Fall 2022

CS 540 - Introduction to Artificial Intelligence

Instructor: Fred Sala

Requirements: Application/Elective for CS, ML requirement for DS.

Workload: There were 10 HW/Projects and 2 exams. I spent about 4hrs per week on this class

Comment: Unlike what most ppl said, I think this is a great class and not an easy class at all. 540 is a great intro to AI class, that covers a lot of topics, but not much into the depth behind them. Content covered includes supervised/unsupervised learning, KNN/Naive Bayes, Neural Network, CNN, games and search algorithm, and RL. You don't need to know much Math to pass the class, but lots of Math was presented. We had prof Sala, he's by far one of the best prof at explaining at UW-Madison. I really recommend taking this class with him if he is teaching it. He has a really strong Math background and is super friendly. HWs for this class were nicely designed, but I do think it was a bit too easy. Exams can be hard but it was online and open notes when I took it, which makes it easier. This class sparked my interest in ML.

Difficulty: 3.5/5

Recommend: 5/5

CS 532 - Matrix Methods in Machine Learning

Instructor: Robert Nowak

Requirements: Elective for CS, ML requirement for DS.

Workload: There were weekly video lectures(optional), activities, and assignments which were all completion based. There were 3 exams. I spent about 8hrs per week on this class

Comment: This was a really great class. It looks at the Math and theory behind the ML algorithm, especially the linear algebra part. For me, the takeaway was huge and this was by far my favorite class at UW and the one I learned a lot about (though I am slowly forgetting the content lol). We covered linear algebra basic (equations, linear dependence, rank etc), different loss function and regularization methods, SVD, PCA, gradient descent, SVM, kernel method, neural network etc. It goes a lot in depth compared to 540. We had prof Nowak. He's lecture was good and all lecture content and exams were all Math, no coding. Some activities and project involve python coding but it's all completion based. You got to try to do ML algorithms from scratch using NumPy. However, one thing I did not like was that the class uses a half-flipped classroom method which we have lecture on Tuesday and activities on Thursday. Since this was a 8-am class nobody comes on Thursday so the activities just basically becomes a group office hour thing. I felt like it would be more valuable to have a lecture from prof Nowak on the topic on Thursday rather than activities. I strongly recommend this class if you are interested in ML and ML research. This class, just like 540, sparked my interest in ML and also in the theoretical foundations of it.

Difficulty: 4.5/5

Recommend: 4.5/5

CS 320 - Data Science Programming II

Instructor: Tyler Caraza-Harter

Requirements: Required for DS.

Workload: There were weekly quizzes, lab, 7 projects and 3 exams. I spent about 3-5 hrs per week.

Comment: A really good intro to DS class. I heard Tyler designed this class (and 220 as well) and it covered a lot of topics that are useful in real-life DS. He is also really good at teaching and explained all concepts clearly. Topics covered include git, performance, graph, web, visualization, and ML(sklearn). The content is really easy for me considering I finished 400 and also doing 540/532. You will be assigned to groups randomly and IMO it is not useful at all. All 7 projects were nicely designed and I became a better python programer from this class. Def recommend taking this class.

Difficulty: 2/5

Recommend: 5/5

MATH 421 - Theory of Single Variable Calculus

Instructor: Andrew Zimmer

Requirements: Intermediate Math class for Math major

Workload: There were weekly HW and 5 exams. I spent about 6 hrs per week.

Comment: Proof-based Calculus class that teaches you how to write proof. It is like a mini-analysis class and prepares you for Math 521. We covered proof-writing, limits, continuity, sup/inf, derivative, and its application (we skipped integration)mathematical rigorously. If you take it with prof Zimmer, he will be mostly copying his note (on canvas) to the board and you will be copying from the board to your notes. His notes are really good though. HWs are hard but exams are easier. He also gives a lot of practice questions and the exams were very similar. Definitely a good class and I recommend taking it with prof Zimmer.

Difficulty: 3/5

Recommend: 4.5/5

ASIALANG 303 - Fifth Semester Japanese

Instructor: Naomi Geyer, TA- Mitsuhiro Kono

Requirements: Humanities Breadth, also required for Japanese Major and advanced elective for Japanese certificate.

Workload: Attendance is mandatory, there were lots of weekly assignments. There were also a few quizzes, and exams. I spent about 2hrs per week.

Comment: This is the last Japanese class in 1st-5th semester sequence. It unlocks upper-level elective Japanese classes. It is quite hard compared to the 4th semester but gradings were generous. We covered more Kanji (not baby Kanji like 3rd and 4th), advanced grammar, and also lots of articles. There was also a group project at the end and I did not like it (interviewing Japanese backgrounds ppl). I didn't find it hard but I definitely did not enjoy it as before. It is still a good class, and taught very well for Japanese majors. I decided to let this be my last Japanese class at UW (it is just not the same as when Hara sensei teaches it).

Difficulty: 2/5

Recommend: 4/5