Pith. sign in

Certified neural net- work watermarks with randomized smoothing

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.CR 1

years

2024 1

verdicts

REJECT 1

representative citing papers

BlockDoor: Blocking Backdoor Based Watermarks in Deep Neural Networks

cs.CR · 2024-12-14 · reject · novelty 4.0

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.

citing papers explorer

Showing 1 of 1 citing paper.

  • BlockDoor: Blocking Backdoor Based Watermarks in Deep Neural Networks cs.CR · 2024-12-14 · reject · none · ref 1

    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.