For the first time, Hamburg University of Technology (TUHH) offered the course Machine Learning I as a three-week Autumn School. The new format was initiated by Prof. Nihat Ay. Together with Dr. Frank Röder, he gave students the opportunity to complete the entire bachelor-level course outside the regular semester in a compact format, including lectures, group sessions and the final exam.
From September 14 to October 2, participants focused intensively on the fundamentals and methods of machine learning. The three-week block combined classroom teaching with independent study and exam preparation, concluding with the final module examination. The Autumn School was limited to 40 participants. Due to high demand, additional students were placed on a waiting list.
The course Machine Learning I covers fundamental and advanced machine learning methods, including discrete und continuous neuronal networks, functional analytic methods, universality of feed-forward networks, the perceptron algorithm, gradient methods, stochastic gradient descent and the backpropagation algorithm. Students applied these methods to specific problems, evaluated trained models and worked both independently and in teams.
The new format offered additional flexibility in structuring studies. As the Autumn School took place during the lecture-free period, students could consciously choose to dedicate three weeks to an intensive focus on machine learning. The bachelor-level course was also opened to students who had come to TUHH for their master’s studies and wanted to acquire or refresh foundational knowledge in machine learning.
Organizing Machine Learning I as a block course also required new approaches to planning and coordination. The format could not simply be integrated into existing structures and required close coordination beyond regular course planning. In addition to Prof. Ay and the Institute for Data Science Foundations the Center for Teaching and Learning (ZLL) as well as the Examination Office and the Program Management Team contributed to the successful implementation.
School of Electrical Engineering, Computer Science and Mathematics
