show that this function satisfies a self-consistency condition and a Bellman equation, which make it learnable via meta-RL. Based on this formulation, we develop a safe meta-RL algorithm that learns the safety value function and leverages it for safety filtering and constrained policy optimization. Experiments on meta-RL benchmarks demonstrate the effectiveness of the proposed method.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Safe and constrained RL, one of 15 directions in Deep reinforcement learning methods. Builds on 12 of its 18 references that are in the map. First posted on arXiv 14 Sep 2026.

Builds on

12 of 18 references.

References · 12

  1. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  2. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
  3. 2019Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learningcited by 1,204
  4. 2019Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variablescited by 389
  5. 2018Learning to Adapt in Dynamic, Real-World Environments Through Meta-Reinforcement Learningcited by 276
  6. 2017Hamilton-Jacobi Reachability: A Brief Overview and Recent Advancescited by 201
  7. 2020Learning to Walk in the Real World with Minimal Human Effortcited by 111
  8. 2023A Tutorial on Meta-Reinforcement Learningcited by 108
  9. 2023State-wise Safe Reinforcement Learning: A Surveycited by 55
  10. 2024A CMDP-within-online framework for Meta-Safe Reinforcement Learningcited by 12
  11. 2023Safe Reinforcement Learning with Dual Robustnesscited by 10
  12. 2023Constrained Meta-Reinforcement Learning for Adaptable Safety Guarantee with Differentiable Convex Programmingcited by 8

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Constrained Meta Reinforcement Learning with Provable Test-Time Safety
  2. 2024Progressive Safeguards for Safe and Model-Agnostic Reinforcement Learning
  3. 2025Provably Optimal Reinforcement Learning under Safety Filtering
  4. 2024Safe Reinforcement Learning in Black-Box Environments via Adaptive Shielding
  5. 2024Safe Reinforcement Learning with Learned Non-Markovian Safety Constraints
  6. 2025Safe Reinforcement Learning using Action Projection: Safeguard the Policy or the Environment?
  7. 2024Meta SAC-Lag: Towards Deployable Safe Reinforcement Learning via MetaGradient-based Hyperparameter Tuning
  8. 2025A Survey of Safe Reinforcement Learning and Constrained MDPs: A Technical Survey on Single-Agent and Multi-Agent Safety