Deep learning is a double-edged sword. The powerful feature learning ability of deep models can effectively improve classification accuracy. Still, when the training samples for each class are limited ...

Multi-label text classification is a critical task in natural language processing, in which each document may belong to multiple categories. This setting is challenging, as it involves complex label ...