wuvin
Always take risks!
Toggle navigation
wuvin
主页
实验室的搬砖生活
机器学习
公开的学术内容
公开的其他内容
About Me
归档
标签
友情链接
ZYQN
ihopenot
enigma_aw
hzwer
杨宗翰
ADDA: Adversarial Discriminative Domain Adaptation
2019-09-20 18:54:09
620
0
0
wuvin
# Introduction CVPR2017. They want to do Adversarial Discriminative Domain Adaptation. # Structure As usual, there's Source Domain(the old one) $S$ and Target Domain(the new one) $T$, and a Classifier $C$ on Source Domain, Target Representations $M_t$ and $M_s$, a Discriminator $D$. The loss is straight forward. ![title](https://leanote.com/api/file/getImage?fileId=5d84b317ab64413335001a6a) # Training method ![title](https://leanote.com/api/file/getImage?fileId=5d84b64dab64413335001a92) 1. Train $M_s$ and $C$ on $S$. 2. Keep $M_s$, initialize $M_t$ wth $M_s$ and train $D$ and $M_t$. 3. Tests with $M_t$ and $C$. # Experiments ![title](https://leanote.com/api/file/getImage?fileId=5d84b5abab6441313d0019da) Recall that CYCADA achived 90% on SVHN → MNIST.
上一篇:
吹爆Colab + Kaggle
下一篇:
CYCADA: CYCLE-CONSISTENT ADVERSARIAL DOMAIN ADAPTATION
0
赞
620 人读过
新浪微博
微信
腾讯微博
QQ空间
人人网
提交评论
立即登录
, 发表评论.
没有帐号?
立即注册
0
条评论
More...
文档导航
没有帐号? 立即注册