BriefGPT.xyz
Jun, 2019
线性回归中的良性过拟合
Benign Overfitting in Linear Regression
HTML
PDF
Peter L. Bartlett, Philip M. Long, Gábor Lugosi, Alexander Tsigler
TL;DR
研究表明，过度拟合是深度学习方法的主要问题之一，但当训练中使用最小规范化规则时，线性回归问题中的过度拟合也可以实现高精度预测，需要超参数数量显著超过样本大小。
Abstract
The phenomenon of
benign overfitting
is one of the key mysteries uncovered by
deep learning
methodology: deep neural networks seem to predict well, even with a perfect fit to noisy training data. Motivated by thi
→