LIIR at SemEval-2021 task 6: Detection of Persuasion Techniques In Texts and Images using CLIP features

Erfan Ghadery, Damien Sileo, Marie-Francine Moens


Abstract
We describe our approach for SemEval-2021 task 6 on detection of persuasion techniques in multimodal content (memes). Our system combines pretrained multimodal models (CLIP) and chained classifiers. Also, we propose to enrich the data by a data augmentation technique. Our submission achieves a rank of 8/16 in terms of F1-micro and 9/16 with F1-macro on the test set.
Anthology ID:
2021.semeval-1.139
Volume:
Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021)
Month:
August
Year:
2021
Address:
Online
Editors:
Alexis Palmer, Nathan Schneider, Natalie Schluter, Guy Emerson, Aurelie Herbelot, Xiaodan Zhu
Venue:
SemEval
SIG:
SIGLEX
Publisher:
Association for Computational Linguistics
Note:
Pages:
1015–1019
Language:
URL:
https://aclanthology.org/2021.semeval-1.139
DOI:
10.18653/v1/2021.semeval-1.139
Bibkey:
Cite (ACL):
Erfan Ghadery, Damien Sileo, and Marie-Francine Moens. 2021. LIIR at SemEval-2021 task 6: Detection of Persuasion Techniques In Texts and Images using CLIP features. In Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021), pages 1015–1019, Online. Association for Computational Linguistics.
Cite (Informal):
LIIR at SemEval-2021 task 6: Detection of Persuasion Techniques In Texts and Images using CLIP features (Ghadery et al., SemEval 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.semeval-1.139.pdf