Generative Model

Flow-Based Model

I believe CS294 teaches these.

Flow-based model learns invertible transformation between data and latent variables: $x=f(z),p(x)=p(z)\left|\det \left(\frac{\partial f^{-1}}{\partial x}\right)\right|$

Resources

• https://lilianweng.github.io/posts/2018-10-13-flow-models/

Flow-based model vs. normalizing flow?

Normalizing Flow is the method.

Flow-based model is a model that uses this method.

Examples:

• RealNVP
• Glow
• NICE

See Normalizing Flow for actual notes.