A Theory to Instruct Differentially-Private Learning via Clipping Bias Reduction

A Theory to Instruct Differentially-Private Learning via Clipping Bias Reduction

Explore a 15-minute IEEE conference talk that delves into the theory of instructing differentially-private learning through clipping bias reduction. Learn about the research conducted by Hanshen Xiao from MIT, Zihang Xiang and Di Wang from KAUST, and Srinivas Devadas from MIT as they present their findings on this important topic in privacy-preserving machine learning.