Enhancing Differentially Private Linear Regression via Public Second-Moment

Zilong Cao, Hai Zhang

Leveraging information from public data has become increasingly crucial in enhancing the utility of differentially private (DP) methods. Traditional DP approaches often require add…