Arabic relation extraction boosted by combined models and ontology constraints
U4RASD at KnowledgeGraphEval-2026: Ontology-Constrained Encoder and LLM Systems for Arabic Relation Extraction
Computation and Language
Summary
Identifying connections between people, places, or things in sentences is important but challenging, especially in Arabic. The authors developed new methods that combine different types of language models and use rules from a knowledge structure (ontology) to improve accuracy. Their combined approach performed best in a competition, showing that mixing different models and using constraints helps understand language relations better.
What this means in practice
- •For natural language processing developers: Build systems that automatically identify and classify relationships in Arabic text using combined encoder and large language models with ontology constraints.
- •For knowledge graph engineers: Enhance Arabic knowledge graph construction by integrating high-accuracy relation extraction models that use ontology constraints during prediction.
Authors
Ahmad Abbas, Hussein Mazeh, Chadi Abou Chakra, Aya Jouni, Mohammad Ballout, Fadi A. Zaraket
Abstract
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.