Sliced-Wasserstein Autoencoder: An Embarrassingly Simple Generative Model | DeepAI
Tom Sercu on Twitter: "We want to generate new peptides (short proteins) which are active anti-microbials and not toxic. To start, we trained a Wasserstein Auto-Encoder on a large diverse set of
The theory behind Latent Variable Models: formulating a Variational Autoencoder | AI Summer
Sliced-Wasserstein Autoencoder: An Embarrassingly Simple Generative Model
Probabilistic Natural Language Generation with Wasserstein Autoencoders | DeepAI