We present an overview of the second edition of the CheckThat! Lab at CLEF 2019. The lab featured two tasks in two different languages: English and Arabic. Task 1 (English) challenged the participating systems to predict which claims in a political debate or speech should be prioritized for fact-checking. Task 2 (Arabic) asked to (A) rank a given set of Web pages with respect to a check-worthy claim based on their usefulness for fact-checking that claim, (B) classify these same Web pages according to their degree of usefulness for fact-checking the target claim, (C) identify useful passages from these pages, and (D) use the useful pages to predict the claim’s factuality. CheckThat! provided a full evaluation framework, consisting of data in English (derived from fact-checking sources) and Arabic (gathered and annotated from scratch) and evaluation based on mean average precision (MAP) and normalized discounted cumulative gain (nDCG) for ranking, and F $$:1$$ for classification. A total of 47 teams registered to participate in this lab, and fourteen of them actually submitted runs (compared to nine last year). The evaluation results show that the most successful approaches to Task 1 used various neural networks and logistic regression. As for Task 2, learning-to-rank was used by the highest scoring runs for subtask A, while different classifiers were used in the other subtasks. We release to the research community all datasets from the lab as well as the evaluation scripts, which should enable further research in the important tasks of check-worthiness estimation and automatic claim verification.

Overview of the CLEF-2019 CheckThat! Lab: Automatic Identification and Verification of Claims

Barron-Cedeno A.;Da San Martino G.;
2019

Abstract

We present an overview of the second edition of the CheckThat! Lab at CLEF 2019. The lab featured two tasks in two different languages: English and Arabic. Task 1 (English) challenged the participating systems to predict which claims in a political debate or speech should be prioritized for fact-checking. Task 2 (Arabic) asked to (A) rank a given set of Web pages with respect to a check-worthy claim based on their usefulness for fact-checking that claim, (B) classify these same Web pages according to their degree of usefulness for fact-checking the target claim, (C) identify useful passages from these pages, and (D) use the useful pages to predict the claim’s factuality. CheckThat! provided a full evaluation framework, consisting of data in English (derived from fact-checking sources) and Arabic (gathered and annotated from scratch) and evaluation based on mean average precision (MAP) and normalized discounted cumulative gain (nDCG) for ranking, and F $$:1$$ for classification. A total of 47 teams registered to participate in this lab, and fourteen of them actually submitted runs (compared to nine last year). The evaluation results show that the most successful approaches to Task 1 used various neural networks and logistic regression. As for Task 2, learning-to-rank was used by the highest scoring runs for subtask A, while different classifiers were used in the other subtasks. We release to the research community all datasets from the lab as well as the evaluation scripts, which should enable further research in the important tasks of check-worthiness estimation and automatic claim verification.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
301
321
Elsayed T.; Nakov P.; Barron-Cedeno A.; Hasanain M.; Suwaileh R.; Da San Martino G.; Atanasova P.
File in questo prodotto:
File Dimensione Formato  
Elsayed_et_al-2019-Experimental_IR_Meets_Multilinguality,_Multimodality,_and_Interaction,_10th....pdf

accesso riservato

Tipo: Versione (PDF) editoriale
Licenza: Licenza per accesso riservato
Dimensione 953.45 kB
Formato Adobe PDF
953.45 kB Adobe PDF   Visualizza/Apri   Contatta l'autore

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11585/709218
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 44
  • ???jsp.display-item.citation.isi??? 10
social impact