Membership Inference Attacks Against Incremental Learning in IoT Devices
- 20 December 2024
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 24 (5) , 4006-4021
Abstract
Internet of Things (IoT) devices are frequently deployed in highly dynamic environments and need to continuously learn new classes from data streams. Incremental Learning (IL) has gained popularity in IoT as it enables devices to learn new classes efficiently without retraining model entirely. IL involves fine-tuning the model using two sources of data: a small amount of representative samples from the original training dataset and samples from the new classes. However, both data sources are vulnerable to Membership Inference Attack (MIA). Fortunately, the existing MIAs result in poor performance against IL, because they ignore features such as the similarity between old and new models at the old classification layer. This paper presents the first MIA against IL, capable of determining not only whether a sample was used for training/fine-tuning but also distinguishing whether it belongs to the representative dataset or the new classes (unique in IL). Extensive experiments validate the effectiveness of our attack across four real-world datasets. Our attack achieves an average attack success rate of 74.03% in the white-box setting (model structure and parameters are known) and 70.08% in the black-box setting. Importantly, our attack is not sensitive to the IL hyper-parameters (e.g., distillation temperature), confirming its accurate, robust, and practical.Keywords
Funding Information
- National Natural Science Foundation of China (62422208, 62202276, 62232010, 62072278, 62202274, 62302265, 62302274, U23A20332, U23A20273)
- Shandong Science Fund for Excellent Young Scholars (2022HWYQ-038, 2024HWYQ-020, 2024HWYQ-021)
- Natural Science Foundation of Shandong Province (ZR2023QF172, ZR2022LZH010, ZR2023QF113)
- Shandong Science Fund (2023TSGC0105)
This publication has 51 references indexed in Scilit:
- A novel IoT intrusion detection framework using Decisive Red Fox optimization and descriptive back propagated radial basis function modelsScientific Reports, 2024
- Sliding principal component and dynamic reward reinforcement learning based IIoT attack detectionScientific Reports, 2023
- Novel Approach of Class Incremental Learning on Internet of Things (IoT) Framework2021
- End-to-End Incremental Learning2018
- Overcoming catastrophic forgetting in neural networksProceedings of the National Academy of Sciences of the United States of America, 2017
- Incremental learning of bidirectional principal components for face recognitionPattern Recognition, 2010
- Direct importance estimation for covariate shift adaptationAnnals of the Institute of Statistical Mathematics, 2008
- Differential Privacy2006
- Self-refreshing memory in artificial neural networks: learning temporal sequences without catastrophic forgettingConnection Science, 2004
- Stolen memories2002