My apologies if this is the wrong subreddit.
I am an undergraduate math major who would like to pursue a graduate degree in computer science, but most importantly I would like to study machine learning, and I have been independently for about half a year.
I have 1 more year of undergraduate studies left, and I can't decide which classes will benefit me the most. Next semester, I will be taking advanced applied linear algebra (a grad course) and either intermediate real analysis or chaos/nonlinear dynamics (after a few Google searches, I couldn't figure out if this would be relevant, though I did find this), and cryptography (a separate interest). The next semester I plan on taking topology and stochastic processes.
My question is this: if you could pick any advanced math courses to help with where machine learning is and where it is going, what would you take? Learning math is a lot easier for me when I know that I will be using it eventually. I have been studying measure theory independently with a professor over the last semester, so I am worried that intermediate real analysis will be boring to me (not because it is easy, but because it will be irrelevant to ML) -- are there topics in analysis I should study independently other than measure theory? I haven't encountered enough higher level ML literature/articles that actively involve analysis, even though I am told that it comes up (and it makes sense that it would, given the optimization-based nature of ML and analysis's relation to probability theory).
Anyway, I figure that this thread can serve two purposes: 1) To discuss which math classes are important in ML/higher level ML (besides basic linear algebra, calculus, and probability) and 2) To discuss the importance of analysis in ML
Thanks in advance.
[link][2 comments]