Enhancing Stateful Detection of Adversarial Attacks with Soft-labels' Temporality and Robust Similarity Approximations
For security practitioners defending black-box ML models, this work improves detection accuracy and robustness against adaptive attacks, though it is an incremental improvement over existing stateful detectors.
This paper enhances stateful detection of adversarial attacks by leveraging temporal correlation of soft labels to reduce false positives and by introducing randomness to prevent evasion of similarity approximations. The proposed method achieves a true positive rate of 1.00 and a false positive rate of at most 0.06 against several attacks.
Stateful Detection (SD) mitigates adversarial attacks by determining whether a sequence of queries contains queries from a black-box adversary. Recent works, such as Blacklight and PIHA utilize query similarity to detect such queries. In this paper, we observe that temporal information, in particular, the temporal correlation of the classification soft labels, is a prominent characteristic of adversarial attacks and can be leveraged to reduce false positive rates. Moreover, we point out a potential vulnerability in SD implementation. Many SD systems identify similar queries according to some implicit, computationally expensive metric. To improve efficiency, these systems often adopt an approximate similarity function as substitute. This discrepancy could be exploited by crafting queries that appear dissimilar under the approximation but are close in the intended metric, thereby evading detection. We refer to this as an ``adversarial attack'' on the approximation function, and demonstrate it through a lightweight attack on Blacklight's similarity function. Based on the above observations, we propose a two-phase approach. The first phase identifies subsequences of queries with high similarity, incorporating randomness to prevent the aforementioned ``adversarial attacks''. The second phase analyzes temporal correlation of the soft-labels to further validate the presence of the adversary's queries. Experimental results show that the framework detects adversarial queries generated by Boundary Attack, HSJA, SimBA, Square Attack with true positive rate (TPR) reaching 1.00, while maintaining a false positive rate (FPR) of at most 0.06. Additionally, the method is robust against OARS which is an adaptive attack.