arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2006.10016stat.MLcs.LG
收藏

用于凸损失函数的 Nyström 方法

The Nyström method for convex loss functions

EPFL(洛桑联邦理工学院);Swiss Finance Institute(瑞士金融研究所);Università di Genova(热那亚大学);CREST(CREST(法国国家统计与经济研究所下属研究中心));ENSAE - Institut Polytechnique de Paris(巴黎综合理工学院下属法国国立统计与经济管理学校);MIT(麻省理工学院);Istituto Italiano di Tecnologia(意大利技术研究院)

Andrea Della Vecchia, Ernesto De Vito, Jaouad Mourtada, Lorenzo Rosasco

AI总结:

本文将 Nyström 随机子空间经验风险最小化的统计—计算权衡分析扩展到一般凸 Lipschitz 损失,证明多种情形下可兼得计算收益与学习性能。

摘要:

我们研究经典经验风险最小化的一种扩展,其中假设空间由给定 Hilbert 空间内的一个随机子空间构成。具体而言,我们考察 Nyström 方法,在该方法中,子空间由数据的一个随机子集定义。这一方法将核方法中使用的 Nyström 近似恢复为一种特例。使用随机子空间自然会带来计算优势,但关键问题在于它是否会损害学习精度。近来,针对平方损失以及 logistic 损失等自协调损失,统计与计算之间的权衡已得到研究。本文将这些分析扩展到一般的凸 Lipschitz 损失,这类损失可能不具备光滑性,例如支持向量机中使用的 hinge 损失。我们的主要结果表明,存在多种情形,可以在不牺牲学习性能的情况下获得计算收益。当专门针对光滑损失函数时,我们的分析恢复了此前的大多数结果。此外,该分析还使我们能够考虑分类问题,并将代理风险界转化为分类误差界。事实上,这提供了机会,可用于比较 Nyström 近似与 hinge 损失或平方损失等不同损失函数结合时的效果。

英文摘要:

We investigate an extension of classical empirical risk minimization, where the hypothesis space consists of a random subspace within a given Hilbert space. Specifically, we examine the Nyström method where the subspaces are defined by a random subset of the data. This approach recovers Nyström approximations used in kernel methods as a specific case. Using random subspaces naturally leads to computational advantages, but a key question is whether it compromises the learning accuracy. Recently, the tradeoffs between statistics and computation have been explored for the square loss and self-concordant losses, such as the logistic loss. In this paper, we extend these analyses to general convex Lipschitz losses, which may lack smoothness, such as the hinge loss used in support vector machines. Our main results show the existence of various scenarios where computational gains can be achieved without sacrificing learning performance. When specialized to smooth loss functions, our analysis recovers most previous results. Moreover, it allows to consider classification problems and translate the surrogate risk bounds into classification error bounds. Indeed, this gives the opportunity to compare the effect of Nyström approximations when combined with different loss functions such as the hinge or the square loss.

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。

↑