If you are motivated by machine learning, I highly suggest taking this class with professor Kassab. I found her willingness to answer questions extremely helpful. She acknowledges the rapidly evolving nature of machine learning, as well as the textbook's dense approach, so lectures offer a chance to cover additional context, such as building the intuition behind theories, calculations omitted from the textbook, (good) hand drawn visuals, examples of modern applications, connections to recent advancements, best practices, and guiding heuristics. She carries a wealth of information and is clearly passionate about the field, so her lectures always ended up feeling "too short." I certainly feel that I got my money's worth out of this class, and I highly disagree with the previous reviewer's perspective. They are a great example of how cheating is only cheating yourself. For the intrigued, these lectures will provide a tremendous opportunity for you to build a solid foundation in machine learning.

My third time having Kassab and since I had her for math 118 and 164, this was the second class I've taken at UCLA where i never attended lecture or discussion (first time being my third Miles Chen class lol his youtube vids are the best). Here's my pro-con breakdown after finishing the final an hour ago and not having received my grade yet:

Pros:
- lecture slides are just watered down textbook notes. she leaves some slides blank where she does stuff like MLE by hand, but the derivations are all in the textbook and really not that hard if you've taken stats100b.
- everything is chatgpt-able. HWs, blank slides in lecture notes, even the project is chatgpt-able. The material isn't necessarily simple, but it's basic ML that's very well documented or basic linear algebra hence GPT-4 can solve pretty much every problem in the course. I also used chatgpt to help me study by explaining concepts, especially since i never went to class lol.
- exams are very conceptual and not too mathy. It's technically a math class, but realistically you'll never need to perform linear regression or SVM or train neural networks by hand irl. what's useful irl is understanding how everything works so you know when/where to use different techniques. Exams reflect this philosophy as the math problems are fairly easy to solve and the rest are conceptual.
- exam topic list. a week before the final, she sent out a list of things to study, things not to study, and things you wouldn't have to memorize. She didn't do this for the other classes i took of hers, so she might've just been feeling generous lol.

cons:
- a lot of kids in my TA's discord complained about how the exams covered so little of the material and they were too conceptual. If you're the type of person who does a lot of practice problems and derivations, the conceptual elements of her exams might be annoying. I recommend truly understanding the concepts (like watch a visualization on YouTube or something) before doing any practice problems.
- lectures are useless. i always felt like her lectures were just a waste of time. i tested my hypothesis by not going to a single class and cramming all the materials from after the midterm to the final in 3 days (excluding topics she specified weren't going to be on the final) by reading thru her lecture slides and previous hws. I finished the final within 2 hours and think i did pretty well (def passed) hence i believe this is a viable strategy for those afflicted with senioritis.

EDIT: Got an A-, not too shabby for a chronic absentee