Research Article
Outsourced Privacy-Preserving SVM Classifier Model over Encrypted Data in IoT
@INPROCEEDINGS{10.1007/978-3-031-30623-5_6, author={Chen Wang and Yan Dong and Jiarun Li and Chen Chen and Jian Xu}, title={Outsourced Privacy-Preserving SVM Classifier Model over Encrypted Data in IoT}, proceedings={Security and Privacy in New Computing Environments. 5th EAI International Conference, SPNCE 2022, Xi’an, China, December 30-31, 2022, Proceedings}, proceedings_a={SPNCE}, year={2023}, month={4}, keywords={Multi-key FHE cloud-edge privacy-preserving Internet of Things (IoT) SVM}, doi={10.1007/978-3-031-30623-5_6} }- Chen Wang
Yan Dong
Jiarun Li
Chen Chen
Jian Xu
Year: 2023
Outsourced Privacy-Preserving SVM Classifier Model over Encrypted Data in IoT
SPNCE
Springer
DOI: 10.1007/978-3-031-30623-5_6
Abstract
The Internet of Things (IoT) enables the development of cloud computing by combining machine learning (ML) and big data technologies. Frameworks supporting ML typically process and classification manufacturing data services via cloud-based technologies. In order to achieve secure and efficient data collection and application, we design, implement and evaluate a new system employing a SVM classifier model over encrypted data (SVMCM-ED) based on multi-key Fully Homomorphic Encryption (multi-key FHE) in IoT. We first propose a new scheme that uses the cloud server and edge node to jointly implement SVM classification over the ciphertext to ensure the security of data and classification model. Our system significantly transfers cloud processing to the edge, and safely uses the model obtained from the data analysis to implement SVM classification among multiple users. We further design secure protocols based on multi-key FHE, which supports multiple users and satisfies the semi-honest security model where cloud computing is outsourced. Our protocol requires no interaction from the data analysis and users during online secure classification. Performance evaluation demonstrates that our new system can securely train a logistic regression model of multiple users. Performance evaluation demonstrates that our new system can securely implement the SVM classification model of multiple users in IoT, as well as share communication and computation overhead of the cloud.
