When i already have money and a position as an maschine learning engineer. Is there anything gained from me studying (maybe part time) Math at an university?I recently wondered if i become a even better maschine learning engineer with more math. I constantly run into mainly statistical and logical problems where i think>man i bet an actual mathmatican would laugh at me nowi am officially le gifted child meme with an 130+ IQ but i still feel like an absolute retard 99% of the timewill studying math fix this?
Well i started studying math precisely because i wanted to get into machine learning ect in high school and found a lot of math i didnt understant. So yea if you want you can enroll but since you work full time i would suggest learning by yourself. You are not bound by time so you can take your time and you dont really need a lot of math since most of it is theory that builds tools for a wide variety of uses.
>>16386720>Is there anything gained from me studying (maybe part time) Math at an university? Sure, it'll make you better at abstract thinking. After taking a real analysis class I got better at leetcode>wondered if i become a even better maschine learning engineer with more mathMaybe , but it would probably be marginal. Ml is not very math heavy outside of a few specialized areas>like an absolute retard 99% of the time >>will studying math fix this?No, if anything learning math will make you feel more retarded
>>16386790NTA.> Maybe , but it would probably be marginal. Ml is not very math heavy outside of a few specialized areas.This could not be further from the truth. The whole world of probabilistic machine learning is very math heavy, and will continue to become even more math heavy as the modeling catches up with the technology (see the research being done into high dimensional statistics/over-parameterization or normalizing flows if you want an example). If your plan is to use existing ML tools in a "tuning" sort of way you don't need a lot of math. If you want to do novel research involving machine learning algorithm development/architectural development you certainly need a ton of math.
>>16386806>If your plan is to use existing ML tools in a "tuning" sort of way you don't need a lot of mathAnd that's what 99% of "Ml engineers" do. I know one "Ml engineer" who didn't even pass calc 2. Just import pytorch, throw data at a model, and tweak it until it works >If you want to do novel research involving machine learning algorithm developmentAlso a meme for applied stuff. Just take a pre-existing successful model, use a different transfer function or architecture or whatever, and see if it works better >certainly need a ton of math.They don't even write proofs
>>16386720studying math will not fix your feelings of inferior retardation
This thread was moved to >>>/adv/32023044
