A classifier trained to be random on clean inputs and accurate on a fixed trigger-shifted input becomes resistant to transferred adversarial attacks, though it remains vulnerable to attackers who replicate the trigger.
Purify unlearnable examples via rate-constrained variational autoencoders,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Towards Model Resistant to Transferable Adversarial Examples via Trigger Activation
A classifier trained to be random on clean inputs and accurate on a fixed trigger-shifted input becomes resistant to transferred adversarial attacks, though it remains vulnerable to attackers who replicate the trigger.