Pith. sign in

REVIEW 5 cited by

Adversarial Policies: Attacking Deep Reinforcement Learning

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

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 1905.10615 v3 pith:HAV22UYR submitted 2019-05-25 cs.LG cs.AIcs.CRstat.ML

Adam Gleave , Michael Dennis , Cody Wild , Neel Kant , Sergey Levine , Stuart Russell This is my paper
classification cs.LGcs.AIcs.CRstat.ML
keywords adversarialpoliciesobservationsagentdeeplearningpolicyreinforcement
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Deep reinforcement learning (RL) policies are known to be vulnerable to adversarial perturbations to their observations, similar to adversarial examples for classifiers. However, an attacker is not usually able to directly modify another agent's observations. This might lead one to wonder: is it possible to attack an RL agent simply by choosing an adversarial policy acting in a multi-agent environment so as to create natural observations that are adversarial? We demonstrate the existence of adversarial policies in zero-sum games between simulated humanoid robots with proprioceptive observations, against state-of-the-art victims trained via self-play to be robust to opponents. The adversarial policies reliably win against the victims but generate seemingly random and uncoordinated behavior. We find that these policies are more successful in high-dimensional environments, and induce substantially different activations in the victim policy network than when the victim plays against a normal opponent. Videos are available at https://adversarialpolicies.github.io/.

Discussion (0). Sign in to comment.

Forward citations

Cited by 5 Pith papers

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

  1. Robust Critics: Defending LLMs Against Multi-Turn Attacks

    cs.AI 2026-05 conditional novelty 6.0 of 10

    Critic-weighted sampling over inferred user intents improves multi-turn LLM defense success while preserving helpfulness, with an expected-Q improvement guarantee and transfer to frontier models.

  2. SoK: The Pitfalls of Deep Reinforcement Learning for Cybersecurity

    cs.LG 2026-02 accept novelty 6.0 of 10

    Across 66 DRL-for-cybersecurity papers, the authors identify 11 recurring methodological pitfalls—averaging 5.8 per paper—and demonstrate their impact in four environments.

  3. Virtual Agent Economies

    cs.AI 2025-09 conditional novelty 5.0 of 10

    Proposes a two-axis framework (emergent versus intentional, permeable versus impermeable) for the coming AI agent economy and argues for proactively designing steerable agent markets.

  4. RIVAL: Reinforcement Learning with Iterative and Adversarial Optimization for Machine Translation

    cs.CL 2025-06 conditional novelty 5.0 of 10

    RIVAL iteratively re-trains a reward model adversarially against the current translator and adds a BLEU-predicting head, improving in-domain WMT and subtitle translation over SFT baselines.

  5. Adversarial Attacks on Reinforcement Learning-based Medical Questionnaire Systems: Input-level Perturbation Strategies and Medical Constraint Validation

    cs.CR 2025-08 reject novelty 4.0 of 10

    On the NHIS mortality task, standard white-box attacks flip the final prediction of the AdaptiveFS RL questionnaire model in 33.1% (FGSM) to 64.7% (AutoAttack) of tested correctly classified cases.

Pith tools