CS 4501

(For Statistical Learning and Graphical Models):

TL;DR - Difficult theory-focused statistical learning class, not for everyone but I personally enjoyed it; definitely hard but not unreasonably so.

Upfront, this is not a CS class really - it's a theory-focused math class. Coming into this you're going to need to be pretty comfortable with linear algebra and especially probability, and having taken a mathematical statistics class like STAT 3120 is probably helpful as well (although less necessary). I largely enjoyed the course but my main criticism of it would be that lacked practical application and contextualization for a lot of the material; I do personally think understanding the theory is cool in and of itself though.

The class is difficult (I'd say comparable to a 4000 or certain 3000s MATH course), there's no getting around that, but it never felt unreasonably so. The TA and professor were helpful on piazza + office hours (and I strongly recommend using both, would not have gotten through the class without those). The midterm (which was the only exam) everyone basically failed, but he gave us reasonable test corrections and extra credit opportunities to make up for it. The psets are pretty time-consuming but doable, generally all of the solutions derive from the notes if you read carefully enough - get a study group too, super useful.

The lectures closely follow his notes (published on his website online which you can check out before taking the course), and are very very thorough. I'd recommend reading the notes before the lecture and using the lecture as a way to clarify. The in-class activities aren't bad at all, just stay on top of things and ask questions if you're confused before getting to class.

(This is for Statistical Learning and Graphical Models)

Please please avoid this class. Not only is it ridiculously worthless, it's insanely difficult.

Grading:
50% Homework: these freaking homeworks took so much time out of my life that I literally can't even. They were assigned roughly once a week, with the occasional off-week. The problems are insane, requiring logic and math that would take a single person an entire semester to solve.
20% In-Class Activities: These were decent, but honest mistakes cost me dearly on some of them.
20% Midterm: Basically a homework assignment, but you have to work on it alone, and you have 2.5 hours to do it. In short, this is where you will lose most of your points. Bring tissues.
10% Project: The professor is ridiculously unclear about what he wants from us, wish me luck.

The Professor: Pretty bad. He "teaches" us by going over his course notes, which is a horrific way to teach a mathematical statistics class, thus rendering the class worthless and boring. He forces us to go to said worthless and boring lectures by giving out participation assignments, which just makes me not like this class even more. He also doesn't seem to recognize the difficulty of this class. For example, one particularly difficult homework got people complaining to him, and his response was "I thought it was pretty easy." He's also condescending, as he mocked the classwork of the class before us. I mean, who does that, let alone a professor?

What you should know: don't take this class if you don't have to. If you do, find a study group oml I would have withdrawn if I didn't find one. Also, you should know probability very well, and I recommend having taken a mathematical statistics class and a linear algebra class beforehand. Good luck :)

#tCFfall2021