arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.04313stat.MLcs.LGstat.ME
收藏

在每用户多样本条件下实现更优的本地差分隐私稀疏估计

Better Locally Private Sparse Estimation Given Multiple Samples Per User

School of Statistics, Renmin University of China(中国人民大学统计学院);Center for Applied Statistics, Renmin University of China(中国人民大学应用统计中心)

Yuheng Ma, Ke Jia, Hanfang Yang

AI总结:

本文针对本地差分隐私稀疏线性回归提出用户级框架,通过先筛选候选变量再在低维空间估计,消除了误差对维度 d 的线性依赖,将上界从 O(s*d/(nmε²)) 改进为 O(s*²/(nmε²)),并在实验中验证了其优越性。

摘要:

先前的研究对带有 $s^*$-稀疏性假设的条目级本地差分隐私线性回归给出了令人失望的结果,其中 $nm$ 个样本的极小极大速率为 $\mathcal{O}(s^{*}d / nm\varepsilon^2)$。这对于维度 $d$ 极大的高维数据而言可能极具挑战性。在本工作中,我们研究了用户级本地差分隐私稀疏线性回归。我们证明,在 $n$ 个用户每人贡献 $m$ 个样本的情况下,可以消除对维度 $d$ 的线性依赖,从而得到 $\mathcal{O}(s^{*2} / nm\varepsilon^2)$ 的误差上界。我们提出了一个框架,该框架首先筛选候选变量,然后在缩窄后的低维空间中进行估计,并且可扩展到具有紧致误差界的一般稀疏估计问题。在合成数据集和真实数据集上的实验证明了所提方法的优越性。理论和实证结果均表明,在样本总数相同的情况下,当每个用户可提供多个样本时,进行本地差分隐私稀疏估计效果更好。

英文摘要:

Previous studies yielded discouraging results for item-level locally differentially private linear regression with $s^*$-sparsity assumption, where the minimax rate for $nm$ samples is $\mathcal{O}(s^{*}d / nm\varepsilon^2)$. This can be challenging for high-dimensional data, where the dimension $d$ is extremely large. In this work, we investigate user-level locally differentially private sparse linear regression. We show that with $n$ users each contributing $m$ samples, the linear dependency of dimension $d$ can be eliminated, yielding an error upper bound of $\mathcal{O}(s^{*2} / nm\varepsilon^2)$. We propose a framework that first selects candidate variables and then conducts estimation in the narrowed low-dimensional space, which is extendable to general sparse estimation problems with tight error bounds. Experiments on both synthetic and real datasets demonstrate the superiority of the proposed methods. Both the theoretical and empirical results suggest that, with the same number of samples, locally private sparse estimation is better conducted when multiple samples per user are available.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