Artificial Intelligence Career

[Anonymous No.3823829]

Hey /sci/,

I want to work with A.I. and now have the following options in college majors:

Computer Science
Computational Mathematics
Mathematical Physics
Mathematics

I personally lean towards Computer Science since I already have 4 years of experience in that field. But I heard A.I. is more math-based so mathematics might be the better choice.

What do you say?

[Anonymous No.3823875]

bump

[Anonymous No.3823908]

different sides of the same coin

captcha: Ph.D. yargis

[Anonymous No.3824013]

>>3823908
yeah, that's why it's so hard to decide :(

[Anonymous No.3824030]

I'm not really sure what Mathematical Physics would be used for in AI, but I've only not just glanced at the definition of it. Probably any of them would be alright, but it depends on what kind of AI you're planning to work on. I'd recommend not becoming too dependent on any particular method, and to make sure to learn at least something related to neuromorphic stuff, but that presumes you want to work on general AI. It's a wide enough field that I have no idea what you're trying to go after unless you describe it in more detail.

[Anonymous No.3824046]

>>3823829
> so I herd A.I. is math-based

No, you heard that AI is based on fucking around with AI concepts but not ever EVER being bold enough to just build AI cores and submitting them to testing. Any "science" that becomes "math-based" is a dead end, ruled by PhD welfare cases who value career safety over everything else.

30 years after Racter and Eliza, we have precisely DICK for conversational AI, and our computational and storage power has increased by at least 3 orders of magnitude. BUY A CLUE ABOUT WHAT'S GOING ON.

With AI of any scope, the three major Jewish Professions would be put out of work (like automation put the laboring goyim out of work): Lawyers, Doctors and Teachers. So those professions have conspired to put the kibosh on true AI development.

[Anonymous No.3824069]

>>3824030
Oh, I'm not too sure myself, I'm more or less asking which one of those would make me more valuable.

[Anonymous No.3824115]

>>3824046
Are you saying that A.I. won't develop any further?

[Anonymous No.3824223]

>>3824069

Machine ethics sounds promising then, as it's quite the problem now and I'm pretty sure it'll remain one for some time. If you really want to do something valuable, though, I'd look for an area of knowledge that's being taken for granted. Research things you're curious about until you find something without enough research to sate that curiosity, then make that research.

In any case, consider reading Creation: Life and How to Make it and/or spending a lot of time playing one of the games int he Creatures series. BEAM robotics is also worth a look just because it's a good demonstration of the usefulness of simple feedback loops. I can probably talk enough about places to learn about AI to make a whole thread on the subject, but I might as well throw this out there too in case you're good at linear algebra and have enough time to take a free online class: http://www.ai-class.com/

>>3824046

Up top. I'm pretty sure it's the secretaries doing it, though. Not only do they run the world, but it's obvious that computers are well suited to clerical work. That, and researchers are prone to being territorial about specializations to the point where anything complex enough to make AI is broken up into artificially exclusive sub-fields.

>>3824115

I don't think so. If anything, he's saying AI will develop rapidly once we get past the barriers to actual research on it.

[Anonymous No.3824277]

>>3824115
> Are you saying that A.I. won't develop any further?

Correct. How can AI develop when the infrastructure to do so, has already been corrupted by people who refuse to allow it to advance?

The point is that by now, we should have had expert systems and chatbots that were 1000 times more capable than those available in 1989. What do we have instead? What we have is a vast desert of expert systems, since those Jewish Professions convulsed and made sure that ES R&D was suppressed from making the threat against their jobs. So today, we have almost zero ability to make use of ES that will suffice as a doctor, lawyer or teacher. In order to get full medical and legal advice, and in order to obtain full instruction, it's NO FUCKING COINCIDENCE that you have to STILL go to a (primarily Jewish) doctor, lawyer or teacher.

And with 1000 times more capacity for intelligent conversation, all we have is Eliza clones. We don't even have a Racter clone (it being too capable at the time, or something). Chatbots died out, and didn't advance. Why? It's because the AI profession REFUSES TO ADVANCE THE STATE OF THE ART. AI dweebs are either being held back externally, or they are themselves afraid to actually create something that can talk with them.

