PPML

STAMP: Lightweight TEE-Assisted MPC for Efficient Privacy-Preserving Machine Learning
Our paper introduces STAMP, an end-to-end 3-party MPC protocol for efficient privacy-preserving machine learning inference. USTAMP combines MPC protocol with a lightweight TEE (LTEE) to reduce MPC overhead while avoiding challenges in a traditional TEE. STAMP achieves significantly lower inference overhead than state-of-the-art MPC protocols with either CPU or GPU, under either a WAN or LAN setting.
Pengzhi Huang, Thang Hoang, Yueying Li, Elaine Shi, G. Edward Suh
PDF Cite