BriefGPT.xyz
Jul, 2022
过拟合的分类:良性,适度和灾难性
Benign, Tempered, or Catastrophic: A Taxonomy of Overfitting
HTML
PDF
Neil Mallinar, James B. Simon, Amirhesam Abedsoltan, Parthe Pandit, Mikhail Belkin...
TL;DR
该论文研究了神经网络等插值方法是否能够在存在噪声的情况下,拟合训练数据而不会表现出灾难性的测试性能,尝试通过“良性过拟合”和“温和过拟合”两个现象进行解释,并首次系统研究了“温和过拟合”的性质及在核(岭)回归中的表现,以及在深度神经网络中的实验结果。
Abstract
The practical success of overparameterized
neural networks
has motivated the recent scientific study of
interpolating methods
, which perfectly fit their training data. Certain
→