Alright Veeky Forums, I'm turning 19 in a few days, and I know I want to work in AI, that's a given. I've taken a year out between secondary school and uni (where I'll be studying mathematics and computer science). I know right now, my best course of action would be to make myself a really strong mathematician if I want to stand any chance in the future, but I do also want to become versed in the theory of AI during that time, too. Can you recommend any resources (particularly research papers or blogs. Pop sci books just sell you dreams) that are accessible?
I know it's a very wide field, and is often misunderstood, not least of all by people like me. If it's simply a bad idea right now, it would be nice to know too.
Are there any easy to set up, very basic neural network programs available?
Im thinking like that greentext about the Doom playing AIs
Logan Clark
Not sure if I'd be able to do much OP, but since we both share the same interest I'll share with you what I'm doing. These will all help improve your data science. Also, I recommend you get familiar with Numpy and SciPy (if you choose python, from there you can get into sckit-learn and Tensorflow. However, if you're going with R, then the kaggle link can help) datacamp.com/home dataquest.io/dashboard kaggle.com/learn/overview
Anaconda (if you want to use Python) will help you organize your libraries and what not. Also, i suggest getting Jupyter Notebook as well anaconda.org/
Best of luck to you mate, and just one thing to keep in mind. To learn this stuff will take some major time and dedication. This isn't something you can understand in a matter of hours; so don't give up too early!
Jayden Williams
This is OP. I'm not too sure about the US levels of math, but I've done up to multivariable calculus and partial derivatives, a bit of abstract algebra too (all formally taught)
There probably are, but I've found Veeky Forums to have pretty direct and honest answers to questions like these
Alexander Foster
You beautiful bastard. Will check all of these links. I completely understand, it'll take time like all good things.
Christopher Anderson
>I know right now, my best course of action would be to make myself a really strong mathematician You're right about that. Get the mathematical preliminaries out of the way before you start. Sure you can "start" sooner, but you'll just be following tutorials and have very little understanding of it.
Connor Ross
Aww, thanks; and make sure to have fun along the way my friend!
Aaron Young
AI is just a meme. Forget about it, you will be happier.
Daniel Baker
real OP here
I've only actually done Calc I
Mason Foster
You great big phony.
I'm not in the business of being happy, it seems b
Jeremiah Gonzalez
Read
Lang's Basic Mathematics Lang's A First Course in Calculus Hubbard and Hubbard's Vector Calculus, Linear Algebra and Differential Forms: A Unified Approach Introduction to Probability by Bertsekas and Tsitsiklis Probability in Electrical Engineering & Computer Science: An Application-Driven Course by Walrand
Why do you say acaemia is a shitshow. Applying to PhD programs this year.
Daniel Thomas
Computer Science degree with a specialization in Machine Learning, then get a Master's in Machine Learning
Don't let anyone tell you otherwise on Veeky Forums - you get paid more to write ML software than to actually train the models because there are few people who have that good intersection, and they are very different skills
Sebastian Anderson
Think long and hard, after looking through others' experiences if academia is what you want to do for the rest of your life. PhDs are, usually, expensive and stable employment is less guaranteed than if you were a regular B.S. Your success is at the mercy of your advisor, and a bad one can make your life real hell. Politics is the name of the game (almost as much as actual politics). That's the only way you're getting funding. And the people fucking suck. There's a saying that the more plenty the resource, the less brutal the man. Well, funding and recognition is Sahara scarce in academia, and the pettiness is word-class.
Pay is shit too. You'll always be making more money in the private sector. Make sure you also think long and hard if you want to stay in academia after your program is over, because once you make a decision, it's hell going back. Companies won't hire someone who's been in academia for longer than a year. Students also suck, bringing home your work sucks. You're not gonna be a patrician going to dinner parties and spending your free time lost in thought.
Something something hours are shit, something something proposal after proposal, something something publish or perish. This is all for the U.S, however. I know it's much better in places like Sweden, Norway, and, especially, Singapore.
Jonathan Fisher
Expense isn't somethong I need to worry about, since I'll only join if I get RA, which mostly I will. The rest of your points are valid though. Though I always had this perception that politics was much brutal in the corporate world. Profs always seem more open and easy going to me, from my experience thus far.
Ryder Johnson
Politics is everywhere. But some would say the academics are fake and don't acknowledge, while the corporate world does.
Carson Sanders
I'm from England. I think I'll end up working my ass off on a problem I have a good chance of not solving/advancing, while being paid peanuts despite the fact I could leave uni after my undergrad and get a quite well paid job. And I literally can't image anything else. Been working in industry during my gap year now and it's taking all my will power not to an hero.
I recommend that you do both strong aswell as weak AI. Also learn a few programming languages. C++, Python are a must but C# will make your life a lot easier and JAVA will open up more job opportunities.
Start with creating boids then try writing A* yourself from scratch. Always recommend doing writing your own ANN.
This will give you some grasp of how different AI's work.
Then after that don't I I really mean this don't reinvent the wheel. That means build on libraries that are already being maintained by large groups of programmers. And your all ready set to go to contribute to AI. Godspeed!
Parker Hall
Buddy, I'm as old as you are and also interested in AI, though for a longer time.
First of all, learn programming, I recommend Python + Numpy. Secondly, you will need to know some math, especially multivariate calculus and some linear algebra. Thirdly, go watch Andrew Ng's course on Machine Learning and implement your own NNs from scratch.
(I'm on the third step.)
Ian Diaz
Cool, how long have you been interested in it? I've done the things you mentioned in the first two steps, so I'll give Andrew ng's course a try. Does it have exercises along with the lectures?
Did a module on machine learning at uni last term and this seems like a pretty good list. If you've just finished school, make sure your statistics is up to scratch (S1/2/3 stuff) and learn some linear algebra. SciPy is easy to quickly implement some algorithms.
Ian Brown
AI seems interesting but I cant motivate myself into making tools without objective, what cool problems could be solved with machine learning?
Ryder Mitchell
>just go read these boring ass math textbooks
lmao?
Nathaniel Ramirez
>particularly research papers or blogs. Pop sci books just sell you dreams your supposed to start with textbooks
Nathan Martin
One of the most interesting books I've read is the Thesaurus. Learn 'word association'. Who knows what word you'll end up on, then try to beat that record!