Devling into Adversarial Transferability on Image Classification: Review, Benchmark, and Evaluation
This work provides a standardized evaluation framework for researchers and practitioners working on adversarial transferability, which is crucial for assessing the security of machine learning models.
This paper addresses the lack of a standardized evaluation framework for transfer-based adversarial attacks in image classification. It proposes a comprehensive benchmark framework and categorizes existing attacks into six groups to enable fair comparisons.
Adversarial transferability refers to the capacity of adversarial examples generated on the surrogate model to deceive alternate, unexposed victim models. This property eliminates the need for direct access to the victim model during an attack, thereby raising considerable security concerns in practical applications and attracting substantial research attention recently. In this work, we discern a lack of a standardized framework and criteria for evaluating transfer-based attacks, leading to potentially biased assessments of existing approaches. To rectify this gap, we have conducted an exhaustive review of hundreds of related works, organizing various transfer-based attacks into six distinct categories. Subsequently, we propose a comprehensive framework designed to serve as a benchmark for evaluating these attacks. In addition, we delineate common strategies that enhance adversarial transferability and highlight prevalent issues that could lead to unfair comparisons. Finally, we provide a brief review of transfer-based attacks beyond image classification.