»A Buster Keaton of Linguistics« – First Automated Approaches for the Extraction of Vossian Antonomasia
M. Schwab, R. Jäschke, F. Fischer, and J. Strötgen. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, page 6239--6244. Association for Computational Linguistics, (November 2019)
DOI: 10.18653/v1/D19-1647
Abstract
Attributing a particular property to a person by naming another person, who is typically well-known for the respective property, is called a Vossian antonomasia (VA). While identifying this subtype of metonymy is of particular interest in the study of stylistics, it is also a source of errors in relation and fact extraction as an explicitly mentioned entity occurs only metaphorically and should not be associated with respective contexts. Despite rather simple syntactic variations, the automatic extraction of VA was never addressed so far as it requires a deep semantic understanding of mentioned entities and underlying relations, and is thus very challenging. In this paper, we propose the first method for the extraction of VA that works completely automatically. Our approaches use distant supervision based on Wikidata, NER and relies on a bi-directional LSTM for postprocessing. The evaluation on 1.8 million articles of the New York Times corpus shows that our approach significantly outperforms the only existing semi-automatic approach for VA identification by more than 30 percent points in precision.
Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing
%0 Conference Paper
%1 schwab2019buster
%A Schwab, Michel
%A Jäschke, Robert
%A Fischer, Frank
%A Strötgen, Jannik
%B Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing
%D 2019
%I Association for Computational Linguistics
%K 2019 blstm dh emnlp lstm myown nlp vossanto
%P 6239--6244
%R 10.18653/v1/D19-1647
%T »A Buster Keaton of Linguistics« – First Automated Approaches for the Extraction of Vossian Antonomasia
%U https://www.aclweb.org/anthology/D19-1647
%X Attributing a particular property to a person by naming another person, who is typically well-known for the respective property, is called a Vossian antonomasia (VA). While identifying this subtype of metonymy is of particular interest in the study of stylistics, it is also a source of errors in relation and fact extraction as an explicitly mentioned entity occurs only metaphorically and should not be associated with respective contexts. Despite rather simple syntactic variations, the automatic extraction of VA was never addressed so far as it requires a deep semantic understanding of mentioned entities and underlying relations, and is thus very challenging. In this paper, we propose the first method for the extraction of VA that works completely automatically. Our approaches use distant supervision based on Wikidata, NER and relies on a bi-directional LSTM for postprocessing. The evaluation on 1.8 million articles of the New York Times corpus shows that our approach significantly outperforms the only existing semi-automatic approach for VA identification by more than 30 percent points in precision.
@inproceedings{schwab2019buster,
abstract = {Attributing a particular property to a person by naming another person, who is typically well-known for the respective property, is called a Vossian antonomasia (VA). While identifying this subtype of metonymy is of particular interest in the study of stylistics, it is also a source of errors in relation and fact extraction as an explicitly mentioned entity occurs only metaphorically and should not be associated with respective contexts. Despite rather simple syntactic variations, the automatic extraction of VA was never addressed so far as it requires a deep semantic understanding of mentioned entities and underlying relations, and is thus very challenging. In this paper, we propose the first method for the extraction of VA that works completely automatically. Our approaches use distant supervision based on Wikidata, NER and relies on a bi-directional LSTM for postprocessing. The evaluation on 1.8 million articles of the New York Times corpus shows that our approach significantly outperforms the only existing semi-automatic approach for VA identification by more than 30 percent points in precision.
},
added-at = {2019-08-16T22:28:06.000+0200},
author = {Schwab, Michel and Jäschke, Robert and Fischer, Frank and Strötgen, Jannik},
biburl = {https://www.bibsonomy.org/bibtex/26018ca7a509f8d549c95bc816dabf738/jaeschke},
booktitle = {Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing},
doi = {10.18653/v1/D19-1647},
interhash = {d5ae16c6fe8dc053e134e882d78d9c42},
intrahash = {6018ca7a509f8d549c95bc816dabf738},
keywords = {2019 blstm dh emnlp lstm myown nlp vossanto},
month = nov,
pages = {6239--6244},
publisher = {Association for Computational Linguistics},
series = {EMNLP '19},
timestamp = {2022-11-16T08:53:02.000+0100},
title = {»A Buster Keaton of Linguistics« – First Automated Approaches for the Extraction of Vossian Antonomasia},
url = {https://www.aclweb.org/anthology/D19-1647},
year = 2019
}