Pith. sign in

REVIEW 2 cited by

On the Robustness of Safe Reinforcement Learning under Observational Perturbations

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2205.14691 v3 pith:KFEOVFFX submitted 2022-05-29 cs.LG cs.AIcs.RO

On the Robustness of Safe Reinforcement Learning under Observational Perturbations

classification cs.LG cs.AIcs.RO
keywords safeobservationalrewardattackadversarialeffectivelearningmaximizes
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Safe reinforcement learning (RL) trains a policy to maximize the task reward while satisfying safety constraints. While prior works focus on the performance optimality, we find that the optimal solutions of many safe RL problems are not robust and safe against carefully designed observational perturbations. We formally analyze the unique properties of designing effective observational adversarial attackers in the safe RL setting. We show that baseline adversarial attack techniques for standard RL tasks are not always effective for safe RL and propose two new approaches - one maximizes the cost and the other maximizes the reward. One interesting and counter-intuitive finding is that the maximum reward attack is strong, as it can both induce unsafe behaviors and make the attack stealthy by maintaining the reward. We further propose a robust training framework for safe RL and evaluate it via comprehensive experiments. This paper provides a pioneer work to investigate the safety and robustness of RL under observational attacks for future safe RL studies. Code is available at: \url{https://github.com/liuzuxin/safe-rl-robustness}

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Why Does Agentic Safety Fail to Generalize Across Tasks?

    cs.LG 2026-05 conditional novelty 6.0

    Agentic safety fails to generalize across tasks because the task-to-safe-controller mapping has a higher Lipschitz constant than the task-to-controller mapping alone, as proven in linear-quadratic control and demonstr...

  2. Safe-RULE: Safe Reinforcement UnLEarning

    cs.LG 2026-06 unverdicted novelty 5.0

    Safe-RULE introduces a reinforcement unlearning defense for offline safe RL that counters data poisoning by removing malicious data influence while preserving task performance and safety.