BriefGPT.xyz
Jul, 2022
自监督学习中我们在最大化什么？
What Do We Maximize in Self-Supervised Learning?
HTML
PDF
Ravid Shwartz-Ziv, Randall Balestriero, Yann LeCun
TL;DR
本文提出通过信息熵理论来推导自监督学习方法VICReg的构建，通过理论与实证验证产生了新的理解以及对其他SSL方法的启示。
Abstract
In this paper, we examine
self-supervised learning
methods, particularly
vicreg
, to provide an information-theoretical understanding of their construction. As a first step, we demonstrate how
→