Saved in:
Bibliographic Details
Main Authors: Yang, Guangmingmei, Miller, David J., Kesidis, George
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.08129
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918239103090688
author Yang, Guangmingmei
Miller, David J.
Kesidis, George
author_facet Yang, Guangmingmei
Miller, David J.
Kesidis, George
contents Most post-training backdoor detection methods rely on attacked models exhibiting extreme outlier detection statistics for the target class of an attack, compared to non-target classes. However, these approaches may fail: (1) when some (non-target) classes are easily discriminable from all others, in which case they may naturally achieve extreme detection statistics (e.g., decision confidence); and (2) when the backdoor is subtle, i.e., with its features weak relative to intrinsic class-discriminative features. A key observation is that the backdoor target class has contributions to its detection statistic from both the backdoor trigger and from its intrinsic features, whereas non-target classes only have contributions from their intrinsic features. To achieve more sensitive detectors, we thus propose to suppress intrinsic features while optimizing the detection statistic for a given class. For non-target classes, such suppression will drastically reduce the achievable statistic, whereas for the target class the (significant) contribution from the backdoor trigger remains. In practice, we formulate a constrained optimization problem, leveraging a small set of clean examples from a given class, and optimizing the detection statistic while orthogonalizing with respect to the class's intrinsic features. We dub this plug-and-play approach Class Subspace Orthogonalization (CSO) and assess it against challenging mixed-label and adaptive attacks.
format Preprint
id arxiv_https___arxiv_org_abs_2512_08129
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Improving the Sensitivity of Backdoor Detectors via Class Subspace Orthogonalization
Yang, Guangmingmei
Miller, David J.
Kesidis, George
Machine Learning
Most post-training backdoor detection methods rely on attacked models exhibiting extreme outlier detection statistics for the target class of an attack, compared to non-target classes. However, these approaches may fail: (1) when some (non-target) classes are easily discriminable from all others, in which case they may naturally achieve extreme detection statistics (e.g., decision confidence); and (2) when the backdoor is subtle, i.e., with its features weak relative to intrinsic class-discriminative features. A key observation is that the backdoor target class has contributions to its detection statistic from both the backdoor trigger and from its intrinsic features, whereas non-target classes only have contributions from their intrinsic features. To achieve more sensitive detectors, we thus propose to suppress intrinsic features while optimizing the detection statistic for a given class. For non-target classes, such suppression will drastically reduce the achievable statistic, whereas for the target class the (significant) contribution from the backdoor trigger remains. In practice, we formulate a constrained optimization problem, leveraging a small set of clean examples from a given class, and optimizing the detection statistic while orthogonalizing with respect to the class's intrinsic features. We dub this plug-and-play approach Class Subspace Orthogonalization (CSO) and assess it against challenging mixed-label and adaptive attacks.
title Improving the Sensitivity of Backdoor Detectors via Class Subspace Orthogonalization
topic Machine Learning
url https://arxiv.org/abs/2512.08129