Analyzing and Improving Representations with the Soft Nearest Neighbor LossEI

摘要

We explore and expand the $textit{Soft Nearest Neighbor Loss}$ to measure the $textit{entanglement}$ of class manifolds in representation space: i.e., how close pairs of points from the same class are relative to pairs of points from different classes. We demonstrate several use cases of the loss. As an analytical tool, it provides insights into the evolution of class similarity structures during learning. Surprisingly, we find that $textit{maximizing}$ the entanglement of representations of different classes in the hidden layers is b...更多
个人信息

 

您的评分 :

2019.

被引用次数0|引用|0
标签
作者
评论