Duty

I served as a Teaching Assistant for the exercise sessions. My responsibilities included designing new exercise sheets and continuous-assessment questions, preparing tutorial material (including a PyTorch tutorial notebook), mentoring student project groups, and grading homeworks, the midterm, the project write-ups, and the written exam. I also contributed to preparing new exam material and supported students through Q&A and office-hour style help.

Topics

• Linear and non-linear networks (activations, backpropagation)
• Approximation theory (including universal approximation)
• Complexity theory
• Optimization for deep learning
• Optimization landscape of neural networks
• Architectures (selected modern architectures)
• Neural Tangent Kernel (NTK)
• Regularization and implicit bias
• Generalization bounds (theory and modern phenomena)
• Adversarial examples and robustness
• Reinforcement learning (introductory module)
• Double descent

Weblink

• Course Webpage
• Course Repository (slides, notes, exercises)