Multi-Label Learning with Highly Incomplete Data via Collaborative Embedding

Tremendous efforts have been dedicated to improving the effectiveness of multi-label learning with incomplete label assignments. Most of the current techniques assume that the input features of dat