李宏毅GAN学习笔记（02）

GAN Lecture 2

---

Conditional Generation by GAN

Algorithm

In each traing iteration:

• Sample m positive examples { ( c 1 , x 1 ) , ( c 2 , x 2 ) , … , ( c m , x m ) } \{(c^1, x^1), (c^2, x^2), \dots, (c^m, x^m)\} {(c1,x1),(c2,x2),…,(cm,xm)} from database
• Sample m noise samples { z 1 , z 2 , … , z m } \{z^1, z^2, \dots, z^m\} {z1,z2,…,zm} from a distribution
• Obtaining generated data { x ~ 1 , x ~ 2 , … , x ~ m } \{\tilde{x}^1, \tilde{x}^2, \dots, \tilde{x}^m\} {x~1,x~2,…,x~m}, ${\tilde{x}}^i=G(c^i,z^i)$
• Sample m objects { x ^ 1 , x ^ 2 , … , x ^ m } \{\hat{x}^1, \hat{x}^2, \dots, \hat{x}^m\} {x^1,x^2,…,x^m} from database
• Update discriminator parameters ${\theta }_d$ to maximize
  • $\tilde{V}=\frac{1}{m}{\sum }_{i=1}^{m}logD(c^i,x^i)+\frac{1}{m}{\sum }_{i=1}^{m}log(1-D(c^i,{\tilde{x}}^i))+{\frac{1}{m}}_{i=1}^{m}log(1-D(c^i,{\hat{x}}^i))$
  • ${\theta }_d\leftarrow {\theta }_d+\eta ▽\tilde{V}({\theta }_d)$

Learning D

• Sample m noise samples { z 1 , z 2 , … , z m } \{z^1,z^2,\dots,z^m\} {z1,z2,…,zm} from a distribution
• Sample m conditions { c 1 , c 2 , … , c m } \{c^1,c^2,\dots,c^m\} {c1,c2,…,cm} from a database
• Update generator parameters ${\theta }_g$ to maximize
  • $\tilde{V}=\frac{1}{m}{\sum }_{i=1}^{m}log(D(G(c^i,z^i)))$, ${\theta }_g\leftarrow \eta ▽\tilde{V}({\theta }_g)$

Learning G

倾向推荐第二种网络架构
参考文献：StackGAN

参考文献：Patch GAN

参考例子：Github