On Convex Optimization with Semi-Sensitive Features

Badih Ghazi, Pritish Kamath, Ravi Kumar, Pasin Manurangsi, Raghu Meka +1 more
6/27/2024
cs.LGcs.CRcs.DS

No confident code implementation found for this paper.

Abstract

We study the differentially private (DP) empirical risk minimization (ERM) problem under the semi-sensitive DP setting where only some features are sensitive. This generalizes the Label DP setting where only the label is sensitive. We give improved upper and lower bounds on the excess risk for DP-ERM. In particular, we show that the error only scales polylogarithmically in terms of the sensitive domain size, improving upon previous results that scale polynomially in the sensitive domain size (Ghazi et al., 2021).

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Adapting to Linear Separable Subsets with Large-Margin in Differentially Private Learning

84% match2025

The Cost of Shuffling in Private Gradient Based Optimization

83% match1 repo2025

Differentially Private Optimization with Sparse Gradients

83% match2024

Differentially Private Bilevel Optimization: Efficient Algorithms with Near-Optimal Rates

82% match2025

On User-Level Private Convex Optimization

82% match2023

Faster Algorithms for User-Level Private Stochastic Convex Optimization

82% match2026

Cite this paper

@article{ghazi2024convex,
  title  = {On Convex Optimization with Semi-Sensitive Features},
  author = {Badih Ghazi and Pritish Kamath and Ravi Kumar and Pasin Manurangsi and Raghu Meka and Chiyuan Zhang},
  year   = {2024},
  eprint = {2406.19040},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2406.19040v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2406.19040.svg)](https://codeofpaper.com/paper/2406.19040)

Discussion