Metraipolis › archive › dc_cd6ee5c649b4

TMT: Runtime Backdoor Detection for Vision-Language-Action Policies on Unseen Tasks — arXiv:2610.09462

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.09462
License (as archived)
CC0-1.0
Archived
2026-10-08T12:49:03.169Z
SHA-256 of the archived text
02e8bbd1346dd62ce8279fa3875bfbebf2b439d6a76fee188e66cb4ff46c2aeb
Stable URL of this page
https://metraipolis.com/documento/dc_cd6ee5c649b4

How to cite

TMT: Runtime Backdoor Detection for Vision-Language-Action Policies on Unseen Tasks — arXiv:2610.09462 — arXiv (en), https://arxiv.org/abs/2610.09462 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cd6ee5c649b4 (Metraipolis archive dc_cd6ee5c649b4, sha256 02e8bbd1346dd62ce8279fa3875bfbebf2b439d6a76fee188e66cb4ff46c2aeb)
BibTeX
@misc{dccd6ee5c649b4,
  title = {TMT: Runtime Backdoor Detection for Vision-Language-Action Policies on Unseen Tasks — arXiv:2610.09462},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.09462},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cd6ee5c649b4; sha256 02e8bbd1346dd62ce8279fa3875bfbebf2b439d6a76fee188e66cb4ff46c2aeb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09462. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Backdoored vision-language-action (VLA) policies can preserve benign task performance while producing malicious actions when a trigger appears. Detecting such activation is difficult because malicious behavior can comprise individually plausible actions, while unfamiliar tasks introduce legitimate changes in observations and behavior. We introduce TMT, a runtime backdoor detector based on Token Manifold and latent Transition modeling. Trained on benign rollouts, its two branches assess input-token structure and prediction errors in adjacent-layer latent dynamics. A suspicious rollout identified by the token manifold branch, once confirmed through latent deviations, guides transition selection for subsequent monitoring. We further explore policy purification through self-distillation: a frozen copy of the backdoored policy provides benign-input actions to supervise a student on paired benign and triggered observations, without requiring a separate clean reference policy. For evaluation, we adapt traditional backdoor detectors and repurpose anomaly and failure detection methods as VLA backdoor detectors. In a post-hoc comparison with ten baselines, TMT achieves state-of-the-art backdoor detection performance on unseen tasks across three VLA backdoor attacks. Our project page is available at this https URL .

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.RO, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2610.09462 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T12:18:30.530Z dalla pagina https://arxiv.org/list/cs.RO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).