Featured Research

Research & Publications

My research focuses on advanced AI architectures, multimodal learning, model efficiency, and trustworthy AI systems. I explore how deep learning models can be designed, optimized, and evaluated for complex real-world applications.

My work spans transformer architectures, computer vision, natural language processing, adaptive transfer learning, and distributed AI systems, with applications in healthcare, accessibility, and sign language translation.

Featured Research

ADAT: Adaptive Transformer Architecture

Developed a time-series-aware Transformer architecture integrating convolutional feature extraction, LogSparse self-attention, and adaptive gating to capture short- and long-range temporal dependencies.
Evaluated across three sign language datasets, ADAT reduced training time by an average of 21% for sign-to-gloss-to-text translation while improving translation quality against evaluated baselines.

Scientific Reports, 2026

Read Publication

HATL: Hierarchical Adaptive Transfer Learning

Introduced an adaptive transfer learning framework that progressively unfreezes pretrained model layers based on training performance, combining dynamic unfreezing, layer-wise learning rate decay, and training stability mechanisms.
Evaluated across multiple sign language translation datasets and model architectures to investigate adaptation and generalization in low-resource learning environments.

arXiv · Preprint, 2026

Read Publication

Trustworthy AI: Edge, Cloud & Blockchain

Proposed and evaluated a privacy-preserving AI framework integrating edge computing, cloud-based model operations, and blockchain-based auditability for sign language translation.
The framework incorporates consent management, data governance, and model lifecycle considerations alongside evaluation of translation quality and system performance.

Mathematics, 2025

Read Publication

VisioSLR: Computer Vision & Model Evaluation

Developed and evaluated a computer vision framework for sign language recognition using fine-tuned YOLO architectures and transfer learning.
The study reported a mean average precision of 97.4% and recognition time of 11 milliseconds using YOLOv8m, demonstrating the performance of the evaluated approach.

Procedia Computer Science, 2025

Read Publication

Journal Publications

Nada Shahin and Leila Ismail (2026). HATL: Hierarchical Adaptive-Transfer Learning Framework for Sign Language Machine Translation. [Under Review].
DOI:10.48550/arXiv.2603.19260

Nada Shahin and Leila Ismail (2026). ADAT novel time-series-aware adaptive transformer architecture for sign language translation. Scientific Reports (Top 10%).
DOI:10.1038/s41598-026-36293-9

Nada Shahin and Leila Ismail (2025). Towards Trustworthy Sign Language Translation System: A Privacy-Preserving Edge–Cloud–Blockchain Approach. Mathematics (Top 8%).
DOI:10.3390/math13233759

Nada Shahin & Leila Ismail (2024). From rule-based models to deep learning transformers architectures for natural language processing and sign language translation systems: survey, taxonomy, and performance evaluation. Artificial Intelligence Review (Top 1%). DOI:10.1007/s10462-024-10895-z

Leila Ismail, Nada Shahin, Huned Materwala, Alain Hennebelle & Lea Frermann (2023). ML-NLPEmot: Machine Learning-Natural Language Processing Event-Based Emotion Detection Proactive Framework Addressing Mental Health. IEEE Access (Top 8%). DOI:10.1109/ACCESS.2023.3343121

Nada Shahin & Haneen Al Arfaj (2022). Learning technologies among academics in the United Arab Emirates: One academic year since COVID-19. International Journal of Information and Education Technology. DOI:10.18178/ijiet.2022.12.2.1597

Mohamed Watfa, Ahmed Abdelrehim, Nada Shahin & Kamal Jaafar (2021). A Structural Equation Model to assess the impact of sustainability management on the success of construction projects. International Journal of Construction Management (Top 2%). DOI:10.1080/15623599.2021.1998302

Nada Shahin & Mohamed Watfa (2020). Deaf and hard of hearing in the United Arab Emirates interacting with Alexa, an intelligent personal assistant. Technology and Disability. DOI:10.3233/tad-200286

Conference Papers

Leila Ismail, Nada Shahin, Henok Tesfaye & Alain Hennebelle. VisioSLR: A Vision Data-Driven Framework for Sign Language Video Recognition and Performance Evaluation on Fine-Tuned YOLO Models. The 16th International Conference on Ambient Systems, Networks and Technologies. April 22-24, 2025, Patras, Greece. DOI:10.1016/j.procs.2025.03.014

Nada Shahin & Leila Ismail. GLoT: A Novel Gated-Logarithmic Transformer for Efficient Sign Language Translation. IEEE Future Networks World Forum. 15-17 October 2024, Dubai, UAE. DOI:10.1109/FNWF63303.2024.11028758

Nada Shahin & Leila Ismail. ChatGPT, Let Us Chat Sign Language: Experiments, Architectural Elements, Challenges and Research Directions. The 10th International Symposium on Networks, Computers and Communications. 23-26 October 2023, Doha, Qatar. DOI:10.1109/ISNCC58260.2023.10323974