Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer

Mina Schütz; Jaqueline Böck; Medina-Petrina Andresel; Armin Kirchknopf; Daria Liakhovets; Djordje Slijepcevic; Alexander Schindler

Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer

Mina Schütz (Speaker), Jaqueline Böck, Medina-Petrina Andresel, Armin Kirchknopf, Daria Liakhovets, Djordje Slijepcevic, Alexander Schindler

Data Science & Artifical Intelligence

St. Pölten University of Applied Sciences

Research output: Chapter in Book or Conference Proceedings › Conference Proceedings with Oral Presentation › peer-review

Abstract

The increase of fake news in today’s society, partially due to the accelerating digital transformation, is a
major problem in today’s world. This year’s CheckThat! Lab 2022 challenge addresses this problem as a
Natural Language Processing (NLP) task aiming to detect fake news in English and German texts. Within
this paper, we present our methodology and results for both, the monolingual (English) and cross-lingual
(German) tasks of the CheckThat! challenge in 2022. We applied the multilingual transformer model
XLM-RoBERTa to solve these tasks by pre-training the models on additional datasets and fine-tuning
them on the original data as well as its translations for the cross-lingual task. Our final model achieves a
macro F1-score of 15,48% and scores the 22𝑡ℎ rank in the benchmark. Regarding the second task, i.e., the
cross-lingual German classification, our final model achieves an F1-score of 19.46% and reaches the 4
𝑡ℎ
rank in the benchmark.

Original language	English
Title of host publication	Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum
Editors	Guglielmo Faggioli, Nicola Ferro, Allan Hanbury, Martin Potthast
Pages	660-670
Number of pages	10
Volume	3180
Publication status	Published - 2022
Event	CheckThat! 2022 - Duration: 5 Sept 2022 → 8 Sept 2022

Conference

Conference	CheckThat! 2022
Period	5/09/22 → 8/09/22

Research Field

Former Research Field - Data Science

Cite this

Schütz, M., Böck, J., Andresel, M.-P., Kirchknopf, A., Liakhovets, D., Slijepcevic, D., & Schindler, A. (2022). Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer. In G. Faggioli, N. Ferro, A. Hanbury, & M. Potthast (Eds.), Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum (Vol. 3180, pp. 660-670). Article 53

@inproceedings{c8281ddbe0b346d8aacd79fed4dacc37,

title = "Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer",

abstract = "The increase of fake news in today{\textquoteright}s society, partially due to the accelerating digital transformation, is amajor problem in today{\textquoteright}s world. This year{\textquoteright}s CheckThat! Lab 2022 challenge addresses this problem as aNatural Language Processing (NLP) task aiming to detect fake news in English and German texts. Withinthis paper, we present our methodology and results for both, the monolingual (English) and cross-lingual(German) tasks of the CheckThat! challenge in 2022. We applied the multilingual transformer modelXLM-RoBERTa to solve these tasks by pre-training the models on additional datasets and fine-tuningthem on the original data as well as its translations for the cross-lingual task. Our final model achieves amacro F1-score of 15,48% and scores the 22푡ℎ rank in the benchmark. Regarding the second task, i.e., thecross-lingual German classification, our final model achieves an F1-score of 19.46% and reaches the 4푡ℎrank in the benchmark.",

author = "Mina Sch{\"u}tz and Jaqueline B{\"o}ck and Medina-Petrina Andresel and Armin Kirchknopf and Daria Liakhovets and Djordje Slijepcevic and Alexander Schindler",

year = "2022",

language = "English",

volume = "3180",

pages = "660--670",

editor = "Faggioli, {Guglielmo } and Ferro, {Nicola } and Hanbury, {Allan } and Potthast, {Martin }",

booktitle = "Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum",

note = "CheckThat! 2022 ; Conference date: 05-09-2022 Through 08-09-2022",

}

Schütz, M, Böck, J, Andresel, M-P, Kirchknopf, A, Liakhovets, D, Slijepcevic, D & Schindler, A 2022, Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer. in G Faggioli, N Ferro, A Hanbury & M Potthast (eds), Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum. vol. 3180, 53, pp. 660-670, CheckThat! 2022, 5/09/22.

Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer. / Schütz, Mina (Speaker); Böck, Jaqueline; Andresel, Medina-Petrina et al.
Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum. ed. / Guglielmo Faggioli; Nicola Ferro; Allan Hanbury; Martin Potthast. Vol. 3180 2022. p. 660-670 53.

Research output: Chapter in Book or Conference Proceedings › Conference Proceedings with Oral Presentation › peer-review

TY - GEN

T1 - Cross-Lingual Fake News Detection with a Large Pre-Trained Transformer

AU - Böck, Jaqueline

AU - Andresel, Medina-Petrina

AU - Kirchknopf, Armin

AU - Liakhovets, Daria

AU - Slijepcevic, Djordje

AU - Schindler, Alexander

A2 - Schütz, Mina

A2 - Faggioli, Guglielmo

A2 - Ferro, Nicola

A2 - Hanbury, Allan

A2 - Potthast, Martin

PY - 2022

Y1 - 2022

N2 - The increase of fake news in today’s society, partially due to the accelerating digital transformation, is amajor problem in today’s world. This year’s CheckThat! Lab 2022 challenge addresses this problem as aNatural Language Processing (NLP) task aiming to detect fake news in English and German texts. Withinthis paper, we present our methodology and results for both, the monolingual (English) and cross-lingual(German) tasks of the CheckThat! challenge in 2022. We applied the multilingual transformer modelXLM-RoBERTa to solve these tasks by pre-training the models on additional datasets and fine-tuningthem on the original data as well as its translations for the cross-lingual task. Our final model achieves amacro F1-score of 15,48% and scores the 22푡ℎ rank in the benchmark. Regarding the second task, i.e., thecross-lingual German classification, our final model achieves an F1-score of 19.46% and reaches the 4푡ℎrank in the benchmark.

AB - The increase of fake news in today’s society, partially due to the accelerating digital transformation, is amajor problem in today’s world. This year’s CheckThat! Lab 2022 challenge addresses this problem as aNatural Language Processing (NLP) task aiming to detect fake news in English and German texts. Withinthis paper, we present our methodology and results for both, the monolingual (English) and cross-lingual(German) tasks of the CheckThat! challenge in 2022. We applied the multilingual transformer modelXLM-RoBERTa to solve these tasks by pre-training the models on additional datasets and fine-tuningthem on the original data as well as its translations for the cross-lingual task. Our final model achieves amacro F1-score of 15,48% and scores the 22푡ℎ rank in the benchmark. Regarding the second task, i.e., thecross-lingual German classification, our final model achieves an F1-score of 19.46% and reaches the 4푡ℎrank in the benchmark.

M3 - Conference Proceedings with Oral Presentation

VL - 3180

SP - 660

EP - 670

BT - Proceedings of the Working Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum

T2 - CheckThat! 2022

Y2 - 5 September 2022 through 8 September 2022

ER -