Lecture part A
Lecture part B
In this section, we covered the implementation of Generative models viz. Undercomplete Autoencoder, Denoising Autoencoders, Variational Autoencoders and Generative Adversarial Networks. We analyze these models from the perspective of the framework of Energy Based Models (EBM). In doing so, we realize that these generative models can be considered as extensions of EBMs and differ from each other with subtle architectural adjustments.