Model Protection: Real-Time Privacy-Preserving Inference Service for Model Privacy at the Edge
- 9 November 2021
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 19 (6) , 4270-4284
Abstract
Major cloud service providers with well-equipped infrastructure, experienced machine learning (ML) expertise, and enriched training datasets are building ML-as-a-Service (MLaaS) systems, in which clients can query ML-based prediction services with their data. Instead of moving private data to the cloud, in this work, we design, implement, and evaluate a novel secure ML system to enable MLaaS on edge devices. To protect the proprietary ML models on edge devices from revealing to the clients while maintaining a real-time inference is challenging. Existing privacy-preserving ML techniques can hardly satisfy real-time requirements. In our solution, we employ a secure enclave (\eg SGX) to offer security and provide better efficiency than cryptographic techniques. However, the enclave alone cannot achieve real-time capability due to its limited capacity. We observe that the ML model imposes a severe accuracy degradation when adding noise to a few model weights. Based on this, we design a suite of novel solutions to optimize the performance of secure enclave-based inference service at the edge by enclosing only $1% computation within secure enclaves. Our work can achieve up to a 7.8 increase in efficiency and a 27 reduction in memory usage compared to the state-of-the-art.Keywords
Funding Information
- National Key R&D Program of China (2018YFB0803400)
- Tsinghua University - AsiaInfo Technologies Inc.
- Joint Rsearch Center (20203910074)
- China National Funds for Distinguished Young Scientists (61625205)
- National Natural Science Foundation of China (62132018, 61751211, 61572347, 61520106007)
- Key Research Program of Frontier Sciences, CAS (QYZDY-SSW-JSC002)
- National Science Foundation (CNS-1526638)
This publication has 29 references indexed in Scilit:
- Oblivious Neural Network Predictions via MiniONN Transformations2017
- Machine Learning Models that Remember Too Much2017
- SecureML: A System for Scalable Privacy-Preserving Machine Learning2017
- Membership Inference Attacks Against Machine Learning Models2017
- Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures2015
- Machine Learning Classification over Encrypted Data2015
- Fully homomorphic encryption using ideal lattices2009
- Calibrating Noise to Sensitivity in Private Data Analysis2006
- Historical Development of the Newton–Raphson MethodSIAM Review, 1995
- Gaussian elimination is not optimalNumerische Mathematik, 1969