ICAART 2025 -17th International Conference on Agents and Artificial Intelligence

Irfan Ali (University of Palermo) and Amina El Ganadi (University of Modena and Reggio Emilia) participated in the 17th International Conference on Agents and Artificial Intelligence, presenting their research on AI applications in linguistic and textual analysis.

Irfan Ali presented ABBIE: Attention-Based BI-Encoders for Predicting Where to Split Compound Sanskrit Words, a deep learning model based on bi-encoders and multi-head attention that achieves high accuracy in segmenting compound Sanskrit words. The study also introduces a new dataset built from the Digital Corpus of Sanskrit and the University of Hyderabad corpus, showing that ABBIE outperforms previous state-of-the-art methods.

Amina El Ganadi presented Generative AI for Islamic Texts: The EMAN Framework for Mitigating GPT Hallucinations, which enhances AI reliability in Islamic studies by integrating verified sources to reduce misinformation. Her research explores the challenges of applying large language models (LLMs) to Islamic studies, addressing issues such as hallucinations and reference inaccuracies. The proposed EMAN framework enhances model reliability through API-based integration with verified sources like Sahih al-Bukhari. The study highlights how embedding-based methodologies improve accuracy and reduce misinformation in AI-generated religious content.