U4RASD at KnowledgeGraphEval-2026: Ontology-Constrained Encoder and LLM Systems for Arabic Relation Extraction
Quick Answer
This paper shows that The U4RASD team achieved a top ranking in Arabic relation extraction at KnowledgeGraphEval 2026, utilizing an ensemble of Yehia-RE, Qwen-RE, and TypedPair models, achieving a micro-F1 score of 94.06.
Quick Take
Their approach combined encoder systems with ontology-constrained , demonstrating the effectiveness of constrained prediction in this task.
Key Points
- Achieved a micro-F1 score of 94.06, ranking 1st in Subtask 2.
- Utilized an ensemble of Yehia-RE, Qwen-RE, and TypedPair models.
- Incorporated typed entity markers and separate heads for relation tasks.
- Fine-tuned LLMs with QLoRA for improved prediction accuracy.
- Highlighted the significance of constrained prediction in Arabic relation extraction.
Paper Resources
📖 Reader Mode
~2 min readAbstract:This paper details our submission to KnowledgeGraphEval 2026 Subtask 2 on directed Arabic relation extraction. We explored encoder-based and large language model approaches for predicting one of 40 relations or no_relation from an Arabic sentence and a subject-object pair. Our encoder systems use typed entity markers and separate relation-existence and relation-classification heads, while our LLM systems fine-tune Yehia and Qwen3.5 with QLoRA and ontology-constrained prediction. Our best system, an ensemble of Yehia-RE, Qwen-RE, and TypedPair, achieved a micro-F1 score of 94.06 and ranked 1st in Subtask 2. Yehia-RE (full-vocabulary) achieved 92.98, while the best encoder and Qwen-RE achieved 92.69 and 91.94 respectively. Our results highlight the importance of constrained prediction and complementary modeling approaches for Arabic relation extraction.
| Comments: | 20 pages. System description paper for the KnowledgeGraphEval 2026 Shared Task at ArabicNLP 2026 |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.12742 [cs.CL] |
| (or arXiv:2610.12742v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.12742 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Ahmad Abbas [view email]
[v1]
Thu, 8 Oct 2026 21:41:29 UTC (429 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →The "10th Juror": Open-Set Standpoint Screening for Bureaucratic Bias Detection
MARS-Gov introduces a framework for detecting bureaucratic bias in Dutch government documents, achieving a new state-of-the-art F1 score of 0.880. This model outperforms existing zero-shot detectors by 20.2 points and reduces unnecessary interventions to just 2.5%. The framework's dynamic '10th juror' adapts to emerging biases, enhancing legal language processing.