Deep Span Representations for Named Entity Recognition
This work addresses performance degradation in NER for long-span and overlapping entities, which is a domain-specific problem for natural language processing applications.
The paper tackles the problem of named entity recognition (NER) by addressing the ineffectiveness of existing span-based models for long-span entities and overlapping spans, proposing DSpERT which achieves competitive or higher performance than state-of-the-art systems on eight NER benchmarks, with particular improvements on long-span entities and nested structures.
Span-based models are one of the most straightforward methods for named entity recognition (NER). Existing span-based NER systems shallowly aggregate the token representations to span representations. However, this typically results in significant ineffectiveness for long-span entities, a coupling between the representations of overlapping spans, and ultimately a performance degradation. In this study, we propose DSpERT (Deep Span Encoder Representations from Transformers), which comprises a standard Transformer and a span Transformer. The latter uses low-layered span representations as queries, and aggregates the token representations as keys and values, layer by layer from bottom to top. Thus, DSpERT produces span representations of deep semantics. With weight initialization from pretrained language models, DSpERT achieves performance higher than or competitive with recent state-of-the-art systems on eight NER benchmarks. Experimental results verify the importance of the depth for span representations, and show that DSpERT performs particularly well on long-span entities and nested structures. Further, the deep span representations are well structured and easily separable in the feature space.