GAN and VA are for unsupervised learning.
Not to map input to output, but understand structure of inputs.
Autoencoders & RBMS also learn structures, but to improve supervise learning.
Last updated 6 years ago