Zero-Shot Learning of Language Models for Describing Human Actions Based on Semantic Compositionality of Actions

19251

Knowledge Triples

Source Entity	Relationship	Target Entity
Natural Language Descriptions Task	isBasedOn	Kinect Videos Material
Semantic Compositionality OtherEntity	composedOf	Elementary Semantic Components OtherEntity
Topic Dependent Language Models Method	appliesTo	Natural Language Descriptions Task
Topic Dependent Language Models Method	leverages	Semantic Compositionality OtherEntity
Zero Shot Learning Method	enables	Topic Dependent Language Models Method