notesum.ai

Published at November 27

MetaphorShare: A Dynamic Collaborative Repository of Open Metaphor Datasets

cs.CL

Released Date: November 27, 2024

Authors: Joanne Boisson¹, Arif Mehmood¹, Jose Camacho-Collados¹

Aff.: ¹Cardiff NLP, School of Computer Science and Informatics, Cardiff University, United Kingdom

Arxiv: http://arxiv.org/abs/2411.18260v1

Refer to caption

Dataset	Reference	License	N	N dist.	N dist.	%	Expr.	Domain/Source
Name				ctxt.	expr.	met.	PoS
Psycholinguistics
JANK	Jankowiak (2020)	CC BY 4.0	240²²2Anomalous sentences are not shown in the table.	240	120	50	N	constructed examples
CARD_V	Cardillo et al. (2010)	CC BY-NC	280	280	140	50	V	constructed examples
CARD_N	Cardillo et al. (2010, 2017)	CC BY-NC	512	512	256	50	N	constructed examples
NLP
MOH	Mohammad et al. (2016)	see data page³³3https://saifmohammad.com/WebPages/metaphor.html	1632⁴⁴4the original dataset contains 1639 instances. A few duplicated example sentences caused by orthographic variants of the target word, such as distil/distill have been removed.	1632	439	25	V	WordNet examples
NewsMet	Joseph et al. (2023)	Apache-2.0	1205⁵⁵5We show the manually labelled sentences (named gold by the authors)	1205	477	49	V	Fake News Corpus⁶⁶6https://github.com/several27/FakeNewsCorpus
TSV_A	Tsvetkov et al. (2014)	see data page⁷⁷7https://github.com/ytsvetko/metaphor/blob/master/LICENSE.md	1945	1072	687	50	A	various websites
GUT	Gutiérrez et al. (2016)	AFL-3.0	8591	3479	23	54	A	Wikipedia, UKWaC⁸⁸8Ferraresi et al. (2008)…
Multi-Word Expressions
PVC	Tu and Roth (2012)	no license⁹⁹9https://cogcomp.seas.upenn.edu/page/resource_view/26	1348	1348	23	65	V-Prep	BNC
MAD	Tayyar Madabushi et al. (2022b)	GPL-3.0	4558	4554	251	48	NC	Common Crawl
MAGPIE	Haagsma et al. (2020)	CC-BY-4.0	48395	47283	9307	75	Various	BNC, PMB¹⁰¹⁰10Abzianidze et al. (2017).
MIPVU
VUAC_BO	Boisson et al. (2023)	CC BY-SA 3.0	39223	11476	8674	52	Various	VUAC
TONG	Tong et al. (2024)	CC-BY-4.0	1428¹¹¹¹11Original VUAC sentences and apt (original dataset label) paraphrases are counted in the table.	739	861	46	Various	VUAC & paraphrases