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 2209.04036math.MGcs.LGmath.GT
收藏

前馈ReLU神经网络的功能维度

Functional dimension of feedforward ReLU neural networks

Boston College(波士顿学院);Department of Mathematics(数学系);University of Wisconsin(威斯康星大学)

J. Elisenda Grigsby, Kathryn Lindsey, Robert Meyerhoff, Chenxi Wu

AI总结:

本文研究了ReLU神经网络的功能维度,揭示其在参数空间中不均匀性,并探讨了功能维度达到理论最大值的条件。

摘要:

已知全连接前馈神经网络中使用ReLU激活函数可表示的函数参数化家族正是有限多个片段的分段线性函数的集合。较少为人所知的是,对于每一个固定的ReLU神经网络架构,参数空间中存在正维的对称性空间,因此在任何给定参数点附近的函数维度低于参数维度。在本工作中,我们严谨地定义了功能维度的概念,证明其在ReLU神经网络函数参数空间中是不均匀的,并继续一项研究——始于[14]和[5]——探讨何时功能维度达到理论最大值。我们还研究了从参数空间到函数空间的实现映射的商空间和纤维,提供了纤维是不连通的、纤维上功能维度不恒定以及纤维上对称群作用非传递的例子。

英文摘要:

It is well-known that the parameterized family of functions representable by fully-connected feedforward neural networks with ReLU activation function is precisely the class of piecewise linear functions with finitely many pieces. It is less well-known that for every fixed architecture of ReLU neural network, the parameter space admits positive-dimensional spaces of symmetries, and hence the local functional dimension near any given parameter is lower than the parametric dimension. In this work we carefully define the notion of functional dimension, show that it is inhomogeneous across the parameter space of ReLU neural network functions, and continue an investigation - initiated in [14] and [5] - into when the functional dimension achieves its theoretical maximum. We also study the quotient space and fibers of the realization map from parameter space to function space, supplying examples of fibers that are disconnected, fibers upon which functional dimension is non-constant, and fibers upon which the symmetry group acts non-transitively.

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。

↑