On Convex Optimization with Semi-Sensitive Features

2024-06-27 · Badih Ghazi, Pritish Kamath, Ravi Kumar, Pasin Manurangsi, Raghu Meka, Chiyuan Zhang

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).

📄 PDF Abstract BibTeX arXiv:2406.19040

Code (0)

등록된 구현이 없습니다.

Private Learning with Public Feature Conditioning

2026-06-17 · Shuli Jiang, Walid Krichene, Nicolas Mayoraz arxiv

We study differentially private (DP) regression in settings where each data sample includes public, non-sensitive features -- common in applications such as recommendation and advertising systems. While such label-DP or …

Geometric Algebra Planes: Convex Implicit Neural Volumes

2024-11-20 · Irmak Sivgin, Sara Fridovich-Keil, Gordon Wetzstein, Mert Pilanci

Volume parameterizations abound in recent literature, from the classic voxel grid to the implicit neural representation and everything in between. While implicit representations have shown impressive capacity and better …

Decoder Video Segmentation Video Semantic Segmentation

Forging The Graphs: A Low Rank and Positive Semidefinite Graph Learning Approach

2012-12-01 · NeurIPS 2012 12 · Dijun Luo, Heng Huang, Feiping Nie, Chris H. Ding

In many graph-based machine learning and data mining approaches, the quality of the graph is critical. However, in real-world applications, especially in semi-supervised learning and unsupervised learning, the evaluation…

BIG-bench Machine Learning Clustering Graph Learning

Randomized Sketches of Convex Programs with Sharp Guarantees

2014-04-29 · Mert Pilanci, Martin J. Wainwright

Random projection (RP) is a classical technique for reducing storage and computational costs. We analyze RP-based approximations of convex programs, in which the original optimization problem is approximated by the solut…

compressed sensing Dimensionality Reduction

Deep Semi-Random Features for Nonlinear Function Approximation

2017-02-28 · Kenji Kawaguchi, Bo Xie, Vikas Verma, Le Song

We propose semi-random features for nonlinear function approximation. The flexibility of semi-random feature lies between the fully adjustable units in deep learning and the random features used in kernel methods. For on…