MAtch, eXpand and Improve: Unsupervised Finetuning for Zero-Shot Action Recognition with Language KnowledgeWei LinLeonid Karlinskyet al.2023ICCV 2023Conference paper