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 1802.06266cs.LGcs.NAmath.NA
收藏

浅网络和深网络的训练和泛化误差分析

An analysis of training and generalization errors in shallow and deep networks

Institute of Mathematical Sciences, Claremont Graduate University(克莱蒙特研究生大学数学科学学院);Center for Brains, Minds, and Machines, McGovern Institute for Brain Research, Massachusetts Institute of Technology(麻省理工学院麦戈文脑研究所脑、心智与机器中心)

Hrushikesh Mhaskar, Tomaso Poggio

摘要:

本文的动机是围绕深度网络的一个开放性问题,即尽管有大量的过度参数化,但明显没有过度拟合,这使得训练数据能够完美拟合。在本文中,我们分析了这种现象的情况下,回归问题时,每个单元评估一个周期性的激活函数。我们认为,平方损失的最小期望值是不适当的,以衡量泛化误差的组合函数的近似,以充分利用的组合结构。相反,我们在最大损失的意义上测量泛化误差,有时,作为逐点误差。我们给出了估计究竟有多少参数确保零训练误差以及良好的泛化误差。我们证明了一个正则化问题的解决方案是保证产生一个良好的训练误差以及良好的泛化误差,并估计有多少误差,预计在测试数据。

英文摘要:

This paper is motivated by an open problem around deep networks, namely, the apparent absence of over-fitting despite large over-parametrization which allows perfect fitting of the training data. In this paper, we analyze this phenomenon in the case of regression problems when each unit evaluates a periodic activation function. We argue that the minimal expected value of the square loss is inappropriate to measure the generalization error in approximation of compositional functions in order to take full advantage of the compositional structure. Instead, we measure the generalization error in the sense of maximum loss, and sometimes, as a pointwise error. We give estimates on exactly how many parameters ensure both zero training error as well as a good generalization error. We prove that a solution of a regularization problem is guaranteed to yield a good training error as well as a good generalization error and estimate how much error to expect at which test data.

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。

↑