MelBERT: Metaphor Detection via Contextualized Late Interaction using Metaphorical Identification Theories

Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1763–1773 June 6–11, 2021. ©2021 Association for Computational Linguistics 1763 MelBERT: Metaphor Detection via Contextualized Late Interaction using Metaphorical Identification Theories Minjin Choi1, Sunkyung Lee1, Eunseong Choi1, Heesoo Park2, Junhyuk Lee1, Dongwon Lee3 and Jongwuk Lee1 1Sungkyunkwan University, Republic of Korea, 2Bering Lab, Republic of Korea, 3The Pennsylvania State University, United States 1{zxcvxd,sk1027,eunseong,ljhjoon00,jongwuklee}@skku.edu 2heesoo.park@beringlab.com 3dongwon@psu.edu Abstract Automated metaphor detection is a challeng- ing task to identify the metaphorical expres- sion of words in a sentence. To tackle this problem, we adopt pre-trained contextualized models, e.g., BERT and RoBERTa. To this end, we propose a novel metaphor detection model, namely metaphor-aware late interac- tion over BERT (MelBERT). Our model not only leverages contextualized word representa- tion but also benefits from linguistic metaphor identification theories to detect whether the tar- get word is metaphorical. Our empirical re- sults demonstrate that MelBERT outperforms several strong baselines on four benchmark datasets, i.e., VUA-18, VUA-20, MOH-X, and TroFi. 1 Introduction As the conceptual and cognitive mapping of words, a metaphor is a common language expression rep- resenting other concepts rather than taking literal meanings of words in context (Lakoff and Johnson, 1980; Lagerwerf and Meijers, 2008). For instance, in the sentence “hope is on the horizon,” the word “horizon” does not literally mean the line at the earth’s surface. It is a metaphorical expression to describe a positive situation. Therefore, the mean- ing of “horizon” is context-specific and different from its literal definition. As the metaphor plays a key role in cognitive and communicative functions, it is essential to un- derstand contextualized and unusual meanings of words (e.g., metaphor, metonymy, and personifica- tion) in various natural language processing (NLP) tasks, e.g., machine translation (Shi et al., 2014), sentiment analysis (Cambria et al., 2017), and dia- logue systems (Dybala and Sayama, 2012). A lot of existing studies have developed various compu- tational models to recognize metaphorical words in a sentence. Automated metaphor detection aims at identify- ing metaphorical expressions using computational models. Existing studies can be categorized into three pillars. First, feature-based models employ various hand-crafted features (Shutova et al., 2010; Turney et al., 2011; Shutova and Sun, 2013; Broad- well et al., 2013; Tsvetkov et al., 2014; Bulat et al., 2017). Although simple and intuitive, they are highly sensitive to the quality of a corpus. Second, some studies (Wu et al., 2018; Gao et al., 2018; Mao et al., 2019) utilize recurrent neural networks (RNNs), which are suitable for analyzing the se- quential structure of words. However, they are limited to understanding the diverse meanings of words in context. Lastly, the pre-trained contex- tualized models, e.g., BERT (Devlin et al., 2019) and RoBERTa (Liu et al., 2019), have been used for detecting metaphors (Chen et al., 2020; Gong et al., 2020; Su et al., 2020). Owing to the power- ful representation capacity, such models have been successful for addressing various NLP tasks (Wang et al., …

Content truncated to 500 words