Saved in:
Bibliographic Details
Main Authors: Xiao, Peiwen, Liu, Xiaohui, Pan, Guangming, Long, Wei
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.16535
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913805547601920
author Xiao, Peiwen
Liu, Xiaohui
Pan, Guangming
Long, Wei
author_facet Xiao, Peiwen
Liu, Xiaohui
Pan, Guangming
Long, Wei
contents In this paper, we introduce a novel decentralized surrogate gradient-based algorithm for quantile regression in a feature-distributed setting, where global features are dispersed across multiple machines within a decentralized network. The proposed algorithm, \texttt{DSG-cqr}, utilizes a convolution-type smoothing approach to address the non-smooth nature of the quantile loss function. \texttt{DSG-cqr} is fully decentralized, conjugate-free, easy to implement, and achieves linear convergence up to statistical precision. To ensure privacy, we adopt the Gaussian mechanism to provide $(ε,δ)$-differential privacy. To overcome the exact residual calculation problem, we estimate residuals using auxiliary variables and develop a confidence interval construction method based on Wald statistics. Theoretical properties are established, and the practical utility of the methods is also demonstrated through extensive simulations and a real-world data application.
format Preprint
id arxiv_https___arxiv_org_abs_2504_16535
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Decentralized Quantile Regression for Feature-Distributed Massive Datasets with Privacy Guarantees
Xiao, Peiwen
Liu, Xiaohui
Pan, Guangming
Long, Wei
Computation
In this paper, we introduce a novel decentralized surrogate gradient-based algorithm for quantile regression in a feature-distributed setting, where global features are dispersed across multiple machines within a decentralized network. The proposed algorithm, \texttt{DSG-cqr}, utilizes a convolution-type smoothing approach to address the non-smooth nature of the quantile loss function. \texttt{DSG-cqr} is fully decentralized, conjugate-free, easy to implement, and achieves linear convergence up to statistical precision. To ensure privacy, we adopt the Gaussian mechanism to provide $(ε,δ)$-differential privacy. To overcome the exact residual calculation problem, we estimate residuals using auxiliary variables and develop a confidence interval construction method based on Wald statistics. Theoretical properties are established, and the practical utility of the methods is also demonstrated through extensive simulations and a real-world data application.
title Decentralized Quantile Regression for Feature-Distributed Massive Datasets with Privacy Guarantees
topic Computation
url https://arxiv.org/abs/2504.16535