Adversarial Auto-Encoders

目录

  • Another Approach: q(z)->p(z)
    • Intuitively comprehend KL(p|q)
  • Minimize KL Divergence
    • How to compute KL between q(z) and p(z)
  • Distribution of hidden code

Adversarial Auto-Encoders_第1张图片

  • Give more details after GAN

Adversarial Auto-Encoders_第2张图片

Another Approach: q(z)->p(z)

  • Explicitly enforce

Adversarial Auto-Encoders_第3张图片

Intuitively comprehend KL(p|q)

\(KL \in{[0,+\infty]}\)

  • KL = 0 --> \(p\approx{q}\)

Adversarial Auto-Encoders_第4张图片

Minimize KL Divergence

  • Evidence Lower BOund

\[ KL(q_\theta(z|x_i)||p(z)) \]

Adversarial Auto-Encoders_第5张图片

How to compute KL between q(z) and p(z)

Adversarial Auto-Encoders_第6张图片

我们所希望的两点:

  • 重建变量
  • 隐藏变量逼近p的分布

你可能感兴趣的:(Adversarial Auto-Encoders)