Customizing Trusted AI Accelerators for Efficient Privacy-Preserving Machine Learning
This addresses privacy concerns for users deploying machine learning on AI accelerators, but it is incremental as it builds on existing trusted hardware and cryptographic methods.
The paper tackles the problem of protecting privacy on AI accelerators by proposing a solution that uses a trusted CPU and a customized trusted AI accelerator with cryptographic primitives, achieving efficient privacy-preserving machine learning with small design cost and moderate performance overhead.
The use of trusted hardware has become a promising solution to enable privacy-preserving machine learning. In particular, users can upload their private data and models to a hardware-enforced trusted execution environment (e.g. an enclave in Intel SGX-enabled CPUs) and run machine learning tasks in it with confidentiality and integrity guaranteed. To improve performance, AI accelerators have been widely employed for modern machine learning tasks. However, how to protect privacy on an AI accelerator remains an open question. To address this question, we propose a solution for efficient privacy-preserving machine learning based on an unmodified trusted CPU and a customized trusted AI accelerator. We carefully leverage cryptographic primitives to establish trust and protect the channel between the CPU and the accelerator. As a case study, we demonstrate our solution based on the open-source versatile tensor accelerator. The result of evaluation shows that the proposed solution provides efficient privacy-preserving machine learning at a small design cost and moderate performance overhead.