Jointly Learning to Label Sentences and Tokens
This addresses the challenge of limited annotated data and compositional language for NLP researchers, though it appears incremental as it builds on existing multi-level supervision methods.
The paper tackles the problem of learning robust and interpretable text representations in end-to-end systems by proposing an architecture that jointly labels sentences and tokens, using attention and explicit supervision. Experiments show substantial improvements in both sentence classification and sequence labeling tasks.
Learning to construct text representations in end-to-end systems can be difficult, as natural languages are highly compositional and task-specific annotated datasets are often limited in size. Methods for directly supervising language composition can allow us to guide the models based on existing knowledge, regularizing them towards more robust and interpretable representations. In this paper, we investigate how objectives at different granularities can be used to learn better language representations and we propose an architecture for jointly learning to label sentences and tokens. The predictions at each level are combined together using an attention mechanism, with token-level labels also acting as explicit supervision for composing sentence-level representations. Our experiments show that by learning to perform these tasks jointly on multiple levels, the model achieves substantial improvements for both sentence classification and sequence labeling.