Biomedical Entity Representations with Synonym Marginalization

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 3641–3650 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 3641 Biomedical Entity Representations with Synonym Marginalization Mujeen Sung Hwisang Jeon Jinhyuk Lee† Jaewoo Kang† Korea University {mujeensung,j hs,jinhyuk lee,kangj}@korea.ac.kr Abstract Biomedical named entities often play impor- tant roles in many biomedical text mining tools. However, due to the incompleteness of provided synonyms and numerous varia- tions in their surface forms, normalization of biomedical entities is very challenging. In this paper, we focus on learning representations of biomedical entities solely based on the syn- onyms of entities. To learn from the incom- plete synonyms, we use a model-based candi- date selection and maximize the marginal like- lihood of the synonyms present in top candi- dates. Our model-based candidates are itera- tively updated to contain more difficult neg- ative samples as our model evolves. In this way, we avoid the explicit pre-selection of negative samples from more than 400K can- didates. On four biomedical entity normal- ization datasets having three different entity types (disease, chemical, adverse reaction), our model BIOSYN consistently outperforms previous state-of-the-art models almost reach- ing the upper bound on each dataset. 1 Introduction Biomedical named entities are frequently used as key features in biomedical text mining. From biomedical relation extraction (Xu et al., 2016; Li et al., 2017a) to literature search engines (Lee et al., 2016), many studies are utilizing biomedical named entities as a basic building block of their methodolo- gies. While the extraction of the biomedical named entities is studied extensively (Sahu and Anand, 2016; Habibi et al., 2017), the normalization of ex- tracted named entities is also crucial for improving the precision of downstream tasks (Leaman et al., 2013; Wei et al., 2015). †Corresponding authors Unlike named entities from general domain text, typical biomedical entities have several different surface forms, making the normalization of biomed- ical entities very challenging. For instance, while two chemical entities ‘motrin’ and ‘ibuprofen’ be- long to the same concept ID (MeSH:D007052), they have completely different surface forms. On the other hand, mentions having similar sur- face forms could also have different meanings (e.g. ‘dystrophinopathy’ (MeSH:D009136) and ‘bestrophinopathy’ (MeSH:C567518)). These ex- amples show a strong need for building latent rep- resentations of biomedical entities that capture se- mantic information of the mentions. In this paper, we propose a novel framework for learning biomedical entity representations based on the synonyms of entities. Previous works on entity normalization mostly train binary classifiers that decide whether the two input entities are the same (positive) or different (negative) (Leaman et al., 2013; Li et al., 2017b; Fakhraei et al., 2019; Phan et al., 2019). Our framework called BIOSYN uses the synonym marginalization technique, which maximizes the probability of all synonym repre- sentations in top candidates. We represent each biomedical entity using both sparse and dense rep- resentations to capture morphological and semantic information, respectively. The candidates are itera- tively updated based on our model’s representations removing the need for an explicit …

Content truncated to 500 words