CLEF CheckThat! · 2024

Fired_from_NLP at CheckThat! 2024: Estimating the Check-Worthiness of Tweets Using a Fine-tuned Transformer-based Approach

Paper Details

Authors: Md. Sajid Alam Chowdhury, Anik Mahmud Shanto, Mostak Mahmud Chowdhury, Hasan Murad, Udoy Das.

Venue: CLEF 2024 Working Notes, CEUR Workshop Proceedings, Vol. 3740, pages 387-395, Grenoble, France, 2024.

Status: Published.

Abstract

Due to the immense use of web-based and social media platforms, people encounter large amounts of information, but not all of it is true. It is therefore important to verify statements before believing them, making check-worthiness a core NLP research topic in both low-resource and resource-rich languages. The CheckThat! Lab at CLEF 2024 organized Task 1 on check-worthiness estimation, providing Arabic, English, and Dutch datasets to determine whether claims in tweets and transcriptions are worth fact-checking. We used several machine learning, deep learning, and transformer-based models to identify the best approach for the task. Our proposed CW-BERT model ranked 7th, 10th, and 12th, with F1 scores of 0.530, 0.543, and 0.745 for Arabic, English, and Dutch respectively.

BibTeX

@inproceedings{chowdhury2024checkthat,
  title = {Fired_from_NLP at CheckThat! 2024: Estimating the Check-Worthiness of Tweets Using a Fine-tuned Transformer-based Approach},
  author = {Chowdhury, Md. Sajid Alam and Shanto, Anik Mahmud and Chowdhury, Mostak Mahmud and Murad, Hasan and Das, Udoy},
  booktitle = {CLEF 2024 Working Notes},
  series = {CEUR Workshop Proceedings},
  volume = {3740},
  pages = {387--395},
  year = {2024},
  url = {https://ceur-ws.org/Vol-3740/paper-34.pdf}
}