- This event has passed.
Statistics Seminar – Canziani
September 11 @ 4:00 pm - 5:00 pm
Title: Unsupervised Deep Learning: Autoencoders and Generative Adversarial Nets
Presenter: Alfredo Canziani
Abstract: The brain has about 10^14 synapses and we only live for about 10^9 seconds. Even just considering binary synapses, a learning algorithm would still require 10^5 bits of information per second to learn all the connections in the brain. This motivates the idea that humans must do a lot of unsupervised learning since the perceptual input (including proprioception) is the only place we can get 10^5 dimensions of constraint per second.» Geoffrey Hinton.
In this lecture we’ll explore three common network architectures (denoising & variational autoencoders and generative adversarial nets) that learn to capture the structure and properties of high dimensional data without supervision, which is usually very expensive to get. Using them as generative models, we can compare interpolation in the latent and input spaces, paying attention to when we end up outside the data manifold. Finally, we’ll see what the research directions are in order to get better and more robust models.