BriefGPT.xyz
Jul, 2017
利用对抗距离的对偶稳定分布对齐
Stable Distribution Alignment Using the Dual of the Adversarial Distance
HTML
PDF
Ben Usman, Kate Saenko, Brian Kulis
TL;DR
通过转换为可优化的对偶形式，将最大化部分替换为其对偶形式，能够更稳定且单调地优化目标函数，用于解决分布对齐问题。实验表明，对于线性鉴别器，对偶形式在平面上对齐合成点云和数字领域自适应问题中均具有更强的稳定性和单调性。
Abstract
Learning to align distributions by minimizing an
adversarial distance
between them has recently achieved impressive results. However, such models are difficult to optimize with
gradient descent
and they often do
→