 My name is Austin Sendek and I'm a Ph.D. student here in the Applied Physics Department of Stanford. And the research problem that I'm trying to solve is the problem of exploding batteries. And we've seen whether it's in the Dreamliner, the Tesla, hoverboards, or cell phones, there's been a major problem with exploding batteries. And part of the problem is because the lithium-ion batteries that we use in these technologies have flammable liquids as the electrolyte. So what my lab wants to do is to replace that flammable liquid with a stable solid that can do the same job. And this is a hard problem. People have been working on this problem looking for good solids for this purpose for several decades, and we don't have several more decades to wait. So what we want to do is take a fundamentally new approach. So let's think about, for example, facial recognition algorithms. We have these two pictures of some very good-looking guy and let's say that we want to know whether or not there's a face in this picture and if so, if it's the same person. Well, we can prime a facial recognition algorithm to learn what faces look like by showing it lots of examples of photos and eventually it'll learn what a face looks like and what one doesn't. So why can't we do this with materials? Let's say we have lots of examples of good materials and bad materials for this purpose and we want to train an algorithm to learn how to discriminate between the two. And so that's what we've done. And why is that cool? Well, it's cool because with these algorithms, we can now go and we can screen every possible known lithium-containing solid to see if, given the data we have, this is a promising material for an electrolyte application. We take these models and we can screen over 12,000 materials, which is every known lithium-containing solid to look for the next FES-1. And once we find it, we can print it out and we can go to our experimental friends, and we can help guide the next generation of experiments so we don't have to rely on trial and error to find the next best materials.