Classification Under Local Differential Privacy with Model Reversal and Model Averaging
Year: 2026
Type: article
Abstract: Local differential privacy (LDP) has become a central topic in data privacy research, offering strong privacy guarantees by perturbing user data at the source and removing the need for a trusted curator. However, the noise introduced by LDP often significantly reduces data utility. To address this issue, we reinterpret private learning under LDP as a transfer learning problem, where the noisy data serve as the source domain and the unobserved clean data as the target. We propose novel techniques specifically designed for LDP to improve classification performance without compromising privacy: (1) a noised binary feedback-based evaluation mechanism for estimating dataset utility; (2) model reversal, which salvages underperforming classifiers by inverting their decision boundaries; and (3) model averaging, which assigns weights to multiple reversed classifiers based on their estimated utility. We provide theoretical excess risk bounds under LDP and demonstrate how our methods reduce this risk. Empirical results on both simulated and real-world datasets show substantial improvements in classification accuracy.
Source: PubMed
Authors: Qin, Caihong, Bai, Yang
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Awards: Measurement error correction approaches to wearable device-based measures of physical activity and self-reported measures of dietary intake in obesity and type 2 diabetes research, Novel Bayesian assessments of device-based physical activity and self-reported dietary intake in joint models of all-cause mortality and type 2 diabetes in a cohort of biracial older US adults
OpenAlex ID: W7127980098
