Conference proceeding
SmolLab_SEU at MAHED Shared Task: Do Arabic-Native Encoders Surpass Multilingual Models in Detecting the Nuances of Hope, Hate, and Emotion?
PROCEEDINGS OF THE THIRD ARABIC NATURAL LANGUAGE PROCESSING CONFERENCE: SHARED TASKS, pp.658-664
Third Arabic Natural Language Processing Conference: Shared Tasks (Suzhou, China, 08/11/2025–09/11/2025)
11/2025
Abstract
The dynamic interplay of hope and hate speech on Arabic social media presents a critical challenge for content moderation and digital discourse analysis. This paper presents our systems for the MAHED 2025 shared task on Multimodal Detection of Hope and Hate Emotions in Arabic Content, addressing the two text-based subtasks. Our approach centers on a systematic, empirical comparison of Arabic-native versus large-scale multilingual Transformer encoders to determine the optimal pretraining strategy for this nuanced domain. Comprehensive evaluations demonstrate the clear superiority of Arabic-native models, with our ARBERTv2-based system achieving the highest performance. We secured 11th place in Sub-task 1 with a macro F1-score of 0.682 and 5th place in Subtask 2 with a macro F1-score of 0.514. Error analysis reveals persistent challenges in interpreting implicit language and overcoming severe class imbalance, particularly in distinguishing targeted hate from general offensiveness. This work contributes a robust benchmark for this comparison and underscores the importance of language-specific pre-training for nuanced affective computing in Arabic.
Details
- Title
- SmolLab_SEU at MAHED Shared Task: Do Arabic-Native Encoders Surpass Multilingual Models in Detecting the Nuances of Hope, Hate, and Emotion?
- Authors/Creators
- Md. Abdur Rahman - United International UniversityMd Sabbir Dewan - Murdoch Univ, Sch IT, Murdoch, AustraliaMd. Tofael Ahmed Bhuiyan - University of DhakaMd. Ashiqur Rahman - Augusta University
- Publication Details
- PROCEEDINGS OF THE THIRD ARABIC NATURAL LANGUAGE PROCESSING CONFERENCE: SHARED TASKS, pp.658-664
- Conference
- Third Arabic Natural Language Processing Conference: Shared Tasks (Suzhou, China, 08/11/2025–09/11/2025)
- Publisher
- Assoc Computational Linguistics-Acl
- Number of pages
- 7
- Grant note
- Southeast University, Bangladesh
- Identifiers
- 991005925418107891
- Murdoch Affiliation
- School of Information Technology
- Language
- English
- Resource Type
- Conference proceeding
Metrics
1 Record Views