DLP: towards active defense against backdoor attacks with decoupled learning process

TL;DR

This paper introduces a novel active defense method against backdoor attacks in deep learning by decoupling the training process into three stages, effectively mitigating backdoor risks across different datasets and attack types.

Contribution

It proposes a general training pipeline that actively defends against backdoor attacks by decoupling learning into supervised, unlearning, and semi-supervised fine-tuning stages.

Findings

Effective against various backdoor attacks

Works across multiple datasets

Improves model robustness

Abstract

Deep learning models are well known to be susceptible to backdoor attack, where the attacker only needs to provide a tampered dataset on which the triggers are injected. Models trained on the dataset will passively implant the backdoor, and triggers on the input can mislead the models during testing. Our study shows that the model shows different learning behaviors in clean and poisoned subsets during training. Based on this observation, we propose a general training pipeline to defend against backdoor attacks actively. Benign models can be trained from the unreliable dataset by decoupling the learning process into three stages, i.e., supervised learning, active unlearning, and active semi-supervised fine-tuning. The effectiveness of our approach has been shown in numerous experiments across various backdoor attacks and datasets.