About
Using variational autoencoders to build a recommender system for fontsWe use a neural network variational autoencoder architecture in order to encode font styles into a latent representation. With the help of this architecture we enable our machine to learn similiarities between font styles. This, in turn, is used to build our recommender system.
Note from the editor
We use a neural network variational autoencoder architecture in order to encode font styles into a latent representation. With the help of this architecture we enable our machine to learn similiarities between font styles. This, in turn, is used to build our recommender system.
Editors
