arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.07238stat.MLcs.LG
收藏

神经网络中的双下降曲线:一种使用高斯过程的新视角

Double-descent curves in neural networks: a new perspective using Gaussian processes

University of Oxford(牛津大学)

Ouns El Harzli, Bernardo Cuenca Grau, Guillermo Valle-Pérez, Ard A. Louis

AI总结:

本文利用随机矩阵理论将经验特征协方差矩阵谱刻画为NNGP核谱的宽度依赖扰动,从而为神经网络双下降现象提供了由经验核与NNGP核差异支配的新解释。

摘要:

神经网络中的双下降曲线描述了一种现象:泛化误差最初随着参数增加而下降,在达到一个小于数据点数量的最优参数数量后上升,但随后在过参数化区域再次下降。在本文中,我们使用随机矩阵理论的技术,将经验特征协方差矩阵的谱分布刻画为神经网络高斯过程(NNGP)核谱的一种依赖宽度的扰动,从而在神经网络的背景下建立了NNGP文献与随机矩阵理论文献之间的新联系。我们的解析表达式使我们能够研究相应核与高斯过程回归的泛化行为,并为双下降现象提供了一种新的解释,即它由依赖宽度的经验核与宽度无关的NNGP核之间的差异所支配。

英文摘要:

Double-descent curves in neural networks describe the phenomenon that the generalisation error initially descends with increasing parameters, then grows after reaching an optimal number of parameters which is less than the number of data points, but then descends again in the overparameterized regime. In this paper, we use techniques from random matrix theory to characterize the spectral distribution of the empirical feature covariance matrix as a width-dependent perturbation of the spectrum of the neural network Gaussian process (NNGP) kernel, thus establishing a novel connection between the NNGP literature and the random matrix theory literature in the context of neural networks. Our analytical expression allows us to study the generalisation behavior of the corresponding kernel and GP regression, and provides a new interpretation of the double-descent phenomenon, namely as governed by the discrepancy between the width-dependent empirical kernel and the width-independent NNGP kernel.

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。

↑