We will mainly follow two existing online courses, namely, John Canny's (UC Berkeley) CS294-129 Designing, Visualizing and Understanding Deep Neural Networks and Fei-Fei Li's (Stanford) CS231n: Convolutional Neural Networks for Visual Recognition. That is you will watch the next video lecture prior to our lecture, prepare a list of 5 questions as main homework for the video you have watched, and then we will try to answer your question in class as a group together. We will also present our own stuff without videos. We may even ask you to read papers. Next to the “5-questions” homework, there might also be written exercises and you can propose and run a deep learning project.

We aim to have the discussions in person, however will also upload the discussions from last year to allow for people to watch online that cannot make it to the in person discussion.

Concerning the exercise sessions, we will try the same principle, i.e. upload the exercise sessions from last year, but will have the exercise session in person. If the majority of students wishes for a complete online version of the exercise sessions, we can definitely adapt.