An Improved Method for Making CNN Immune to Backdoor Attack by Activating Clustering

Yuang Zhou, Yichen Lei, Limin Yu, Xianyao Li, Dingding Chen, Tongpo Zhang*

*Corresponding author for this work

Research output: Chapter in Book or Report/Conference proceedingConference Proceedingpeer-review

Abstract

When a neural network is trained with a data set from an untrusted source, an attacker can insert poisoned data with a backdoor trigger into the data set to make the neural network make wrong decisions. By using Activation Clustering over convolutional neural networks, we propose an improved method for defensing backdoor attacks in the process of data collection and preparation. Experimental results show that this method can reliably protect neural networks from the interference of malicious data during training. The essence of this method is making a neural network to learn the feature of the trigger and classify the toxic data into a separate class. The structure of the existing model is also optimized to make the size of the model lightweight.

Original languageEnglish
Title of host publicationProceedings - 2022 6th International Symposium on Computer Science and Intelligent Control, ISCSIC 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1-6
Number of pages6
ISBN (Electronic)9781665454889
DOIs
Publication statusPublished - 11 Nov 2022
Event6th International Symposium on Computer Science and Intelligent Control, ISCSIC 2022 - Virtual, Online, China
Duration: 11 Nov 202213 Nov 2022

Publication series

NameProceedings - 2022 6th International Symposium on Computer Science and Intelligent Control, ISCSIC 2022

Conference

Conference6th International Symposium on Computer Science and Intelligent Control, ISCSIC 2022
Country/TerritoryChina
CityVirtual, Online
Period11/11/2213/11/22

Keywords

  • Activation Clustering
  • Backdoor Attack
  • Machine learning
  • Neural Network
  • Poison data

Fingerprint

Dive into the research topics of 'An Improved Method for Making CNN Immune to Backdoor Attack by Activating Clustering'. Together they form a unique fingerprint.

Cite this