Definition Modelling for Appropriate Specificity

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 2499–2509 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 2499 Definition Modelling for Appropriate Specificity Han Huang†, Tomoyuki Kajiwara‡, Yuki Arase† †Graduate School of Information Science and Technology, Osaka University ‡Graduate School of Science and Engineering, Ehime University †{huang.han, arase}@ist.osaka-u.ac.jp ‡kajiwara@cs.ehime-u.ac.jp Abstract Definition generation techniques aim to gen- erate a definition of a target word or phrase given a context. In previous studies, re- searchers have faced various issues such as the out-of-vocabulary problem and over/under- specificity problems. Over-specific defini- tions present narrow word meanings, whereas under-specific definitions present general and context-insensitive meanings. Herein, we propose a method for definition generation with appropriate specificity. The proposed method addresses the aforementioned prob- lems by leveraging a pre-trained encoder- decoder model, namely Text-to-Text Trans- fer Transformer, and introducing a re-ranking mechanism to model specificity in defini- tions.1 Experimental results on standard eval- uation datasets indicate that our method signif- icantly outperforms the previous state-of-the- art method. Moreover, manual evaluation con- firms that our method effectively addresses the over/under-specificity problems. 1 Introduction The usage of a word or phrase changes over time and new words and phrases emerge every day; therefore, the maintenance of their meanings in dic- tionaries is crucial but labour-intensive and time- consuming. Such definitions are also useful for computer-aided language learning (CALL), which helps language learners learn a target word or phrase (Shardlow, 2014; Srikanth and Li, 2021). A definition generation technique aims to automatically generate a textual definition for a target word or phrase (referred to as ‘target’ herein) in a given sentence containing the target (referred to as ‘local context’ herein). Noraset et al. (2017) employed a static word embedding that models the usage of a target word or phrase, and Ni and Wang 1Code is available at https://github.com/ amanotaiga/Definition_Modeling_Project Target Hammer Local Context Health professionals are mo- bilising to condemn the gov- ernment , propose major struc- tural reforms , and hammer the ineffectual minister . Reference attack or criticize forcefully and relentlessly (Ishiwatari et al., 2019) a person who hits something Proposed method attack or criticize severely Target Bang Local Context Young andrew wilson , until a bang on the head necessitated his withdrawal , again played very well . Reference a sudden painful blow (Ishiwatari et al., 2019) ( of a person ) strike or strike ( something ) with a sudden sharp noise Proposed method a sudden sharp blow Table 1: Examples of generated definitions by a previ- ous study and our method; the previous study struggles with under- and over-specific generations. (2017), Gadetsky et al. (2018) and Ishiwatari et al. (2019) used an encoder-decoder model to generate a definition for a given sentence containing the tar- get. However, these previous studies are limited by two problems: out-of-vocabulary (OOV) and over- and under-specificity (Noraset et al., 2017; Mickus et al., 2019; Li et al., 2020), as shown in Table 1. An under-specific definition denotes a general definition wherein part of …

Content truncated to 500 words