IEEE Mobile Section Hybrid Technical Meeting
Deep Learning: Overview, Applications and Optimization
Dr. Mohamed Shaban
ECE, University of South Alabama
Deep learning (DL) has emerged as a transformative paradigm within artificial intelligence, driving significant advancements across diverse sectors, including computer vision, natural language processing, healthcare, and autonomous systems. By utilizing hierarchical, multi-layer artificial neural networks—such as CNNs, RNNs, and Transformers—DL models possess the capability to automatically extract complex patterns and high-level features directly from raw, large-scale data. This presentation explores the foundational techniques of deep learning, highlighting its applications in critical domains such as medical image analysis. Despite their high predictive accuracy, the significant computational, memory, and energy requirements of modern deep learning models restrict their deployment on resource-constrained edge devices. To address this, the presentation surveys various complexity reduction techniques—including model pruning, quantization, low-rank approximation, and knowledge distillation—designed to accelerate training and inference speeds. It will also be shown that these optimization approaches can substantially reduce model size and computational operations (FLOPs) while maintaining acceptable performance levels.
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
Loading virtual attendance info...
- University of South Alabama
- 150 Student Services Dr.
- Mobile, Alabama
- United States 36688
- Building: Shelby Hall
- Room Number: 2119
- Contact Event Host
- Co-sponsored by Mobile chapter of IEEE
Speakers
Mohamed Shaban
Deep Learning: Overview, Applications and Optimization
Deep learning (DL) has emerged as a transformative paradigm within artificial intelligence, driving significant advancements across diverse sectors, including computer vision, natural language processing, healthcare, and autonomous systems. By utilizing hierarchical, multi-layer artificial neural networks—such as CNNs, RNNs, and Transformers—DL models possess the capability to automatically extract complex patterns and high-level features directly from raw, large-scale data. This presentation explores the foundational techniques of deep learning, highlighting its applications in critical domains such as medical image analysis. Despite their high predictive accuracy, the significant computational, memory, and energy requirements of modern deep learning models restrict their deployment on resource-constrained edge devices. To address this, the presentation surveys various complexity reduction techniques—including model pruning, quantization, low-rank approximation, and knowledge distillation—designed to accelerate training and inference speeds. It will also be shown that these optimization approaches can substantially reduce model size and computational operations (FLOPs) while maintaining acceptable performance levels.
Biography:
Dr. Mohamed Shaban is the director of the Computer Vision Lab. He is currently an Assistant Professor in the Electrical, and Computer Engineering department at the University of South Alabama (USA). He also has a joint appointment in the department of Pathology at the Whiddon College of Medicine, USA. He has previously served as an Assistant Professor of Computer Science at Southern Arkansas University, Graduate Teaching Assistant at the University of Louisiana at Lafayette and an Assistant Lecturer at Mansoura University.
Dr. Shaban has received the Ph.D., and M.S. degrees in Computer Engineering from the University of Louisiana at Lafayette in 2016, and 2012 respectively. He has also received the M.S. degree in Electrical Communications Engineering, and the B.S. degree (Excellent with Honors Degree) in Electronics, and Communications Engineering from Mansoura University, Egypt in 2010, and 2006 respectively.
His current research interests are in the fields of Signal, and Image Processing for Biomedical Applications, Machine, and Deep Learning Applications, Edge AI Optimization and Time-Series Analysis.
Email: