aiqbench analysis
This research identifies a critical vulnerability in Vision-Language-Action (VLA) policies where adversarial patches induce persistent state corruption that survives patch removal, significantly degrading long-term task recoverability. The study demonstrates that timely intervention via specialized recovery adapters is essential for restoring performance, as delays drastically reduce the efficacy of these mitigation strategies. These findings highlight a fundamental gap in current security evaluations that often overlook post-attack persistence in embodied AI systems.
Beyond Patch Removal: Persistent Adversarial Effects in Vision-Language-Action Policies
arXiv:2609.19669v1 Announce Type: cross Abstract: Adversarial patches to Vision-Language-Action (VLA) policies can cause both immediate action corruption and persistent state effects that remain after the patch is removed. Existing evaluations largely focus on continuous attacks and do not separate these two effects. We introduce a state-restoration protocol that removes the patch at matched action-chunk boundaries and measures subsequent recoverability under the same remaining step budget. Clean, random-patch, deviation-matched, and fixed-direction controls distinguish adversarial effects from occlusion, action-error magnitude, and directional persistence. We also evaluate a recovery adapter trained on attack-induced states under controlled intervention latency. On OpenVLA-OFT with EDPA attacks, only 36.2% of LIBERO-Long episodes remain recoverable after five chunks, compared with 89.9% and 87.0% for the deviation-matched and fixed-direction controls. Similar persistent effects are observed on autoregressive OpenVLA. The recovery adapter improves recovery from 7.7% to 47.4% at one-chunk latency, but its benefit decreases substantially with delayed intervention. These results show that adversarial effects can persist after patch removal and that timely intervention is critical for recovery.
Abstract:Adversarial patches to Vision-Language-Action (VLA) policies can cause both immediate action corruption and persistent state effects that remain after the patch is removed. Existing evaluations largely focus on continuous attacks and do not separate these two effects. We introduce a state-restoration protocol that removes the patch at matched action-chunk boundaries and measures subsequent recoverability under the same remaining step budget. Clean, random-patch, deviation-matched, and fixed-direction controls distinguish adversarial effects from occlusion, action-error magnitude, and directional persistence. We also evaluate a recovery adapter trained on attack-induced states under controlled intervention latency. On OpenVLA-OFT with EDPA attacks, only 36.2% of LIBERO-Long episodes remain recoverable after five chunks, compared with 89.9% and 87.0% for the deviation-matched and fixed-direction controls. Similar persistent effects are observed on autoregressive OpenVLA. The recovery adapter improves recovery from 7.7% to 47.4% at one-chunk latency, but its benefit decreases substantially with delayed intervention. These results show that adversarial effects can persist after patch removal and that timely intervention is critical for recovery.
Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
arXiv:2609.19669 [cs.CV]
(or arXiv:2609.19669v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2609.19669 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 262 words
Related coverage
[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model
Latent Space · Oct 13, 2026
Lattice-Based Cryptography Explained
Quantum Zeitgeist · Aug 15, 2026
Tiered KV cache for large LLMs on Amazon SageMaker HyperPod with Curvine
AWS · Aug 12, 2026
How might continual learning affect safety and alignment?
LessWrong · Jun 13, 2026