UCL x DeepMind DL Lecture - Variational Inference in Generative Models

UCL Centre for AI is partnering with DEEPMIND TO DELIVER A DEEP LEARNING LECTURE SERIES. This lecture will cover two types of powerful modern generative models with latent variables: variational autoencoders and flow-based models. After introducing the idea of latent variable modelling and the central role of inference in it, I will cover flow-based models which combine high expressive power with tractable inference. I will then introduce variational inference for efficient training of intractable models and go over the highly flexible VAE modelling framework that relies on it. Speaker bio: coming soon PLEASE NOTE THAT PHOTOGRAPHS AND FOOTAGE WILL BE TAKEN THROUGHOUT THIS LECTURE.