Advanced Deep Learning for Visual Understanding: Representation, Compression, and Explainability

This course offers an introduction to advanced deep learning methods for visual data, combining foundational topics with emerging research directions. Students will explore core concepts in computer vision alongside cutting-edge techniques such as knowledge distillation. Emphasis will be placed on representation learning, model compression strategies, and explainable AI (XAI), addressing the dual needs for efficient and interpretable models. Through lectures and practical examples, students will gain insights into building deep learning models that are not only accurate but also compact, explainable, and better aligned with the structure of real-world visual information.