I just finished up my undergrad and am starting a PhD in comp vision/machine learning this fall. I don't have as much math maturity as I'd like.

Do you recommend taking math classes early on or classes such as computer vision/nlp/robotics early on to get a better idea for my research focus?

  • 1
    Can you contact the instructors for the CS classes you want to take? If so, you could ask about math classes that will actually help you in the CS classes. Commented Jul 23, 2017 at 21:31

1 Answer 1


Based on vague phrases such as "[not] as much math maturity as I'd like", you may get a spectrum of advice on this site. You should ask people who know you and your background and an adviser in the PhD program.

Presumably, if your background (in math or other subject matter) is lacking, your new adviser will have noticed and will have recommendations how to remedy that.

Issues to consider:

(1) There are probably some CS and ML courses you will need to take at the beginning in order to be on track to finish your degree during the time the department is prepared to support you. My guess is that these will not have heavy math requirements.

(2) You will not know everything you need to know even when you finish your PhD, much less at the start. Learning much of what you need to know is a lifelong adventure.

(3) Your primary objective should be to find an interesting and do-able thesis topic as soon as possible, to finish a high quality thesis on schedule, and to get your degree.

(4) A thesis requires original research. If you know exactly what you're doing, how long it will take, or how much support you will need, then the project is not research. So you need to anticipate uncertainty and delay. Accordingly, you should not build any avoidable delays into your program.

(5) All other things being equal (which they never are), you might consider looking for a thesis adviser who is in the middle of his or her career, and heavily engaged in pathrbeadking work. Assistant professors may be so frantic trying to do their own research to get tenure that they have little time to worry about yours. Some senior professors are not entirely up to date, and may suggest topics that would have been more appropriate five or ten years ago. A really promising adviser may have great advice what courses to take early in the PhD program

(6) Most people who get into PhD programs are from the top of their undergraduate classes. They cannot all be at the top of their PhD class. There may even be some areas in which you are near the bottom of your PhD class and you can hope for some in which you are near the top. You may have as much to learn from some of your classmates as from some of your professors.

(7) Soon there may be other Answers disagreeing with everything I have said. Maybe they will be right. But these are some issues worth considering.

Not the answer you're looking for? Browse other questions tagged .