arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2408.11267cs.LG
收藏

反转杠杆分数梯度:一种高效的近似牛顿方法

Inverting the Leverage Score Gradient: An Efficient Approximate Newton Method

Fuzhou University(福州大学);Adobe Research(奥多比研究院);Wuhan University(武汉大学);Boston University(波士顿大学)

Chenyang Li, Zhao Song, Zhaoxing Xu, Junze Yin

AI总结:

本文提出一种基于子采样杠杆分数分布的近似牛顿迭代算法,用于高效反转杠杆分数梯度,恢复模型参数,显著降低计算复杂度,兼顾数据隐私与对抗安全。

摘要:

杠杆分数在统计学和机器学习中已成为不可或缺的工具,助力回归分析、随机矩阵计算以及各种其他任务。本文深入探讨逆问题,旨在根据杠杆分数梯度恢复内在模型参数。这一努力不仅丰富了对使用杠杆分数技术训练的模型的理论理解,而且对数据隐私和对抗性安全具有重大意义。我们特别审视了杠杆分数梯度(记为$g(x)$)的反转。我们引入了一种创新的迭代算法,用于近似求解正则化最小二乘问题,表述为$\min_{x \in \mathbb{R}^d} 0.5 \\|g(x) - c\\|_2^2 + 0.5\\|\mathrm{diag}(w)Ax\\|_2^2$。我们的算法在每次迭代中采用子采样杠杆分数分布来计算近似Hessian矩阵,在标准假设下,显著降低了时间复杂度。鉴于总共需要$T = \log(\\| x_0 - x^* \\|_2/ ε)$次迭代,每次迭代的成本被优化至$O( (\mathrm{nnz}(A) + d^ω ) \cdot \mathrm{poly}(\log(n/δ))$的数量级,其中$\mathrm{nnz}(A)$表示$A$的非零条目数。

英文摘要:

Leverage scores have become essential in statistics and machine learning, aiding regression analysis, randomized matrix computations, and various other tasks. This paper delves into the inverse problem, aiming to recover the intrinsic model parameters given the leverage scores gradient. This endeavor not only enriches the theoretical understanding of models trained with leverage score techniques but also has substantial implications for data privacy and adversarial security. We specifically scrutinize the inversion of the leverage score gradient, denoted as $g(x)$. An innovative iterative algorithm is introduced for the approximate resolution of the regularized least squares problem stated as $\min_{x \in \mathbb{R}^d} 0.5 \|g(x) - c\|_2^2 + 0.5\|\mathrm{diag}(w)Ax\|_2^2$. Our algorithm employs subsampled leverage score distributions to compute an approximate Hessian in each iteration, under standard assumptions, considerably mitigating the time complexity. Given that a total of $T = \log(\| x_0 - x^* \|_2/ ε)$ iterations are required, the cost per iteration is optimized to the order of $O( (\mathrm{nnz}(A) + d^ω ) \cdot \mathrm{poly}(\log(n/δ))$, where $\mathrm{nnz}(A)$ denotes the number of non-zero entries of $A$.

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。

↑