Anyone else a machine learning researcher? For such a popular

Do you listen to Data Skeptic, OP? It would be directly up your alley

hi OP, care to give an outline of what maths are relevant to the subject? Or what maths should be known to grasp some measure of subtlety in the field?

>>8371899
Buy Nvidia cards. We'll talk after that.

One of Fei-Fei Lee's (Stanford prof) recent grad students (Andrej Karpathy) got multiple job offers upon graduating a few for more than a million dollars a year: Deep/Machine Learning they are just throwing money at right now.

>>8372000
Go on libgen and get a book on the Tensor Flow library https://bleedingedgepress.com/tensor-flow-for-machine-intelligence/ this one is good. It gives a crash course in the math required and you can go from there (or just look up university calendars)

>>8372000
Linear algebra is a must. It's basically the vocabulary of machine learning. I would say learn as much as you can in linear algebra.

Calculus to the point where you are comfortable with vector calculus. Jacobians, Hessians etc. You rarely see super advanced calc stuff but it's good to be familiar.

Probability and stats. Honestly this is my weakest area. Still learning so can't give solid advice. But it plays a huge role in machine learning.

From there it really depends. If you working with audio you'll probably want to learn some audio relevant maths, like FFT and signal analysis stuff. Working with graphics/images requires another set of knowledge.

I would say in terms of a degree, an applied mathematics degree would probably be the most relevant. If I could go back in time I would have majors in applied math instead.

Why are so many ML researchers SJWs?

>>8372027
because math is waycist
http://money.cnn.com/2016/09/06/technology/weapons-of-math-destruction/index.html

Being a first year CS undergrad and compiling Caffe does not make you an ML researcher. Nobody cares about your latest "Deep Meme" project.

>>8373023
>Nobody cares about your latest "Deep Meme" project
I care.

>>8372027
It's not just a ML researcher thing, it's a general academic thing. Actually, it's a general "educated people" thing.

>>8373103
Only for olds (5-20 years ago). Most of the really bright younger undergrads I know (sophomores and freshman) are shitlords who like Milo posts and Trump compilations on Facebook. And bright kids usually predict cultural trends berore the normals get them.

( ‿ ) leftism is so passé

>>8373155
>Most of the really bright younger undergrads I know (sophomores and freshman) are shitlords who like Milo posts and Trump compilations on Facebook.
I guess it varies, but in my experience it's always the opposite. I have to agree with >>8373103 here, academia and the upcoming new academics look even more left-leaning overall than before. Studying in the UK here, and every prominent uni I've visited it's mostly the same. Speaking for natural science though, not computer science, I don't know if it's any different in that field.

>>8373163
Oh, I'm talking about east coast muricans so maybe it is different elsewhere. Its interesting that it's not just white kids, either--the most prolific shitlords ive seen from the young math/cs cohort at my alma mater are all two Americans with Indian parents and a Pakistani raised in a ghetto area of the states.

File: a-new-kind-of-science.jpg (6KB, 200x265px) Image search: [Google]

6KB, 200x265px

>>8373023
>he thinks I do "deep learning"

>>8371899
I figure no one talks about it because almost everyone here is an undergrad, or about to start their undergrad.
I'm using machine learning to discover biomarkers in diseases/disorders like alzheimers, depression, schizophrenia, addiction, etc... People have been using machine learning for this kind of stuff for years, but we still haven't discovered any robust biomarkers. I'm hoping my project will work out alright due to the fact that I have huge amounts of data to work with; unlike most studies.

>>8372000
>>8372021

This. Honestly, it only comes down to maybe second year undergraduate mathematics. You could go a little further and look at functional analysis for optimisation.

>>8373103
>>8373163
It's only a subset of Generation Y that are SJWs. It was only towards the end of my undergrad that I saw SJW nonsense actually start to become mainstream. Now I'm beginning to see it fade out of existence. It's only going to be a small number of people born within a 5-year window who are cucked SJW faggots.

>>8372000
Literally no maths are not relevant.

>>8374117
Fluid mechanics.

>>8374121
If you study this you'll have something interesting to think about while your machine learning algorithms give no results.