Task guided representation learning using compositional models for zero-shot domain adaptation

Shuang Liu, Mete Ozay