Query-Free Adversarial Transfer via Undertrained Surrogates

TL;DR

This paper proposes a novel adversarial attack method that uses undertrained surrogate models to improve transferability in black-box settings, outperforming existing techniques across multiple datasets and architectures.

Contribution

Introducing a simple yet effective approach of undertraining surrogate models to enhance adversarial transferability in black-box attacks.

Findings

Outperforms state-of-the-art methods significantly

Reduces local loss maxima hindering transferability

Effective across diverse datasets and model architectures

Abstract

Deep neural networks are vulnerable to adversarial examples -- minor perturbations added to a model's input which cause the model to output an incorrect prediction. We introduce a new method for improving the efficacy of adversarial attacks in a black-box setting by undertraining the surrogate model which the attacks are generated on. Using two datasets and five model architectures, we show that this method transfers well across architectures and outperforms state-of-the-art methods by a wide margin. We interpret the effectiveness of our approach as a function of reduced surrogate model loss function curvature and increased universal gradient characteristics, and show that our approach reduces the presence of local loss maxima which hinder transferability. Our results suggest that finding strong single surrogate models is a highly effective and simple method for generating transferable adversarial attacks, and that this method represents a valuable route for future study in this field.