[AIRG] Fall 2019 Schedule, Locations, and More

[Ronak Mehta <ronakrm@xxxxxxxxxxx>]

Hey everyone,

Again, welcome back to Madison! We'll be starting up AIRG soon, but are interested in your thoughts and opinions as to what you would like AIRG to be. In the next few days we'll be sending out a survey about how AIRG may better serve your needs and interests.

Tentatively we will be sticking to the 4pm-5pm Wednesday time, with the location dependent on survey feedback. 

If you've already blocked out this afternoon, you're encouraged to check out the statistics seminar this week, it looks like it will be a nice overview of standard deep generative models.

https://stat.wisc.edu/event/statistics-seminar-2019-09-11/

Statistics Seminar â 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 weâll 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, weâll see what the research directions are in order to get better and more robust models.

If you have any questions or thoughts feel free to contact me, and look out for the survey later this week!

-Ronak