A Neural Multi-digraph Model for Chinese NER with Gazetteers

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1462–1467 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 1462 A Neural Multi-digraph Model for Chinese NER with Gazetteers Ruixue Ding1, Pengjun Xie1, Xiaoyan Zhang2, Wei Lu3, Linlin Li1 and Luo Si1 1Alibaba Group 2Beihang University, China 3Singapore University of Technology and Design {ada.drx,chengchen.xpj,linyan.lll,luo.si}@alibaba-inc.com xiaoyan.loic@gmail.com, luwei@sutd.edu.sg Abstract Gazetteers were shown to be useful resources for named entity recognition (NER) (Ratinov and Roth, 2009). Many existing approaches to incorporating gazetteers into machine learn- ing based NER systems rely on manually de- fined selection strategies or handcrafted tem- plates, which may not always lead to opti- mal effectiveness, especially when multiple gazetteers are involved. This is especially the case for the task of Chinese NER, where the words are not naturally tokenized, leading to additional ambiguities. To automatically learn how to incorporate multiple gazetteers into an NER system, we propose a novel approach based on graph neural networks with a multi- digraph structure that captures the information that the gazetteers offer. Experiments on vari- ous datasets show that our model is effective in incorporating rich gazetteer information while resolving ambiguities, outperforming previous approaches. 1 Introduction Previous work (Ratinov and Roth, 2009) shows that NER is a knowledge intensive task. Back- ground knowledge is often incorporated into an NER system in the form of named entity (NE) gazetteers (Seyler et al., 2018). Each gazetteer is typically a list containing NEs of the same type. Many earlier research efforts show that an NER model can benefit from the use of gazetteers (Li et al., 2005). On the one hand, the use of NE gazetteers alleviates the need of manually label- ing the data and can handle rare and unseen cases (Wang et al., 2018). On the other hand, resources of gazetteers are abundant. Many gazetteers have been manually created by previous studies (Za- min and Oxley, 2011). Besides, gazetteers can also be easily constructed from knowledge bases (e.g., Freebase (Bollacker et al., 2008)) or com- � � � � � � � � � Open Three At North Capital Human People Public Park Zhang San PER2 People’s Park Beijing LOC1 LOC2 PER2 Zhang Sanzai PER1 Beijing citizen Wrong matches Correct matches The actual translation: Zhang San is at the Beijing People’s Park LOC1 Figure 1: Example of Entity Matching mercial data sources (e.g., product catalogues of e-commence websites). While such background knowledge can be help- ful, in practice the gazetteers may also con- tain irrelevant and even erroneous information which harms the system’s performance (Chiu and Nichols, 2016). This is especially the case for Chi- nese NER, where enormous errors can be intro- duced due to wrongly matched entities. Chinese language is inherently ambiguous since the granu- larity of words is less well defined than other lan- guages (such as English). Thus massive wrongly matched entities can be generated with the use of gazetteers. As we can see from the example shown in Figure 1, matching …

Content truncated to 500 words