Part-aware Unified Representation of Language and Skeleton for Zero-shot Action Recognition | Read Paper on Bytez