Zero shot learning

Zero shot learning is the ability of a model to generalize to unseen inputs without any additional information being passed to them. It is in comparison to one shot learning and few shot learning.