If I want to work on artificial intelligence...

If I want to work on artificial intelligence, is Computer Science the right major or should I choose something more specific?

get a BS in compsci and masters in something more specific.

Do something like compsci or applied maths and do a degree in philosophy on the side.

Companies and labs will get thousands of application from random pajeet compsci majors (possibly even with better grades than you), but a degree in philosophy will set you apart; and knowledge of philosophy is, according to a lot of people who work in AI, necessary to understand what you'll actually be doing.

Depending on what exactly you want to do in A.I. , I guess cog sci with the right specifications could work too.

Forget philosophy. What you want for machine learning is coding experience, solid knowledge of statistics and linear algebra, and skill in proving—so take some real maths at a high level too.

The actual major you put all this under probably doesn't matter that much, so long as it doesn't restrict your courses in a bad way.

>coding experience
>coding
neck yourself my man

>coding experience
LMAO

If you want to go to grad school in CS, then CS is the worst major you can pick. Major in math, statistics, EE, and/or CpE.

Huh? Care to explain?

Artificial intelligence requires a whole fucking bunch of code to be written, so with a CompSci major you'll probably find people who'll pay you to sit at a desk and write and implement the code they thought and planned out. Though that probably won't be fulfilling, and there's low chances you'll actually understand the theory behind the A.I you'll be droning about to build. If you want to "work on A.I. ", as in actually do research, develop it, and not be alienated by it, you're going to have to do what said at some point. Cog sci or philosophy will be necessary in creating new breakthroughs in AI.
However, that post said to do philosophy ON THE SIDE, and I could not stress this enough. Acquire a solid understanding of what fucking computer science is, and how a computer solves problems before going into philomemes.

Computer science has nothing to do with software, that would be software engineering.

I don't see how the distinction is relevant to my post. Generally speaking, when working in ML, you're going to actually have to implement some things, or you wont be able to test jack. So you want some experience in coding. That's all I've said on the matter. It shouldn't be controversial.

You said "coding" which is usually said by web designers or people with no programming experience

lmao
only fa/g/gots are this concerned about muh proper nomenclature. time to go back to the

most AI are just a dressed-up optimization or statistics problem.

you can go for math or EE as well as CS

the nomenclature says a lot about you though. "coding" is to programming what "building" is to architecture

Don't do pure compsci. It's notoriously horrible. They're pretty much vocational courses churning out code monkeys.

At least do CompEng or CS&Engeneering. It's more rigorous.

I found the (you) that can talk good game but can't program his way out of a wet tissue.

cs will gook ya up

Computer Science, Mathematics, Electronics, Physics and Philosophy

even if this were true, so what? programming is a monkey's job

One of my friends is a mechatronics engineer and has been doing post grad work on machine learning and such.

He gets pissy when I tell him that a philosopher will lead to way bigger contributions to the whole thing than some random coder.

>He gets pissy when I tell him that a philosopher will lead to way bigger contributions to the whole thing than some random coder.
the ultimate red pill

Philosophers are useless. Don't insult the man by implying that they could do his job

>Electronics

Applied Mathematics and/or Computer Science. Just focus on "machine learning" (AKA: analysis, Linear Algebra, Probability, Statistics and Optimization - applied to making something learn and develop without being explicitly programmed).

So much bad advice in this thread.
If you want to do AI then compsci is your bread and butter. Maths to Physics is what compsci is to AI. Philosophy is something that comes out of character and thinking about things as you progress and understand your topic more.

>He gets pissy when I tell him that a philosopher will lead to way bigger contributions to the whole thing than some random coder.
and rightly so.
What is a philosopher? A generalist and (usually) an uneducated retard, and no, knowing a lot of old pseudo-scientific books, that amount to hyper-ideological formalisation of commonsense, doesn't constitute being educated.
What is a "random coder"? Someone who can program decently (I hope), but also potentially someone who knows relevant Mathematics and Logic.

So what we've got is a generalist with no understanding of anything, nor any technical ability and a programmer with potential "things".

For AI:
Theoretical CS > Math > Philosophy

Bunch of retards.

>"""AI""" is just optimization and statistics
ftfy.

