MEDIC: Model Backdoor Removal by Importance Driven Cloning

Qiuling Xu, Guanhong Tao, Jean Honorio, Yingqi Liu, Shengwei An, Guangyu Shen, Siyuan Cheng, Xiangyu Zhang

Keywords: Backdoor Removal, Cloning

TL;DR: We propose importance driven cloning to remove backdoor in machine learning models.

Abstract:

We develop a novel method to remove injected backdoors in Deep Learning models. It works by cloning the benign behaviors of a trojaned model to a new model of the same structure. It trains the clone model from scratch on a very small subset of samples and aims to minimize a cloning loss that denotes the differences between the activations of important neurons across the two models. The set of important neurons varies for each input, depending on their magnitude of activations and their impact on the classification result.
Our experiments show that our technique can effectively remove nine different types of backdoors with minor benign accuracy degradation, outperforming the state-of-the-art backdoor removal techniques that are based on fine-tuning, knowledge distillation, and neuron pruning.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Social Aspects of Machine Learning (eg, AI safety, fairness, privacy, interpretability, human-AI interaction, ethics)