Publication Details

KInITVeraAI at SemEval-2023 Task 3: Simple yet Powerful Multilingual Fine-Tuning for Persuasion Techniques Detection

HROMÁDKA Timo, SMOLEŇ Timotej, REMIŠ Tomáš, PECHER Branislav and SRBA Ivan. KInITVeraAI at SemEval-2023 Task 3: Simple yet Powerful Multilingual Fine-Tuning for Persuasion Techniques Detection. In: 17th International Workshop on Semantic Evaluation, SemEval 2023 - Proceedings of the Workshop. Toronto: Association for Computational Linguistics, 2023, pp. 629-637. ISBN 978-1-959429-99-9. Available from: https://aclanthology.org/2023.semeval-1.86/
Czech title
KInITVeraAI na SemEval-2023 úlohe 3: Jednoduché ale výkonné viacjazyčný fine-tuning pre detekciu techník presviedčania.
Type
conference paper
Language
english
Authors
Hromádka Timo ()
Smoleň Timotej ()
Remiš Tomáš ()
Pecher Branislav, Ing. (DCGM FIT BUT)
Srba Ivan ()
URL
Keywords

multilingual persuasion technique detection, fine-tuning, SemEval

Abstract

This paper presents the best-performing solution to the SemEval 2023 Task 3 on the subtask 3 dedicated to persuasion techniques detection. Due to a high multilingual character of the input data and a large number of 23 predicted labels (causing a lack of labelled data for some language-label combinations), we opted for fine-tuning pre-trained transformer-based language models. Conducting multiple experiments, we find the best configuration, which consists of large multilingual model (XLM-RoBERTa large) trained jointly on all input data, with carefully calibrated confidence thresholds for seen and surprise languages separately. Our final system performed the best on 6 out of 9 languages (including two surprise languages) and achieved highly competitive results on the remaining three languages.

Published
2023
Pages
629-637
Proceedings
17th International Workshop on Semantic Evaluation, SemEval 2023 - Proceedings of the Workshop
Conference
The 61st Annual Meeting of the Association for Computational Linguistics, Toronto, CA
ISBN
978-1-959429-99-9
Publisher
Association for Computational Linguistics
Place
Toronto, CA
DOI
10.18653/v1/2023.semeval-1.86
EID Scopus
2-s2.0-85168849823
BibTeX 
@INPROCEEDINGS{FITPUB12968,
   author = "Timo Hrom\'{a}dka and Timotej Smole\v{n} and Tom\'{a}\v{s} Remi\v{s} and Branislav Pecher and Ivan Srba",
   title = "KInITVeraAI at SemEval-2023 Task 3: Simple yet Powerful Multilingual Fine-Tuning for Persuasion Techniques Detection",
   pages = "629--637",
   booktitle = "17th International Workshop on Semantic Evaluation, SemEval 2023 - Proceedings of the Workshop",
   year = 2023,
   location = "Toronto, CA",
   publisher = "Association for Computational Linguistics",
   ISBN = "978-1-959429-99-9",
   doi = "10.18653/v1/2023.semeval-1.86",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/12968"
}
Back to top