# Classification Under Local Differential Privacy with Model Reversal and Model Averaging

arXiv: https://arxiv.org/abs/2602.05797 · HTML: https://mlsanity.com/paper/2602.05797/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 76 references that are in the map.
- First posted on arXiv 5 Feb 2026.

## Builds on (most cited first, 27 of 27)

- [RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response](https://mlsanity.com/paper/1407.6981/) (2014)
- [Collecting Telemetry Data Privately](https://mlsanity.com/paper/1712.01524/) (2017)
- [LDP-Fed: Federated Learning with Local Differential Privacy](https://mlsanity.com/paper/2006.03637/) (2020)
- [Minimax Optimal Procedures for Locally Private Estimation](https://mlsanity.com/paper/1604.02390/) (2016)
- [Transfer Learning for High-dimensional Linear Regression: Prediction, Estimation, and Minimax Optimality](https://mlsanity.com/paper/2006.10593/) (2020)
- [Transfer Learning under High-dimensional Generalized Linear Models](https://mlsanity.com/paper/2105.14328/) (2021)
- [Transfer Learning for Nonparametric Classification: Minimax Rate and Adaptive Classifier](https://mlsanity.com/paper/1906.02903/) (2019)
- [LDP-FL: Practical Private Aggregation in Federated Learning with Local Differential Privacy](https://mlsanity.com/paper/2007.15789/) (2020)
- [Collecting and Analyzing Multidimensional Data with Local Differential Privacy](https://mlsanity.com/paper/1907.00782/) (2019)
- [Local Differential Privacy and Its Applications: A Comprehensive Survey](https://mlsanity.com/paper/2008.03686/) (2020)
- [Encode, Shuffle, Analyze Privacy Revisited: Formalizations and Empirical Evaluation](https://mlsanity.com/paper/2001.03618/) (2020)
- [FedSel: Federated SGD under Local Differential Privacy with Top-k Dimension Selection](https://mlsanity.com/paper/2003.10637/) (2020)
- [Collecting and Analyzing Data from Smart Device Users with Local Differential Privacy](https://mlsanity.com/paper/1606.05053/) (2016)
- [Classification under local differential privacy](https://mlsanity.com/paper/1912.04629/) (2019)
- [Optimal Locally Private Nonparametric Classification with Public Data](https://mlsanity.com/paper/2311.11369/) (2023)
- [Random Sampling Plus Fake Data: Multidimensional Frequency Estimates With Local Differential Privacy](https://mlsanity.com/paper/2109.07269/) (2021)
- [Optimal Bayes Classifiers for Functional Data and Density Ratios](https://mlsanity.com/paper/1605.03707/) (2016)
- [Profile-Based Privacy for Locally Private Computations](https://mlsanity.com/paper/1903.09084/) (2019)
- [Improving the utility of locally differentially private protocols for longitudinal and multidimensional frequency estimates](https://mlsanity.com/paper/2111.04636/) (2021)
- [Practical Solutions in Fully Homomorphic Encryption -- A Survey Analyzing Existing Acceleration Methods](https://mlsanity.com/paper/2303.10877/) (2023)
- [Geometry of Sensitivity: Twice Sampling and Hybrid Clipping in Differential Privacy with Optimal Gaussian Noise and Application to Deep Learning](https://mlsanity.com/paper/2309.02672/) (2023)
- [On Excess Risk Convergence Rates of Neural Network Classifiers](https://mlsanity.com/paper/2309.15075/) (2023)
- [Classification of functional fragments by regularized linear classifiers with domain selection](https://mlsanity.com/paper/1708.08257/) (2017)
- [Danger of using fully homomorphic encryption: A look at Microsoft SEAL](https://mlsanity.com/paper/1906.07127/) (2019)
- [A reproducing kernel Hilbert space framework for functional data classification](https://mlsanity.com/paper/2103.04504/) (2021)
- [Lemmas of Differential Privacy](https://mlsanity.com/paper/2211.11189/) (2022)
- [A Survey of Local Differential Privacy and Its Variants](https://mlsanity.com/paper/2309.00861/) (2023)

## Abstract (arXiv)

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.