BlockDoor wraps a watermarked network with detectors for adversarial, out-of-distribution, and randomly labeled triggers, but its own results show the OOD branch fails on unseen trigger sources and the random-label branch degrades clean accuracy.
Stealing machine learning models via prediction apis
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
BlockDoor: Blocking Backdoor Based Watermarks in Deep Neural Networks
BlockDoor wraps a watermarked network with detectors for adversarial, out-of-distribution, and randomly labeled triggers, but its own results show the OOD branch fails on unseen trigger sources and the random-label branch degrades clean accuracy.