Whenever I see stuff like this I regret not taking any graphics courses at school. Modeling reality is very hard. I remember a few years ago learning about how Pixar did hair simulation and finding it fascinating. The Incredibles was one of the first ones in which they seriously took on the physics of moving hair, and then went all out with their physics engine for Brave. In this article they discuss surface modeling, and a new technique they’re using to speed up rendering.
The application of these simple machine learning concepts keep impressing me more and more. Autoencoders are a very simple idea. If anything, click through to see the side-by-side video.