Is cognitive science a smart option for somebody wanting to work in artificial intelligence...

Is cognitive science a smart option for somebody wanting to work in artificial intelligence? Any cognitive science majors care to weigh in on there experiences in this field? It seems to blend a lot of fields together that I'm interested in!

My idea would be studying cognitive science and then a masters program in artificial intelligence

Not to get things mixed up. As far as i know at my uni cognitive science is rather about complex systems and chaotic behaviour therefore still trying to understand how the brain works. Since psychology is about statistics i would consider taking an approach from informatics. And ofc cognitive sciense doesnt have to be the same at every uni so at the right place this could also be the way to go.

No. Go for CS or statistics.

Care to extrapolate on your reasoning? To me it seems like within my career/lifetime that the computer science/technology for AI will be in place but people who can help translate that into structured thought and human-like cognition will be more important.

Cognitive science is already pretty heavily steeped in statistical models and computer science already.

It also leaves me open for further study in biological neuroscience, which is another field that will be important going forward when computer to brain interfaces become prevalent.

There are four main approaches in AI. The goal is to get a computer to either
1) think humanlike
2) act humanlike
3) think rationally
4) act rationally

Currently pretty much any notable or serious person in AI works in column 4. Some techniques are "biologically inspired" but most are formulated as information/optimization/probabilistic problems.

I think there are a handful of people in cognitive science are doing things like trying to simulate a neuron or brain, but we really don't know enough about the brain for that to effective at this point. If you want to advance AI go into CS stuff or go into neuroscience and increase our understanding of the brain.

The most advanced AI algorithm out there right now is heavily inspired by the brain.
And I'm not talking about deep learning

Which "most advanced" algorithm?

HTM

I'm not a cognitive science major, but the answer is no.

For example, look at the authors of the white papers behind alpha go. They pretty much exclusively are computer scientists with PhDs.

The exception being the CEO who has a degree in neuroscience and a degree in computer science.

I really wouldn't call that advanced, especially not in a general sense. it's covered under deep learning, and it's analogous to ANN which is not at all biologically inspired.

not sure where you are reading that, but it's completely wrong

Before we go any further what's your background?

CS

HTM is not analogous to deep learning really at all. There are some specialized ANN architectures that emulate similar capabilities in the temporal aspect, but the actual computation is completely different.

It's still a machine learning algorithm that deals heavily with abstract representations. It's really not that big in the field.

Look into it. Trust me.

It's the real deal

Oh, I like this idea! If OP was to try this route, he/she would need to make sure to take electives and gain experience aligning with the requirements and goals of the graduate program though.

>it's analogous to ANN which is not at all biologically inspired

Wrong.

Go as quantitative as possible. The general rule is more quantitative people I. Etc. Math comp sci physical find it easy to transfer to other fields I. Etc. Bio psych etc compared to the other way round. It's definitely an advantage with modelling in neuroscience. Or a.i. the leading computational neuroscientists often have math (and on one notable occasion, physics) degrees and I expect anyone in a.i will be math or compsci. Cognitive science has quantitive stuff in it but the bsc wont be as indepth as others. I also found that in cog neuroscience and psych or stuff like that, youre not really going to need anything you learned in bsc except naybe methods and youre probably better doing your own indepth reading on a topic if u want to do research rather than carrying over from something you learnt in bsc.

In this boat with you too OP. Am a senior in in high school waiting to hear back from colleges. I'm leaning towards computational neuroscience or just doing cs with a psych minor, then doing a strict AI plan in grad school.

What the hell are you talking about?

half of your cog sci curriculum will be bullshit like psych and linguistics

better to do CS and then take an appropriate minor or some grad courses and a senior thesis

Do CS. Don't do some non-thorough hours d'ouerves degree. Then pick your concentration in grad school.

Nobody in the cognitive science field has any precedence in the a.I. community. The devs just use robots to cuck themselves with.

user is right
study computer science and or mathematics while reading cognitive science stuff in your free time

I would say that a better option is go math/stats and then masters in cs.

I think cognitive science actually has a lot of ideas that mainstream AI/Deep Learning researchers will eventually need to take into account.

But as for a undergraduate degree cognitive science is a bad idea. At most schools its basically just a psychology degree with a few more technical classes sprinkled in. In order to stand a chance in the current landscape of AI, you really need the mathematical background from a math/cs/engineering degree.

A great math/cs background, and then doing a PhD in cogsci... that's when things get interesting. Look at Joshua Tenenbaums work for example.

Basically use your undergraduate to become as comfortable and familiar with math as possible. You can always learn the other stuff, but building a mathematical mind becomes harder as you age.

If you want to do a masters program in AI then you have to have very strong foundations in linear algebra, probability theory & stats, logic, general computer science.