Hessian Informed Mirror Descent

Hessian Informed Mirror Descent
复制标题

黑森知情镜后裔

DOI:
10.1007/s10915-022-01933-5
复制
发表时间:
2022
影响因子:
2.5
通讯作者:
Yan, Ming
Yan, Ming
中科院分区:
数学2区
文献类型:
--
作者:
Wang, Li;Yan, Ming

文献摘要

参考文献

被引文献

相似文献

受近期论文(L. Ying, Journal of Scientific Computing, 84, 1-14(2020))的启发,我们探讨了镜像下降与变度量方法之间的关系。当镜像坐标系中的度量由一个与目标函数的Hessian接近的凸函数诱导时,该方法既具有镜像下降的鲁棒性,又具有牛顿型方法的超线性收敛性。当应用于一个线性约束最小化问题时,我们证明了在连续和离散情况下的全局和局部收敛性。作为应用,我们计算了具有退化迁移率的Wasserstein梯度流和Cahn-Hillard方程。当使用相对于可变度量的最小化运动方案来制定这些问题时,我们的镜像下降算法为潜在的优化问题提供了快速的收敛速度,同时保持了解决方案的总质量和边界。
Inspired by the recent paper (L. Ying, Journal of Scientific Computing, 84, 1–14 (2020), we explore the relationship between the mirror descent and the variable metric method. When the metric in the mirror decent is induced by a convex function, whose Hessian is close to the Hessian of the objective function, this method enjoys both robustness from the mirror descent and superlinear convergence for Newton type methods. When applied to a linearly constrained minimization problem, we prove the global and local convergence, both in the continuous and discrete settings. As applications, we compute the Wasserstein gradient flows and Cahn-Hillard equation with degenerate mobility. When formulating these problems using a minimizing movement scheme with respect to a variable metric, our mirror descent algorithm offers a fast convergence speed for the underlying optimization problem while maintaining the total mass and bounds of the solution.
Wasserstein 梯度流的增强拉格朗日方法及其应用
DOI: --
发表时间: 2016
期刊:
影响因子: --
作者:
J. Benamou;G. Carlier;M. Laborde
通讯作者: M. Laborde
DOI: --
发表时间: 2020
期刊: E S A I M: Control, Optimisation and Calculus of Variations
影响因子: --
作者:
M. Jacobs;Wonjun Lee;F. L'eger
通讯作者: F. L'eger
用于最小化相互作用自由能的镜像下降算法
DOI: --
发表时间: 2020
影响因子: 2.5
作者:
Lexing Ying
通讯作者: Lexing Ying