English

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

Artificial Intelligence 2023-10-31 v1

Abstract

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.

Keywords

Cite

@article{arxiv.2304.11823,
  title  = {Enhancing Fine-Tuning Based Backdoor Defense with Sharpness-Aware Minimization},
  author = {Mingli Zhu and Shaokui Wei and Li Shen and Yanbo Fan and Baoyuan Wu},
  journal= {arXiv preprint arXiv:2304.11823},
  year   = {2023}
}
R2 v1 2026-06-28T10:15:18.997Z