PulseAugur
EN
LIVE 10:34:53
tool · [1 source] ·

New algorithm improves differentially private linear regression accuracy

Researchers have introduced a new algorithm called Iterative Hessian Mixing (IHM) for differentially private ordinary least squares (DP-OLS). This method builds upon existing Gaussian sketching techniques and offers improved accuracy guarantees compared to prior approaches like Adaptive Sufficient Statistics Perturbation (AdaSSP). IHM demonstrates superior performance in empirical evaluations across various datasets, outperforming existing baselines. AI

Summary written by gemini-2.5-flash-lite from 1 sources. How we write summaries →

IMPACT This research advances privacy-preserving techniques in machine learning, potentially enabling more secure data analysis for sensitive datasets.

RANK_REASON The cluster contains a new academic paper detailing a novel algorithm for a specific machine learning task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv stat.ML →

COVERAGE [1]

  1. arXiv stat.ML TIER_1 · Omri Lev, Moshe Shenfeld, Vishwak Srinivasan, Katrina Ligett, Ashia C. Wilson ·

    Near-Optimal Private Linear Regression via Iterative Hessian Mixing

    arXiv:2601.07545v2 Announce Type: replace-cross Abstract: We study differentially private ordinary least squares (DP-OLS) with bounded data $(X,Y)$ via sketching-based mechanisms. While Gaussian sketching approaches have been explored for DP-OLS \citep{sheffet2017differentially},…