Train Robert to Open a Cabinet in the Google Kitchen

In the future i will have rober that can do more than just openin a cabinet right now, we don't hapand have that. We want systems to generalize like that t tackle diverse trains. On both quadruped and biped robots, we now have very effective reinforcement learning,. Thereis this paper called rapid motion adaptation from the papata slab at sime. The tested t on a bunch of different things. And if you like, that's exactly the kind of generalization we want.