YMAI at SemEval-2022 Task 5: Detecting Misogyny in Memes using VisualBERT and MMBT MultiModal Pre-trained Models
Mohammad Habash, Yahya Daqour, Malak Abdullah, Mahmoud Al-Ayyoub
Abstract
This paper presents a deep learning system that contends at SemEval-2022 Task 5. The goal is to detect the existence of misogynous memes in sub-task A. At the same time, the advanced multi-label sub-task B categorizes the misogyny of misogynous memes into one of four types: stereotype, shaming, objectification, and violence. The Ensemble technique has been used for three multi-modal deep learning models: two MMBT models and VisualBERT. Our proposed system ranked 17 place out of 83 participant teams with an F1-score of 0.722 in sub-task A, which shows a significant performance improvement over the baseline model’s F1-score of 0.65.- Anthology ID:
- 2022.semeval-1.108
- Volume:
- Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)
- Month:
- July
- Year:
- 2022
- Address:
- Seattle, United States
- Venue:
- SemEval
- SIGs:
- SIGLEX | SIGSEM
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 780–784
- Language:
- URL:
- https://aclanthology.org/2022.semeval-1.108
- DOI:
- 10.18653/v1/2022.semeval-1.108
- Cite (ACL):
- Mohammad Habash, Yahya Daqour, Malak Abdullah, and Mahmoud Al-Ayyoub. 2022. YMAI at SemEval-2022 Task 5: Detecting Misogyny in Memes using VisualBERT and MMBT MultiModal Pre-trained Models. In Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022), pages 780–784, Seattle, United States. Association for Computational Linguistics.
- Cite (Informal):
- YMAI at SemEval-2022 Task 5: Detecting Misogyny in Memes using VisualBERT and MMBT MultiModal Pre-trained Models (Habash et al., SemEval 2022)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2022.semeval-1.108.pdf