I create notebooks related to this topics for my own learing about AutoEncoder, Variational AutoEncoder and their application in topic modeling.
-
Notebook 1: "autoencoder_with_single_dense_layer" Understand the structure of an AutoEncoder (Encoder and Decoder) with a simple implementation with a single dense layer
-
Notebook 2: "Variational_Auto_Encoder" More advance topic with "Variational" in which the final layer of the encoders present the distribution of parameters of a distribution Z, we will use the network to learn p(Z|X). Finally we will sample from this distribution and recontruct in the decoder (implement the reparameterization tricks)
-
Notebook 3: An application topic with "Topic Modeling with VAE" with the application of the probabistics LDA model.
Check out these 3 notebooks if you want to understand more about Variational Auto Encoding (I have watch tons of document and youtube videos to implement these 3 but forgot to refer them, thank you all the publish materials and creators, I would try to add the reference later)
Email me if you want to contact me : [email protected]