From Individual Computation to Allied Optimization: Remodeling Privacy-Preserving Neural Inference with Function Input Tuning
Qiao Zhang, Tao Xiang, Chunsheng Xin, Hongyi Wu
IEEE Symposium on Security and Privacy 2024 · Day 3 · Continental Ballroom 6
The proliferation of Machine Learning as a Service (MLaaS) has democratized access to powerful AI capabilities, yet it introduces significant privacy challenges, particularly when handling sensitive data like medical records. This talk, presented by Qiao Zhang from Tongji University, delves into a novel approach to enhance the efficiency of **privacy-preserving neural inference** within MLaaS environments. Co-authored with Tao Xiang, Chunsheng Xin, and Hongyi Wu, the research addresses the inherent computational bottlenecks in existing privacy-preserving machine learning (PPML) solutions.
AI review
This research presents a technically robust and practically significant advancement in privacy-preserving neural inference. By introducing the Function Input Tuning (FIT) framework, it cleverly re-engineers the interaction between linear and non-linear operations, yielding over 10x speedup and reducing communication overhead, which is critical for real-world MLaaS adoption.