Zero-shot learning (ZSL)

Problem setup in machine learning Zero-shot learning (ZSL) is a problem setup in machine learning where, at test time, a learner observes samples from classes which were not observed during