A variational autoencoder (VAE) is one of several generative models that use deep learning to generate new content, detect anomalies and remove noise. Suited for generating synthetic time series data ...