PrivaScissors: Enhance the Privacy of Collaborative Inference through the Lens of Mutual Information

2023/05/17 by Lin Duan, Duan, Lin, Jingwei Sun +5 · 4 citations
Computer Science · Social Sciences · #Adversarial Robustness in Machine Learning #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Privacy, Security, and Data Protection #Privacy-Preserving Technologies in Data

paper · pdf · doi:10.48550/arxiv.2306.07973

openalex publication_date 2023/05/17 · openalex created_date 2023/06/17 · openalex updated_date 2026/07/28

Abstract

Edge-cloud collaborative inference empowers resource-limited IoT devices to support deep learning applications without disclosing their raw data to the cloud server, thus preserving privacy. Nevertheless, prior research has shown that collaborative inference still results in the exposure of data and predictions from edge devices. To enhance the privacy of collaborative inference, we introduce a defense strategy called PrivaScissors, which is designed to reduce the mutual information between a model's intermediate outcomes and the device's data and predictions. We evaluate PrivaScissors's performance on several datasets in the context of diverse attacks and offer a theoretical robustness guarantee.

Cited by

• A Survey on Cloud-Edge-Terminal Collaborative Intelligence in AIoT Networks
• VMask: Tunable Label Privacy Protection for Vertical Federated Learning via Layer Masking
• Partitioning for Intrinsic Model Inversion Resistance in Collaborative Inference
• Edge-Cloud Collaborative Computing on Distributed Intelligence and Model Optimization: A Survey

Related

• PRICURE: Privacy-Preserving Collaborative Inference in a Multi-Party Setting
• Can Differential Privacy Practically Protect Collaborative Deep Learning Inference for the Internet of Things?
• How Breakable Is Privacy: Probing and Resisting Model Inversion Attacks in Collaborative Inference
• Privacy-preserving Security Inference Towards Cloud-Edge Collaborative Using Differential Privacy
• PATROL: Privacy-Oriented Pruning for Collaborative Inference Against Model Inversion Attacks