TY - JOUR

T1 - “The Rodney Dangerfield of Stylistic Devices”

T2 - End-to-End Detection and Extraction of Vossian Antonomasia Using Neural Networks

AU - Schwab, Michel

AU - Jäschke, Robert

AU - Fischer, Frank

N1 - Funding Information: We acknowledge support by the German Research Foundation (DFG) and the Open Access Publication Fund of Humboldt-Universität zu Berlin.

PY - 2022/6/9

Y1 - 2022/6/9

N2 - Vossian Antonomasia (VA) is a well-known stylistic device based on attributing a certain property to a person by relating them to another person who is famous for this property. Although the morphological and semantic characteristics of this phenomenon have long been the subject of linguistic research, little is known about its distribution. In this paper, we describe end-to-end approaches for detecting and extracting VA expressions from large news corpora in order to study VA more broadly. We present two types of approaches: binary sentence classifiers that detect whether or not a sentence contains a VA expression, and sequence tagging of all parts of a VA on the word level, enabling their extraction. All models are based on neural networks and outperform previous approaches, best results are obtained with a fine-tuned BERT model. Furthermore, we study the impact of training data size and class imbalance by adding negative (and possibly noisy) instances to the training data. We also evaluate the models' performance on out-of-corpus and real-world data and analyze the ability of the sequence tagging model to generalize in terms of new entity types and syntactic patterns.

AB - Vossian Antonomasia (VA) is a well-known stylistic device based on attributing a certain property to a person by relating them to another person who is famous for this property. Although the morphological and semantic characteristics of this phenomenon have long been the subject of linguistic research, little is known about its distribution. In this paper, we describe end-to-end approaches for detecting and extracting VA expressions from large news corpora in order to study VA more broadly. We present two types of approaches: binary sentence classifiers that detect whether or not a sentence contains a VA expression, and sequence tagging of all parts of a VA on the word level, enabling their extraction. All models are based on neural networks and outperform previous approaches, best results are obtained with a fine-tuned BERT model. Furthermore, we study the impact of training data size and class imbalance by adding negative (and possibly noisy) instances to the training data. We also evaluate the models' performance on out-of-corpus and real-world data and analyze the ability of the sequence tagging model to generalize in terms of new entity types and syntactic patterns.

KW - BERT

KW - binary classification

KW - information extraction

KW - metaphor

KW - metonymy

KW - neural network

KW - sequence tagging

KW - Vossian Antonomasia

UR - http://www.scopus.com/inward/record.url?scp=85133463750&partnerID=8YFLogxK

U2 - 10.3389/frai.2022.868249

DO - 10.3389/frai.2022.868249

M3 - Article

AN - SCOPUS:85133463750

VL - 5

JO - Frontiers in Artificial Intelligence

JF - Frontiers in Artificial Intelligence

M1 - 868249

ER -