Data-Driven Metaphor Recognition and Explanation

Abstract

Recognizing metaphors and identifying the source-target mappings is an important task as metaphorical text poses a big challenge for machine reading. To address this problem, we automatically acquire a metaphor knowledge base and an isA knowledge base from billions of web pages. Using the knowledge bases, we develop an inference mechanism to recognize and explain the metaphors in the text. To our knowledge, this is the first purely data-driven approach of probabilistic metaphor acquisition, recognition, and explanation. Our results shows that it significantly outperforms other state-of-the-art methods in recognizing and explaining metaphors.

Design and Evaluation of Metaphor Processing Systems

“…The word or relation levels provide the most information and have been the focus of the majority of approaches (Gedigian et al 2006;Shutova, Sun, and Korhonen 2010;Turney et al 2011;Gandy et al 2013;Heintz et al 2013;Hovy et al 2013;Neuman et al 2013;Shutova 2013;Shutova and Sun 2013;Wilks et al 2013). However, some works annotated metaphor at the sentence level (Krishnakumaran and Zhu 2007;Dunn 2013a;Li, Zhu, and Wang 2013;Mohler et al 2013;Tsvetkov, Mukomel, and Gershman 2013).…”

“…Recent years have seen a growing interest in statistical modeling of metaphor (Mason 2004;Gedigian et al 2006;Shutova 2010;Shutova, Sun, and Korhonen 2010;Turney et al 2011;Dunn 2013a;Heintz et al 2013;Hovy et al 2013;Li, Zhu, and Wang 2013;Mohler et al 2013;Shutova and Sun 2013;Shutova Design and Evaluation of Metaphor Processing Systems Tsvetkov, Mukomel, and Gershman 2013), with many new techniques opening routes for improving system accuracy and robustness. A wide range of methods have been proposed and investigated by the community, including supervised (Gedigian et al 2006;Dunn 2013a;Hovy et al 2013;Mohler et al 2013;Tsvetkov, Mukomel, and Gershman 2013) and unsupervised (Heintz et al 2013;Shutova and Sun 2013) learning, distributional approaches (Shutova 2010(Shutova , 2013Shutova, Van de Cruys, and Korhonen 2012), lexical resource-based methods (Krishnakumaran and Zhu 2007;Wilks et al 2013), psycholinguistic features (Turney et al 2011;Gandy et al 2013;Neuman et al 2013;, and Web search (Veale and Hao 2008;Bollegala and Shutova 2013;Li, Zhu, and Wang 2013). Although individual approaches tackling individual aspects of metaphor have met with success, the insights gained from these experiments are still difficult to integrate into a single computational metaphor modeling landscape, because of the lack of a unified task definition, a shared data set, and well-defined evaluation standards.…”

Déjà Image-Captions: A Corpus of Expressive Descriptions in Repetition

“…Figurative language: There has been substantial work for detecting and interpreting figurative language (Shutova, 2010;Li et al, 2013;Kuznetsova et al, 2013a;Tsvetkov et al, 2014), while relatively less work on generating creative or figurative language (Veale, 2011;Ozbal and Strapparava, 2012). We probe data-driven approaches to creative language generation in the context of image captioning.…”

Bigrams and BiLSTMs Two Neural Networks for Sequential Metaphor Detection

“…In this frame, semantic spaces appear to be a very flexible and powerful frame to model such semantic domains in terms of words' clustering and distributional similarity (Mohler et al, 2014). Also, semantic spaces are relatively easy to build and handle, giving them an advantage over more time-consuming resources, such as very large knowledge bases and "is A" bases from web corpora, as in Li et al (2013). Gutierrez et al (2016) use the flexibility of word vectors to study the compositional nature of metaphors and the possibility of modeling it in a semantic space.…”