BriefGPT.xyz
Jul, 2024
深度神经网络如何摆脱维数灾难：组合性和对称学习
How DNNs break the Curse of Dimensionality: Compositionality and Symmetry Learning
HTML
PDF
Arthur Jacot, Seok Hoan Choi, Yuxiao Wen
TL;DR
深度神经网络可以有效地学习有界F1-范数的任何函数组合，从而打破维度诅咒，得到具有创新性的广义限制。
Abstract
We show that
deep neural networks
(DNNs) can efficiently learn any
composition of functions
with bounded $F_{1}$-norm, which allows DNNs to break the
→