Title: "Memory Augmented Generative Models for Few-shot Learning"

Over the last couple of years there has been significant progress in building better generative models that learn to generate complex and high dimensional data like images or sound. Much of this progress has been driven by advances in deep learning and by applying variational inference techniques. One of the most important algorithms in deep learning is stochastic gradient descent and its variants: slowly adapting the models parameters one mini-batch at a time. But we sometimes face situations where we would like to rapidly adapt our models based on only very few training examples. Here I will talk about recent approaches to augment variational autoencoder based models with memory subsystems, how they add few-shot learning capabilities to these models, and how to generate new samples based on very few training examples.