Multi-Task Learning for Chemical Named Entity Recognition with Chemical Compound Paraphrasing

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 6244–6249, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 6244 Multi-Task Learning for Chemical Named Entity Recognition with Chemical Compound Paraphrasing Taiki Watanabe1,3∗ watanabe-taiki@fujitsu.com Akihiro Tamura2,3 Takashi Ninomiya2,3 {tamura,ninomiya}@cs.ehime-u.ac.jp Takuya Makino1,3 Tomoya Iwakura1,3 {makino.takuya,iwakura.tomoya}@fujitsu.com 1 Fujitsu Laboratories, Ltd. 2 Ehime University 3 RIKEN AIP-FUJITSU Collaboration Center Abstract We propose a method to improve named entity recognition (NER) for chemical compounds using multi-task learning by jointly training a chemical NER model and a chemical com- pound paraphrase model. Our method en- ables the long short-term memory (LSTM) of the NER model to capture chemical com- pound paraphrases by sharing the parameters of the LSTM and character embeddings be- tween the two models. The experimental re- sults on the BioCreative IV’s CHEMDNER task show that our method improves chemi- cal NER and achieves state-of-the-art perfor- mance (+1.43 F-score). 1 Introduction Named Entity Recognition (NER) is one of the important basic technologies for Natural Lan- guage Processing (NLP) such as Information Extraction and Entity Linking. LSTM-CRF NER models, which combine a conditional ran- dom field (CRF) and a long short-term mem- ory (LSTM), have achieved high performance (Lample et al., 2016; Ma and Hovy, 2016). The LSTM-CRF models that use a neural language model (NLM) pre-trained from a large-scale un- labeled corpus (Akbik et al., 2018; Peters et al., 2018) have shown state-of-the-art performances on the CoNLL 2003 shared task (Sang et al., 2003). Chemical compound databases are widely used for investigating properties of chemical com- pounds or for developing new chemical products. However, updating the databases by hand is hard because new findings on chemical compounds are mainly reported in scientific papers and patents ev- ery day. Hence, chemical NER has been studied to recognize chemical compounds from chemical ∗Taiki Watanabe belonged to Ehime University when this work was done. documents (Leaman et al., 2015; Lu et al., 2015; Lin et al., 2018). One of the problems in chemical NER is nota- tion variants of chemical compounds. For exam- ple, Phenylalanine has different notations such as L-β-phenylalanine and (S)-2-Benzylglycine. The average number of notations of each compound in PubChem1 is approximately 3.88.2 If these expressions are dealt with differently, the statistics of the same compound can be dis- persed, especially for low frequency compounds. In other words, the more distinct these representa- tions are, the more difficult identifying chemical compound entities becomes. However, existing chemical NER methods do not deal with notation variants of chemical compounds, derived from the partial structures or the notation fluctuation pecu- liar to these chemical compounds. We propose HanPaNE, which Handles Paraphrase in NER by utilizing chemical com- pound paraphrase pairs as multi-task learning. To train expression identity of different nota- tions, HanPaNE learns shared parameters between paraphrases using multi-task learning on NER and paraphrase generation. This contrasts with widely used approaches that automatically aug- ment training data by replacing NEs …

Content truncated to 500 words