As a PhD student in a research lab doing literal artificial intelligence research: computer science is good, pure math would also be good (if not better) as long as you took a respectable amount of CS but not necessarily things like OS, networking, etc.

Also mitchell's machine learning book

Yes, computer science is the right major might want a Masters if you are serious though

Lol. You should try telling him that a philosopher will have way more insight into artificial intelligence and its ramifications and ethical problems than some lowly engineering monkey.

Why not software engineering?

>ramifications and ethical problems

Give us a fucking century or two. ML papers today: "Hurr durr this bag of numbers never figured out how to tell cats from rabbits before because we've been propagating gradients like a fucking idiot." It's 2017.

Programming, Statistics, Linear Algebra, Functional Analysis, one of half a dozen areas of pure maths. All these things will be 100 times as helpful to an aspiring ML researcher as Philosophy.

FYI I have a paper getting accepted into ICML and I say 'coding' all the time. My supervisor does too. Probably because neither of us've taken a CS or SWEN paper in our lives.

>do a degree in philosophy

>I have a paper getting accepted

This is some odd phrasing. Has your paper actually been accepted, or is this just speculation?

To simplify, acceptance is basically a majority vote from your reviewers. A while from now, someone will finish looking through all the reviews and decide exactly what papers are getting accepted and what ones aren't. If your paper has rejects or only weak accepts, then at this point you don't know if it's in yet. Otherwise, you basically do.

CS studies algorithms, algorithms apply to anything that processes, anything that thinks and can process, anything that can be solved, anything that can be computed or done by algorithm or through a countable number of sets

that's CS, it's not games or just math

i guess youll have to sit around and wait for a monkey to implement your precious ideas. i hope the monkey doesnt fuck up

>Math PHD

Yes. AI aka Machine Learning is mostly CS theory with some math and fucking about with libraries like Tensorflow.

Applied mathematics is the best degree for AI.

If your schools program is good, you will learn enough CS to get basically anything done, and your math education will be much more rigorous.

Artificial intelligente is a meme

Math yes, EE no. EE has absolutely nothing to do with AI.

>potentially someone who knows relevant Mathematics and Logic.

So a philosopher, then.
Either this is bait or you clearly don't know what a philosopher is.

>brainlets need to make "artificial intelligence" because their own intelligence is insufficient

Most important is maths and statistics, but computer science is good too. Many big names in the community have maths or physics background. Philosophy is imo not that appropriate, unless you want to deal with meta-issues and ethical (how to prevent AI from harming humans, implications etc). Personally have combined computer science and neuroscience background, and can say that maths and statistics is what I lack the most. Programming is not very important for theoretical, and should be easy to learn anyway if you can do the rest. It very much depends on what you want to do with AI - general intelligence, pattern recognition, specific machine learning application...

Taking from psychology (I dabble, fight me) intelligence is simply the ability to learn, solve, and adapt.
These are the goals of creating ai.
Philosophers theorize an ai is to have 3 levels-the core (which can't be changed), the values (learned from experience and hard to alter), and working memory (sensory input, communicating, etc)

This is the basic idea of ai, but how to implement it is the hard part. Does anyone have any ideas?

Programming the code is trivial. Anybody can implement the ML algorithms.

What you want to do is to design the reasoning behind it all.

Any engineering will do.
Computer science won't.

> programming code is trivial

If all you're doing is acting as a human compiler for some mathematical formalism that has already been developed then that is true.
Working on areas such as compilers, kernels or protocols or even just simple but stable and reliable software systems for drudge work actually takes some substantial brainpower because your problem domain is not circumscribed and well defined ahead of time.

Well, yes, but talking about ML, you're not going to be interested in doing the actual programming, it's not what the field is about.

So long as you don't need to weld whatever library you're using into a larger system then I guess not.

\thread

>> If I want to work on artificial intelligence, is Computer Science the right major or should I choose something more specific?

Yes, Computer Science is the best choice. Don't listen the meme makers here.

Do Pure Math undergrad and Statistics graduate school.

>EE has absolutely nothing to do with AI.
Nice joke. Who designs the computational pipelines that allow us to actually train the AI, so we can actually apply it to something beyond a publication?

Are there online forums specifically for ai?