Label-Enhanced Hierarchical Contextualized Representation for Sequential Metaphor Identification

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3533–3543 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 3533 Label-Enhanced Hierarchical Contextualized Representation for Sequential Metaphor Identification Shuqun Li∗, Liang Yang†∗, Weidong He, Shiqi Zhang, Jingjie Zeng and Hongfei LIN Department of Computer Science and Technology, Dalian University of Technology, China shuqunli@mail.dlut.edu.cn, liang@dlut.edu.cn {heweidong,22009191,jjwind}@mail.dlut.edu.cn hflin@dlut.edu.cn Abstract Recent metaphor identification approaches mainly consider the contextual text features within a sentence or introduce external lin- guistic features to the model. But they usu- ally ignore the extra information that the data can provide, such as the contextual metaphor information and broader discourse informa- tion. In this paper, we propose a model aug- mented with hierarchical contextualized rep- resentation to extract more information from both sentence-level and discourse-level. At the sentence level, we leverage the metaphor infor- mation of words that except the target word in the sentence to strengthen the reasoning ability of our model via a novel label-enhanced con- textualized representation. At the discourse level, the position-aware global memory net- work is adopted to learn long-range depen- dency among the same words within a dis- course. Finally, our model combines the repre- sentations obtained from these two parts. The experiment results on two tasks of the VUA dataset show that our model outperforms ev- ery other state-of-the-art method that also does not use any external knowledge except what the pre-trained language model contains. 1 Introduction Metaphor is a type of figurative language and its essence is understanding and experiencing one kind of thing in terms of another(Lakoff and Johnson, 1980). As a common language expression, we of- ten use metaphors to express our thoughts vividly and concisely in daily communication. For exam- ple, in the sentence It is one of the keys for suc- cess of a commercial product, keys is used to help understand the importance of It. However, this characteristic of metaphor makes it challenging to identify metaphors in texts. But the identification *Both authors contributed equally to this research †Corresponding author of metaphors is meaningful and can help us to un- derstand the meaning of the texts, from which many downstream applications such as machine transla- tion(Koglin, 2015) and opinion mining(Shutova et al., 2013) can benefit. Recent metaphor researches(Gao et al., 2018; Mao et al., 2019), and ACL 2020 Metaphor Shared Task(Leong et al., 2020) regard it as a sequence labeling task. Although many previous works have explored ways to enhance the contextualized repre- sentation within a sentence(Gao et al., 2018; Mao et al., 2019), or to introduce some external knowl- edge(Rohanian et al., 2020; Chen et al., 2020; Wan and Xing, 2020), most of them do not make full use of the information in the dataset, from which the metaphor identification process may benefit. Firstly, when considering the metaphoricity of the target word, the metaphor information of other words in the sentence can also be helpful. E.g., in the sentence He find himself in the position of the gambler who gambled all and …

Content truncated to 500 words