Personal notes for Computational Intelligence Lab (CIL) 2020 by Prof. Thomas Hofmann at ETH Zurich
By Yu Fei
Motivated by some friends of mine, I wrote my personal lecture notes for this course which
- cover all the stuff (I believe) on the slides;
- add important remarks (I think) mentioned by Prof. Thomas Hofmann during the lectures but not on the slides;
- connect things in each chapter together with motivations and detailed discussions (if they are correct) in a plain language;
- add some supplementary materials (also reorganized with my own words) that I think are useful to get a deeper insight into some points.
The materials are mainly based on
- lecture slides
- lecture recordings (2020, 2019)
- exercises, solutions and tutorial slides (2020)
- piazza discussions (2020)
- Pattern Recognition and Machine Learning (Bishop 2006)
- original papers of the algorithms mentioned in the lectures
- other courses I took (mainly mathematics of data science (2020) by Prof. Afonso at ETH Zurich).
This course covers many interesting topics, but sometimes it goes too fast to get all the things just by looking at the slides and watching the recordings. This motivates me to try to organize things together with extra materials fulfilling the details. As a result, I feel I gain a much deeper understanding than before. To be honest, I would say either writing things like this or reading it might not be the most useful thing for preparing for the final exam. However, I believe it will provide some useful insights that are helpful to understand the fruitful contents prepared by Prof. Hofmann and the TAs during the semester. It is also a good way, from my point of view, to make this course better by accumulating meaningful discussions, comments, or thoughts by students in each year, since students tend to encounter similar problems through the attempt towards understanding. This is also another reason for me to write this stuff before I forget them, and I tried to make everything in an easily editable way to allow later amendments.
I am totally aware that this note is far far from perfection or even being able to be called a well-written material. I wrote this note with an average speed of 0.5 to 1 chapter per day, and my English writing skills are not that good. As a result, there are many typos and grammar mistakes, and thanks to my friend Lixin and my classmate Guillaume Wang many of them have been fixed. I don't attempt to create a substitute for the lecture materials, but if it can help someone learn this course better, that's all that matters.