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 2005.04507math.OCcs.LG
收藏

利用占用时间自适应扰动有效地逃离鞍点

Escaping Saddle Points Efficiently with Occupation-Time-Adapted Perturbations

UC Berkeley(加州大学伯克利分校);Flatiron Institute(熨斗研究院);Columbia University(哥伦比亚大学)

Xin Guo, Jiequn Han, Mahan Tajrobehkar, Wenpin Tang

摘要:

基于统计物理学中自排斥随机游动的超扩散性,本文提出了一种新的优化算法扰动机制。在这种机制中,扰动通过占据时间的概念适应于状态的历史。将该机制引入扰动梯度下降(PGD)和扰动加速梯度下降(PAGD)框架,提出了两种新的算法:适应占用时间的扰动梯度下降(PGDOT)及其加速算法(PAGDOT). PGDOT和PAGDOT收敛到二阶稳定点至少一样快PGD和PAGD,分别,因此,他们保证避免陷入非退化鞍点。理论分析证实了实证研究中,新的算法始终逃脱鞍点和超越不仅是他们的同行,PGD和PAGD,但也包括其他流行的替代品,随机梯度下降,亚当,AMSGrad,和RMSProp。

英文摘要:

Motivated by the super-diffusivity of self-repelling random walk, which has roots in statistical physics, this paper develops a new perturbation mechanism for optimization algorithms. In this mechanism, perturbations are adapted to the history of states via the notion of occupation time. After integrating this mechanism into the framework of perturbed gradient descent (PGD) and perturbed accelerated gradient descent (PAGD), two new algorithms are proposed: perturbed gradient descent adapted to occupation time (PGDOT) and its accelerated version (PAGDOT). PGDOT and PAGDOT are shown to converge to second-order stationary points at least as fast as PGD and PAGD, respectively, and thus they are guaranteed to avoid getting stuck at non-degenerate saddle points. The theoretical analysis is corroborated by empirical studies in which the new algorithms consistently escape saddle points and outperform not only their counterparts, PGD and PAGD, but also other popular alternatives including stochastic gradient descent, Adam, AMSGrad, and RMSProp.

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。

↑