Trying to learn ML

I'm a guy trying to learn Machine Learning all by myself and I went to college to get a useless degree in physics. Fast forward to today, I understand how the python syntax works and some algorithms but what the fuck does discrete math have to do with it? I was looking at some CS engineering curriculum and most of them mention about graph theory and group theory, I don't understand why I would ever need that? Can someone tell me if this is important to know?

Attached: Capture.png (755x618, 51K)

Other urls found in this thread:

mml-book.github.io,
twitter.com/NSFWRedditImage

u dont, mostly these topics matter:
mml-book.github.io,

>why is graph theory needed to understand algorithms built on graph theory?
Gee, I wonder.

Most large companies use graph related methods to determine relationships between different entities.

Group theory / finite fields form the mathematical basis for most data security, integrity and cryptography applications.

You should learn ML, any by that I mean a meta language.
Maybe SML, maybe CakeML. Maybe even OCaml.

Attached: IMG_20190607_154237_437.jpg (898x902, 158K)

>what the fuck does discrete math have to do with it?
Machine learning is all about math.

Teutonic Tongues on Jow Forums???

If you think math has no place in computer science, then you are literally too retarded to do programming. Go back to your video games, dumbass.

And by the way, coding TensorFlow in Python doesn't count as writing AI. You're USING AI, you're not WRITING AI. If you were WRITING AI, you would be using something like C++, Lisp, or Prolog. Using TensorFlow means you're tacking together AI programs that other people wrote and calling it AI programming.

hey whatsup phys major, i just finished a useful phys undergrad degree, too yayayay

csci is fundamentally about counting. therefore discrete math forms it's logical mathematical basis.

I don't know much about machine learning, but I think it's basically AI which is just algorithms. so if you wanna know how the algorithms actually work, you should probably be comfortable with some discrete math. coming from a physics background, discrete math should be pretty simple. What I would do is take an actual algorithms course at the upper division level (if you have the time, money, inclination, etc) and learn the discrete math as you go along, or in the appendix or something.

I ended up taking more csci courses than a normal phys major does cuz one of the instructors in the csci dept where I'm at is a goddamn joy to learn from.

But yeah, coming from physics it's weird having to do discrete math stuff cuz you really don't ever interact with that world in physics

>useful phys
meant to say useless, as well. OOPS

Have you heard of data structures?

bump

focus on programming, it's easier to get a coding job than a ml job and it's easy to turn that experience into ml later

>to get a useless degree in physics
>doesn't know group theory

Why do you lie on the internet?

Coursera, motherfucker, do you know it?

And after that Udacity.

Daily reminder that no one would hire a nigger without a proper degree for a reputable tech school.

>math

I am thinking OP is specifically wondering why the fuck discrete math, though, not just math in general.

Coming from a physics background, having to learn Discrete Math-y stuff is like learning downy math, so they are probably are wondering if it's totally necessary to pick up an entire additional course on the subject vs learn via application

im not OP tho so i dunno

Discrete math is important for recursion, graphs, counting, and proving if something is correct or even computable. All of those are central problems in CS and appear often in programming.

right right, I get that but I am just trying to explain what OP's thought process might be as someone who is coming from a physics background

Discrete math is taught at University to build a framework to think through CS problems. In higher level algorithms courses, you'll use discrete math type proofs for analysis. It also is the building block for cryptography, usually you'll be introduced to OTP, RSA, Galois Fields which all use modular arithmetic and are proved using discrete math frameworks.