![[2602.15139] CGRA-DeBERTa Concept Guided Residual Augmentation Transformer for Theologically Islamic Understanding](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.15139] CGRA-DeBERTa Concept Guided Residual Augmentation Transformer for Theologically Islamic Understanding
Summary
The paper presents CGRA-DeBERTa, a novel transformer model designed to enhance question-answering over classical Islamic texts by integrating theological concepts and improving semantic representation.
Why It Matters
This research addresses the challenges in accurately interpreting Islamic texts, which are often complex due to their domain-specific language. By improving QA systems for Hadith, the study contributes to both AI advancements and the accessibility of theological education.
Key Takeaways
- CGRA-DeBERTa outperforms existing models like BERT and DeBERTa in theological QA tasks.
- The model incorporates a Concept Guided Residual Block that enhances semantic understanding using an Islamic Concept Dictionary.
- It achieves a significant improvement in accuracy while maintaining computational efficiency.
- The study emphasizes the importance of domain-specific adaptations in AI models for specialized fields.
- Results indicate potential applications in educational tools for Islamic studies.
Computer Science > Computation and Language arXiv:2602.15139 (cs) [Submitted on 16 Feb 2026] Title:CGRA-DeBERTa Concept Guided Residual Augmentation Transformer for Theologically Islamic Understanding Authors:Tahir Hussain (1), Saddam Hussain Khan (2) ((1) Artificial Intelligence Lab, Department of Computer Systems Engineering, University of Engineering and Applied Sciences (UEAS), Swat, Pakistan (2) Interdisciplinary Research Center for Smart Mobility and Logistics (IRC-SML), King Fahad University of Petroleum and Minerals (KFUPM), Dhahran, Saudi Arabia) View a PDF of the paper titled CGRA-DeBERTa Concept Guided Residual Augmentation Transformer for Theologically Islamic Understanding, by Tahir Hussain (1) and 8 other authors View PDF Abstract:Accurate QA over classical Islamic texts remains challenging due to domain specific semantics, long context dependencies, and concept sensitive reasoning. Therefore, a new CGRA DeBERTa, a concept guided residual domain augmentation transformer framework, is proposed that enhances theological QA over Hadith corpora. The CGRA DeBERTa builds on a customized DeBERTa transformer backbone with lightweight LoRA based adaptations and a residual concept aware gating mechanism. The customized DeBERTa embedding block learns global and positional context, while Concept Guided Residual Blocks incorporate theological priors from a curated Islamic Concept Dictionary of 12 core terms. Moreover, the Concept Gating Mechanism selectively amplifies sem...