Fine-grained Information Status Classification Using Discourse Context-Aware Self-Attention
This work addresses the need for more accurate natural language processing tools for discourse analysis, particularly in bridging anaphora recognition, but it is incremental as it builds on prior methods with neural enhancements.
The paper tackled the problem of fine-grained information status classification and bridging anaphora recognition by proposing a discourse context-aware self-attention neural network model with BERT, achieving a 4.1% absolute accuracy improvement on the ISNotes corpus and a 3.9% F1 improvement for bridging anaphora recognition.
Previous work on bridging anaphora recognition (Hou et al., 2013a) casts the problem as a subtask of learning fine-grained information status (IS). However, these systems heavily depend on many hand-crafted linguistic features. In this paper, we propose a discourse context-aware self-attention neural network model for fine-grained IS classification. On the ISNotes corpus (Markert et al., 2012), our model with the contextually-encoded word representations (BERT) (Devlin et al., 2018) achieves new state-of-the-art performances on fine-grained IS classification, obtaining a 4.1% absolute overall accuracy improvement compared to Hou et al. (2013a). More importantly, we also show an improvement of 3.9% F1 for bridging anaphora recognition without using any complex hand-crafted semantic features designed for capturing the bridging phenomenon.