ADS-C: Antidistillation Sampling for Classification
Machine LearningCryptography and Security
Summary
The authors studied a way to stop attackers from copying a classifier by tweaking the output probabilities in a way that doesn’t change the classifier’s predictions. Their adapted method, ADS-C, ensures the classifier's accuracy remains the same while making it much harder for attackers to successfully mimic it. Unlike previous methods that hurt the classifier’s performance, ADS-C blocks knowledge stealing without any loss in accuracy. This method works by carefully controlling how much the probabilities can be changed without affecting the top prediction, effectively reversing the advantage for attackers.
Authors
Khawaja Abaid Ullah, Mohammad Javad Khojasteh
Abstract
Knowledge distillation enables an adversary to replicate a proprietary classifier by querying its prediction interface and training a surrogate on the returned probability vectors. Antidistillation sampling, proposed for large language models, counters this threat with an input-dependent, gradient-directed perturbation of the served distribution; its transfer to classification has not been studied. Adapting the defense to classification, we show its behavior is governed by the distribution of the teacher's per-input confidence margins. Because well-trained classifiers are severely overconfident, the direct transfer exhibits an inert window: below a closed-form-predictable threshold, it affects neither attacker nor defender; beyond it, the defense undergoes a phase transition and degrades the teacher faster than the attacker's student. Temperature softening rescales the transition in closed form, and every temperature configuration lies on the same unfavorable trade-off curve. Our method, ADS-C, composes the perturbation under a closed-form, per-input margin budget that provably preserves every served top-1 prediction, so the defended teacher's accuracy equals the undefended teacher's identically. Under this guarantee the distilled student still loses 17.4 percentage points on CIFAR-100, 29.6 on CIFAR-10, and 13.3 on Tiny-ImageNet; matching this degradation with the unmodified defense costs 27.5, 32.9, and 22.2 points of teacher accuracy. Because served labels are unchanged, a hard-label attacker gains nothing, while the defended soft output trains a student up to 29.7 points below that floor: the incentive to distill served probabilities is not merely removed but reversed. To our knowledge, ADS-C is the first antidistillation defense for classification whose utility cost is exactly zero.