Recent advancements in machine learning methods based on deep neural networks have led to performance breakthroughs in many fields, especially within transportation systems and autonomous vehicles. This class will explore how these technologies are revolutionizing the fields of robotics and transportation, equipping students with the knowledge to contribute to the advancement of smart navigation systems and the development of autonomous vehicles. This class will not only go over the principles of deep learning but also provide students with an understanding of the most recent research in the field. The first part of the class will begin with a high-level introduction to neural models.After that, the fundamental strategies and cutting-edge methods for constructing, training, and visually depicting different types of deep learning models to:Develop object, detection, recognition, localization, and segmentation models.Understand the basics of attention mechanisms in deep learning.Overcome the "lack of training data" problem using zero and few-shot learning techniques.Through hands-on projects, participants will also learn to apply deep learning techniques to real-world transportation challenges, preparing them for the future of intelligent mobility and robotic autonomy.

The sessions will combine theoretical background, seminars about recent papers and mathematical modeling of deep learning architectures  focusing on convolutionAl neural networks.

Introduction to machine learningShallow Neural Networks And Deep Neural Networks - OverviewBackpropagation algorithm - OverviewConvolutional Neural NetworksSemantic segmentation (U-NET)Transfer learningObject localization (YOLO)Semantic embedding (Overview)Zero shot learning Few shot Learning (Siamese Neural Network)Attention + transformers Application in Robotics and Intelligent Transportation                      

Applied Statistics + calculus + coding skills in python 

Python Machine Learning, 3rd edition. By Raschka & Mirjalili

Format:An oral exam using a digital platform (e.g., BBB via Moodle) or a written one.Scheduling:Students will be given a specific time slot at least one week before the exam date.It is crucial for students to check their technical setup before the exam day to avoid any technical difficulties.

Review course materials, including lecture notes, readings, and assignments.Think critically about the topics and anticipate potential counterarguments or challenges.

Depth of Understanding: Demonstrated thorough comprehension of course content.Ability to connect course concepts to real-world applications or personal experiences.