Machine Learning

Thoughts on current state of machine learning/AI?
Is it just a meme or can it actually do some good?

not a meme but super autismo genetically engineered whipped-by-parents pajeet mode

what?

It's good for CS majors who need a new field to rot in

If you're reading about it on sites like Hacker Jews and other popsci outlets, it's already a meme at that point.

Just a female assblasted cuz their place in society is getting replaced by a fucking machine

it's a weird meme in the sense that since it's math heavy so normie reddit shit can't get easily into it like they can with webdev

google "cm1k congnistix"

It's a meme in the sense that's it has gained mainstream support and even normalfags are interested in it. But it's actual computer science so no faggot web devs that "self-taught," themselves JavaScript are allowed.

Here's how machine learning works:

there are maybe 10 people with PhDs in the WORLD who really understand this shit and implement new algorithms

these people work for huge companies like google, and do research for top universities

each advancement in AI is a slight improvement in performance over the previous iteration

the smart people create frameworks which simplify their extensive research, so anyone can sort of use it but not as well as they could

once processors are fast enough computers will take over, but in the mean time machine learning is very fucking useful for many different things. i'm not going to go in depth here, figure it out for yourself.

Do you honestly believe that AI and ML have no practical applications?

...

What's a good source if not hacker Jews.

>there are maybe 10 people with PhDs in the WORLD who really understand this shit and implement new algorithms

It's just applied discreet maths

Convex optimization is not really discrete and many learning problems can be formulated as such.

You have no idea what you are talking about.

It's pretty fun, probably a field I want to stay in.

looks pretty cool. Is it a much-used thing?

I'm going to be a CS freshman this year thinking of getting into it. Seems like a good career option plus it does sound fun.

probably better than front end web development

current methods are good, but there is room for improvement.

>Is it just a meme?
It is a branch of mathematics, so if you believe math is all just dank memes, nothing is of value and you might as well kill yourself right now.

>Can it do some good?
Google seems to make a lot of money.
But you can use it for a lot of things (decision making, classification, control, etc)

I feel like they're deliberately making it inaccessible to most people.

dumb nigger

what kind of math do i need to know to get into it?

Look at how fractured the frameworks are.

Also, books do really shit job of explaining which model to use to apply for which types of ML problems.

>It is a branch of mathematics
Then you can consider computer science a branch of mathematics. It sure doesn't feel a whole lot like a branch of mathematics. Most of the time it's just a sophisticated form of intuition, trial and error.

Look at stuff like scikit-learn or tensorflow or tori. It literally couldn't be made more idiot proof. If you have some data ready it takes maybe half an hour to get your first program running. The details are often cumbersome to research as there are hundreds of different algorithms (that by themselves are for the most part actually pretty simple to understand) in use that can also be combined in smart ways etc, but getting into it is simple, fun and kind of rewarding.

Not really

Watch their demos. They have one where it shows how fisherman use one to identify species of fish.

Also, one feature built into it is the ability to run as many in parallel as you want. They sell a board that will mate with itself so you can stack them. They have a "daisy chain" pin for this purpose IIRC