Multimodal Variational Autoencoders for Semi-Supervised Learning: In Defense of Product-of-Experts

Open in new window