A Boundary-aware Neural Model for Nested Named Entity Recognition

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 357–366, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 357 A Boundary-aware Neural Model for Nested Named Entity Recognition Changmeng Zheng1, Yi Cai1∗, Jingyun Xu1, Ho-fung Leung2 and Guandong Xu3 1School of Software Engineering, South China University of Technology, Guangzhou, China 2The Chinese University of Hong Kong, Hong Kong SAR, China 3Advanced Analytics Institute, University of Technology Sydney, Australia sethecharm@mail.scut.edu.cn, ycai@scut.edu.cn Abstract In natural language processing, it is common that many entities contain other entities in- side them. Most existing works on named en- tity recognition (NER) only deal with flat en- tities but ignore nested ones. We propose a boundary-aware neural model for nested NER which leverages entity boundaries to predict entity categorical labels. Our model can lo- cate entities precisely by detecting boundaries using sequence labeling models. Based on the detected boundaries, our model utilizes the boundary-relevant regions to predict entity cat- egorical labels, which can decrease computa- tion cost and relieve error propagation prob- lem in layered sequence labeling model. We introduce multitask learning to capture the de- pendencies of entity boundaries and their cate- gorical labels, which helps to improve the per- formance of identifying entities. We conduct our experiments on nested NER datasets and the experimental results demonstrate that our model outperforms other state-of-the-art meth- ods. 1 Introduction Named entity recognition (NER) is a task that seeks to locate and classify named entities in un- structured texts into pre-defined categories such as person names, locations or medical codes. NER is generally treated as single-layer sequence label- ing problem (Lafferty et al., 2001; Lample et al., 2016) where each token is tagged with one label. The label is composed by an entity boundary la- bel and a categorical label. For example, a to- ken can be tagged with B-PER, where B indi- cates the boundary of an entity and PER indicates the corresponding entity categorical label. How- ever, when entities are nested within one another, single-layer sequence labeling models can not ex- ∗Corresponding author tract both entities simultaneously. A token con- tained inside many entities has more than one cat- egorical label. Consider an example in Figure 1 from GENIA corpus (Kim et al., 2003), “Human TR Beta 1” is an protein and it is also a part of a DNA “Human TR Beta 1 mRNA”. Both enti- ties contain the same token “Human”. Thus the token should have two different categorical labels. In that case, assigning a single categorical label for “Human” is improper. Figure 1: An example of nested entities and their boundary labels. “B” and “E” indicate the beginning and end of an entity. They are the boundary labels. “I” and “O” denote tokens inside and outside entities, respectively. protein and RNA are categories of enti- ties. Traditional methods coping with nested entities rely on hand-craft features (Shen et al., 2003; Alex et al., 2007) and suffer …

Content truncated to 500 words