A new benchmark for group distribution shifts in hand grasp regression for object manipulation. Can meta-learning raise the bar?

Open in new window