Adding Grafts to Natural Language Processing Is a Big Thing

This kind of data modality is something that you almost always want to reach for. It's so common, at least in images, that it's often just like a switch that you flip. You really don't have to think very hard about adding these kinds of augmentations with modern tools such as tenser flow and pye torch. And eno hasn't made it entirely into natural language processing either because those are all so discreet objects. But there's a lot of other opportunities in grafts as well. So the just, the field is kind of open. One that's kind of especially exciting in chemistry is cirality.