Review of 2025 Spring

Yesterday, the first term at SFU is finished. I took five courses and it was rewarding to do as all of the courses were closely related to my interests. Some of them were already known before, while others were not. However, all of them helped me filling the bumps between what I knew and what I want to do in research.

This webpage is to leave some traces while I’m in school. I’m not sure if I can find time to do this frequently, but I hope this helps someone who has similar interests.

Review of 2025 spring

I was fortunate to have these lectures in this semester, and I recommend taking these by following instructors.

cmpt 410 Machine Learning - Ke Li

This course introduces theoretical background of machine learning. I don’t know other professor’s version of this course but professor Ke Li’s version of this course is very well structured and gave very good insight and backgrounds to go further in the field. The first month mostly used to cover mathematical background like linear algebra, calculus and probability, but I guess it’s better to have them by macm316, math151 (math251) and math240. Two Assignments were very heavy but they translates the lecture into the practical problems to understand topics well, so if you do the assignments properly, you could get good ideas on the final exam.

cmpt 412 computer vision - Andrea Tagliasacchi

This course introduces common computer vision problems as projects, such as object detection, segmentation, and panorama stitching. The lectures covers each topics of the projects well, and also the instructions have some additional materials to cover more. I used colab pro to get enough resources to compute relatively large networks than example, and I recommend it if you want to get a good score in kaggle. Also, always start early. It’s hard to determine how long it would take at the beginning.

cmpt 361 computer vision - Yagiz Aksoy and Jason Peng

This course introduces a part of computer vision and computer graphics. The lecture is most likely a type of seminar, so they were easy to follow. Also, the assignments are related to the each topics and very straightforward (most of students get full marks). However, the midterm and finals were different story. They expected solid understanding the concepts, and sometimes requires the derivation of the concepts with mathematical equations. If you just read the slides and remember some of the concepts there, the exams will betray you. Even though the professors does not explain very clearly sometimes, you have to understand and know the steps of important concepts for the exam.

macm 361 Numerical method - Steven Ruuth

This course introduces computational approach to solve mathematical problems like linear problems, polynomials, and a little bit of ODEs. The lecture was good and have records on youtube to cover mostly same contents. The pace was good to follow, and tutorial session helped me a lot to learn how to solve the problem actually. So, I recommend not to miss the tutorial and lecture, because if you lose the pace, it’s hard to follow up later. Midterm was easy enough but the final was relatively hard, but it scaled very well eventually.

math 308 Linear programming - Maxwell Levit

This course introduces the linear programming, mainly focuses on the simplex algorithms and its dual theory. The concept itself is not difficult but straight-forward, and the lecture was clear. Wordload was reasonable, but the quizes and exams are quite easy, so making mistake could feel like a big negatives. Overall, the course contents were well-structured, grading were fair enough. The theories in the lecture are very basic and old, so might not be useful later on, but it would help at least to understand the background of the optimization problems.