BriefGPT.xyz
Jun, 2021
生成模型作为多视角表示学习的数据源
Generative Models as a Data Source for Multiview Representation Learning
HTML
PDF
Ali Jahanian, Xavier Puig, Yonglong Tian, Phillip Isola
TL;DR
研究了从黑盒生成模型而不是数据直接学习通用视觉表达式的设置,通过该生成器的样本输出训练,比较了几种可应用于该设置的表征学习方法,使用生成器的潜空间来生成相同语义内容的多个“视图”,表明多视图数据可以自然地用于识别正面对和负面对。
Abstract
generative models
are now capable of producing highly realistic images that look nearly indistinguishable from the data on which they are trained. This raises the question: if we have good enough
generative models
→