Arabic argument recognition improved with new sequence tagging system
TTLab at Daleel 2026: STAR-Ar, Sequence Tagging for Argument Recognition in Arabic
Computation and LanguageArtificial IntelligenceMachine Learning
Summary
Understanding arguments in Arabic texts is hard because there aren't many tools for it. The authors built STAR-Ar, a system that detects and classifies argumentative parts of debate and editorial texts by looking at every word and its context. They combined a transformer model called BERT with other methods to recognize argument spans accurately. Their system scored around 73% in tests and found that training on debates worked better than on editorials, mainly because editorials had less data.
What this means in practice
- •For arabic nlp developers: Build better tools to automatically identify and classify arguments in Arabic debate and editorial texts using token-level tagging.
- •For social media analysis teams: Automatically detect argumentative parts in Arabic user discussions to improve content moderation and sentiment analysis.
Authors
Bhuvanesh Verma, Ali Abusaleh, Alexander Mehler
Abstract
Argument Mining (AM) is a critical NLP task that remains significantly under-resourced in Arabic. This paper presents $\testtt{STAR-Ar}$, a BERT-BiLSTM-CRF architecture for argument discourse detection and classification, as our system for Daleel 2026, the inaugural Arabic argument mining shared task. The task requires the identification and classification of argumentative discourse units (ADUs) in debate and editorial texts.We jointly model these two objectives as a token-level sequence labeling task using a BERT-BiLSTM-CRF architecture that combines contextual transformer embeddings with structural transition constraints to support accurate span detection. $\testtt{STAR-Ar}$ achieves an F1-score of 72.69 on validation and 73.7 on test data. Our domain-specific analysis shows that models trained exclusively on editorials underperform those trained on debates, a disparity we primarily attribute to the smaller size of the editorial dataset. The code for $\testtt{STAR-Ar}$ is available at ${\href{https://github.com/ENTAILab/daleel_2026_Arabic-Argumentative-Discourse-Mining}{\faGithub~TTLab at Daleel 2026}}$