In today's information-rich era, learners have access to abundant educational resources, but the lack of practice materials tailored to these resources presents a significant challenge. This project addresses that gap by developing a multi-modal question generation system that can automatically generate diverse question types from various content formats. The system features four major components: multi-modal input handling, question generation, reinforcement learning from human feedback (RLHF), and an end-to-end interactive interface. This project lays the foundation for automated, scalable, and intelligent question generation, carefully balancing resource efficiency, robust functionality and a smooth user experience.
@article{arxiv.2509.03535,
title = {QuesGenie: Intelligent Multimodal Question Generation},
author = {Ahmed Mubarak and Amna Ahmed and Amira Nasser and Aya Mohamed and Fares El-Sadek and Mohammed Ahmed and Ahmed Salah and Youssef Sobhy},
journal= {arXiv preprint arXiv:2509.03535},
year = {2025}
}
Comments
7 pages, 8 figures, 12 tables. Supervised by Dr. Ahmed Salah and TA Youssef Sobhy