I urge all those who are going into the profession to stop listening to the (primarily Jewish) leaders in their so-called field of knowledge, and just ADVANCE THE STATE OF THE ART. Build BEFORE testing. Stop testing BEFORE building, since that just leads to building NOTHING.

[Anonymous No.3824284]

>>3824046
Um...lawyers are going to be the very last thing to go, like there will literally be machines, mechanics, lawyers and politicians and nothing else, doctors aren't going anywhere until at least 50 years into meaningful AI but probably less, and how the fuck do you figure Jews are teachers?

[Anonymous No.3824299]

Fuck A.I tech.
let me build something to so everything so i dont so nothing while doing what i want when i want= A.I

Enjoy being butpooped by rapist robots you fucking faggots

[Anonymous No.3824313]

>>3824299

Enjoy not being able to spell the word "do". And besides, I personally plan to build a robot that attacks me when I least expect it, like Cato Fong does for Inspector Clouseau.

[Anonymous No.3824326]

>>3824313
AHH you may not have this in mind but if you build something that can learn. What God like abilities do you have to keep it from secrectly learning to out learn you. If it learns it can keep secrects then what will what butt fucked the what outta what.
If you cant keep your shit together..dont shit.

[Anonymous No.3824382]

>>3824326
>hurr dystopian machine ruled world

[Anonymous No.3824936]

OP here, can someone answer my initial question? I really just wanted to know which one of those gets me closer to A.I.

[Anonymous No.3825034]

>>3823829
It really depends on what kind of AI you're aiming for. Soft computing and symbolic AI are completely different worldviews (even though both camps sometimes refuse to accept this).

If you lean towards the symbolic side, you ought to pick CS or logic if you can. I know a few professors who came from linguistics and made their way into AI, too.

If it's the non-symbolic camp (think Rodney Brooks), physics and maths are fine too. CS won't be wrong but it depends on your teachers, your uni and what you make from of it.

[Anonymous No.3825058]

>>3825034
Can you please give some examples for symbolic and non-symbolic A.I:?

[Anonymous No.3825137]

>>3825058
symbolic: Newell & Simon, GPS -- a Program that simulates the Human Mind

non symbolic: Rodney Brooks, Intelligence without Representation (there's a similar named one, I confuse them, better have a look at both)

symbolic: expert systems, logic, case-based reasoning, knowledge-based configuration, qualitative simulation

non symbolic: artificial neural networks, EC, swarm intelligence, often using biological metaphors -- see PPSN proceedings for instance.

[Anonymous No.3825162]

>>3825137
forgot to say: semantic web is another example of symbolic AI from, say, 30 years ago. This kind knowledge representation in general.

[Anonymous No.3825412]

>>3824936
> OP here, can someone answer my initial question? I really just wanted to know which one of those gets me closer to A.I.

I already answered your question, in the implied meta form. If you're not going to actually BUILD an AI, then it doesn't matter HOW you approach the field. If you're not going to actually perform AI research, then it doesn't matter what path you take. Your effectiveness will be ZERO, hence your trip will be POINTLESS.

[Anonymous No.3825489]

Fact is that funding in fields that are called "AI" is dead because of AI winter. Funding is full of polotics and how you can get funding agencies to believe you etc.

Fields that are funded well right now is things like machine learning, pattern recognition, etc. I am in computer vision right now, and there is LOTS of funding in that field.

Clearly the reason is that funding agencies stopped believing in AI or because only someone builds something previously called AI, it stops becomes AI (I guess because the mechanics are already known.) So in the end the word AI becomes avoided.

I have no idea what "symbolics" mean so I am just going to say that preparing for a field like pattern recognition requires you to know lots of linear algebra. I would advice you to take Computational Mathematics as an undergrad degree because it is going to make it a LOT easier to learn pattern recognition. You are also going to be prepared to actually code the stuff.

For grad. school, I would say go for either CS or EE because that's where all the groups who are interested in AI are. Of course they never call it AI. It's always robotics or whatever. And ofcourse choose the group that you will go to wisely. If you are even remotely interested you would want to start reading papers by junior year, just to explore, see how you like it, etc.