Biomedical Concept Normalization by Leveraging Hypernyms

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3512–3517 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 3512 Biomedical Concept Normalization by Leveraging Hypernyms Cheng Yan1,2, Yuanzhe Zhang1,2, Kang Liu1,2, Jun Zhao1,2, Yafei Shi3, Shengping Liu3 1National Laboratory of Pattern Recognition, Institute of Automation, CAS, Beijing, China 2School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China 3Unisound AI Technology Co., Ltd. {cheng.yan, yzzhang, kliu, jzhao}@nlpr.ia.ac.cn, {shiyafei, liushengping}@unisound.com Abstract Biomedical Concept Normalization (BCN) is widely used in biomedical text processing as a fundamental module. Owing to numerous sur- face variants of biomedical concepts, BCN still remains challenging and unsolved. In this pa- per, we exploit biomedical concept hypernyms to facilitate BCN. We propose Biomedical Con- cept Normalizer with Hypernyms (BCNH), a novel framework that adopts list-wise training to make use of both hypernyms and synonyms, and also employs norm constraint on the representation of hypernym-hyponym entity pairs. The experimental results show that BCNH outperform the previous state-of-the-art model on the NCBI dataset. Code will be available at https://github.com/ yan-cheng/BCNH. 1 Introduction Biomedical Concept Normalization (BCN) plays an important and prerequisite role in biomedi- cal text processing. The goal of BCN is to link the entity mention in the context to its nor- malized CUI (Unique Concept Identifier) in the biomedical dictionaries such as UMLS (Bodenrei- der, 2004), SNOMED-CT (Spackman et al., 1997) and MedDRA (Brown et al., 1999). Figure 1 is an example of BCN from NCBI dataset (Do˘gan et al., 2014), the mention B-cell non-Hodgkins lym- phomas should be linked to D016393 Lymphoma, B-Cell in the MEDIC (Davis et al., 2012) dictio- nary. Recent works on BCN usually adopt encoders like CNN (Li et al., 2017), LSTM (Phan et al., 2019), ELMo (Peters et al., 2018; Schumacher et al., 2020) or BioBERT (Lee et al., 2020; Fakhraei et al., 2019; Ji et al., 2020) to embed both the men- tion and the concept’s name entities, and then feed the representations to the following classifier or ranking network to determine the corresponding concept in the biomedical dictionary. However, biomedical dictionaries are generally sparse in na- ture: a concept is usually provided with only CUI, referred name (recommended concept name string), synonyms (acceptable name variants, synonyms), and related concepts (mainly hypernym concepts). Therefore, effectively using the limited information in the biomedical dictionary where the candidate entities came from is paramount for the BCN task. For concept’s synonym entities, recent BNE (Phan et al., 2019) and BIOSYN (Sung et al., 2020) tries to make full use of them by synonym marginal- ization to enhance biomedical entity representation and achieved consistent performance improvement. Unfortunately, previous works generally ignore concept hypernym hierarchy structure, which is exactly the initial motivation of biomedical dictio- nary: organization of thousands of concepts under a unified and multi-level hierarchical classification schema. We believe that leveraging hypernym informa- tion in the biomedical dictionary can improve the BCN performance based on two intuitions. First, hard negative sampling (Fakhraei et al., 2019; Phan et al., …

Content truncated to 500 words