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
AUGMENTED CYCLIC ADVERSARIAL LEARNING FOR LOW RESOURCE DOMAIN ADAPTATION
2019-09-30 14:43:45
753
0
0
wuvin
# Background * Solving low resource domain adaptation. * Assume has few lable for Target Domain. # Structure ![title](https://leanote.com/api/file/getImage?fileId=5d91a51bab6441478f0014c7) * the ultimate goal of our approach is to use the mapped Source → Target samples ($X_{S→T}$ ) to augment the limited data of the target domain ($X_T$ ). * let $M_S$ and $M_T$ be the task specific models trained on domains $P_S(X, Y)$ and $P_T (X, Y)$. ## Relaxed Cycle Consistency . ![title](https://leanote.com/api/file/getImage?fileId=5d91abfdab6441498d001599) ## In supervised case ![title](https://leanote.com/api/file/getImage?fileId=5d91b2d0ab6441498d0015e6) Similarly, there's ![title](https://leanote.com/api/file/getImage?fileId=5d91b33fab6441478f001577) ## In unsupervised case ![title](https://leanote.com/api/file/getImage?fileId=5d91b639ab6441478f0015a8) # Performance $MNIST$ domain is limited to only 10 samples per class, denoted $MNIST-(10)$ ![title](https://leanote.com/api/file/getImage?fileId=5d91b721ab6441478f0015b3) ![title](https://leanote.com/api/file/getImage?fileId=5d91b78cab6441498d001627) ![title](https://leanote.com/api/file/getImage?fileId=5d91d2c7ab6441478f0016cd)
上一篇:
小米运动APP 注册地无法使用NFC问题
下一篇:
Compressed Sensing using Generative Models
0
赞
753 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