Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Enhancing Fine-Tuning Based Backdoor Defense with Sharpness-Aware Minimization

About

Backdoor defense, which aims to detect or mitigate the effect of malicious triggers introduced by attackers, is becoming increasingly critical for machine learning security and integrity. Fine-tuning based on benign data is a natural defense to erase the backdoor effect in a backdoored model. However, recent studies show that, given limited benign data, vanilla fine-tuning has poor defense performance. In this work, we provide a deep study of fine-tuning the backdoored model from the neuron perspective and find that backdoorrelated neurons fail to escape the local minimum in the fine-tuning process. Inspired by observing that the backdoorrelated neurons often have larger norms, we propose FTSAM, a novel backdoor defense paradigm that aims to shrink the norms of backdoor-related neurons by incorporating sharpness-aware minimization with fine-tuning. We demonstrate the effectiveness of our method on several benchmark datasets and network architectures, where it achieves state-of-the-art defense performance. Overall, our work provides a promising avenue for improving the robustness of machine learning models against backdoor attacks.

Mingli Zhu, Shaokui Wei, Li Shen, Yanbo Fan, Baoyuan Wu• 2023

Related benchmarks

TaskDatasetResultRank
Backdoor DefenseCIFAR10 (test)--
322
Backdoor DetectionCIFAR-10
Bd. Rate20
120
Backdoor DefenseCIFAR-10 (test)--
40
Backdoor DetectionGTSRB--
39
Backdoor DefenseCIFAR-10 Blended v1 (test)
Clean Accuracy92.28
34
Backdoor DefenseCIFAR-10 BadNet v1 (test)
Clean Accuracy91.53
20
Backdoor DefenseCIFAR-10
BadNet C-Acc92.63
17
Backdoor DefenseCIFAR-10 LC DenseNet-161 (test)
Clean Accuracy87.79
17
Backdoor DefenseCIFAR-10 SSBA DenseNet-161 (test)
Clean Accuracy86.98
17
Backdoor DefenseCIFAR-10 SSBA v1 (test)
Clean Accuracy91.77
17
Showing 10 of 16 rows

Other info

Follow for update