I've been interested in some work where people have used GANs to increase the amount of data in use (for training a classification model for example). Are GANs in a place where it is realistic enough to do so and can we minimise effectively propagating the error from the GAN into the new model?
You showed up just in time for the waffles! And this tiny ad:
Building a Better World in your Backyard by Paul Wheaton and Shawn Klassen-Koop