Picking an Area of Interest

[garbledy]

Hi everyone,

I'm going to be a Senior in CS next year, and am looking to attend a Ph. D. program for Fall 2011.

Right now, I'm trying to create a list of 8-11 schools to apply to, but I'm having difficulty doing so because I'm not even sure exactly to look for in a school based on my interests. I figured you guys are all very knowledgeable in this and would really appreciate it if you could help me out here!

I am getting a B.S. in CS with a minor in math. At my school, we have to pock a "concentration" - mine is theory of computation (lots of math and logic courses)

I am VERY interested in the following:

- typical theory- and math-related things (logic, automata, graph theory)

- optimizing algorithms based on complexity (as opposed to hardware/architecture/compilers)

- linguistics and NLP (the research I am doing now focuses on this)

I am NOT interested in:

- architecture

- operating systems

- compilers

- robotics

- networks

- databases

- security

- being someone else's code monkey

I am curious about, but don't know if I am interested in:

- cognitive science

- human-computer interaction

- category theory

My questions:

- What schools have good programs in these areas? Any professors in particular I should read more about?

- Are there any more areas that you think I might enjoy based on what I've written above?

And one last thing!

My goal after grad school is to continue working on research, but not in academia. Ideally, I want to be implementing what I've researched as useful software, instead of just publishing papers about it. I'm worried that since my interests are primarily theory related, that I will not be able to fulfill my dream of writing awesome software. Is this the case, or are there applications for high-level research out there?

Sorry for the ridiculously long post! Any help would be greatly appreciated, especially any recommendations on schools or professors I should look into!

Edited August 4, 2010 by meow

[OH YEAH]

I don't know too much about areas other than PL, but here's what I know:

If you're interested in theory:

MIT has a gigantic theory group.

Princeton has a very strong theory group. For instance, they have Sanjeev Arora, who came up with probabilistically checkable proofs, and Robert Tarjan, who invented splay trees and Fibonacci heaps and won the Turing award.

Berkeley also has a very strong theory group, especially if you're interested in game theory--they have Christos Papadimitriou.

NLP: Berkeley, CMU, Brown (Eugene Charniak), Johns Hopkins.

If you're interested in HCI:

CMU (a ton of people), MIT (they have Rob Miller and the Media Lab), Stanford (Scott Klemmer, some others I can't remember of the top of my head), I think GaTech? I'm not sure. I think the University of Maryland has a good program too.

Category theory:

I can actually give you competent advice about this because category theory fits under programming languages and that is my area. No one in the US is really doing category theory anymore. Your best bet would be the University of Pennsylvania (Benjamin Pierce). Some people at CMU might still be doing this as well. I know of a student who is interested in category theory at Princeton, but I don't think anyone is actively doing research there on it. Most of the work in category theory is being done overseas.

As for other areas: you might be interested in programming language semantics or type theory if you like category theory. A ton of groups are working in this area.

Edited August 4, 2010 by OH YEAH

[timuralp]

This is general advice: ask your professors. When I was applying to grad school they had the best knowledge/advice/people to consider at the different schools. I asked 3 different ones and made a list of 8 schools. All in all, pretty painless and by far the best way to go about it.

Also, most students don't have a clear area of focus picked out before coming into grad school and a fair share change advisors, but you might want to at list consider the broader terms: HCI vs Theory vs NLP and pick one of those. I'm not sure what software you're going to write while pursuing Theory of Computation in depth, but those sort of goals are too far removed at this point anyway.

Finally, there is a reason grad student code is grad student code. If you're interested in graduating somewhat on time, you won't have time to make your code "production quality", unless you're writing something very small. When you have a research project that's between 5,000 and 10,000 lines of code, getting it to work with good results is often a challenge in of itself, barring the whole "I want others to use it" aspect.

Edited August 4, 2010 by timuralp

[OH YEAH]

Depends. If you come from a non-research institution and not many of the grads go on to grad school, your professors may not know the best schools for you. Mine did not.

[garbledy]

@OH YEAH Thanks for your reply. It's really helpful, I will definitely look into the professors and departments you mentioned. All the schools you mentioned are ones that I would LOVE to go to.. now it's just a matter if I'm competitive enough to be accepted

@timuralp I asked my advisors, but didn't think to ask more than the two... Now that I have a better idea about what areas to look into, I'll probably contact professors at my school in these areas. Thank you!

On another note - does anyone have any good ways to determine which schools I would be competitive for? I've tried scouring the forums and finding a student with a similar profile for me, but my stats seem to be in the grey area of . I wish there was some sort of database for this type of thing! I know "chance me" threads can get annoying/are generally useless, so I didn't want to post one of those, but at the same time I don't want to waste my time/energy aiming for a school that I have no chance of getting into!

[timuralp]

Grad school admissions are a bit of a crap shoot. For instance, a not so impressive GPA/GRE scores are easily overcome with research experience and a great letter of recommendation. If you get a letter from someone who is either at a school you're applying to or graduated from there, it helps even more (although, you could argue it shouldn't). When I applied, I went through the top-20 schools in my area with the professors and they highlighted the schools I should get into with reasonable certainty, ones where it's a toss up, and ones that were a reach. They also pointed out some schools that were ranked highly but should not be considered for application, because of their admission reputation.

Now, to highlight that it's a crapshoot, at the time I applied to: #2, #4, #5, #7, #11, #13, #14, #15, if I recall correctly, and got into, with full funding, #5, #7, #11, and #14. The problem is that every year, even if you're a great candidate, the professors in your area may or may not be looking for students. Further, if you're not as great of a candidate but somehow few people applied in your field and there is a new faculty or a lot of students graduating, you may find yourself with much better chances at an offer.

Finally, the rankings and even the people, won't tell you which school is best for YOU. But you can figure that out when you visit. Sometimes, after a visit, you're left thinking "Who would I ever work with here? It looked so great on paper!". Don't get set on any one school, visit each one with open mind, see how it feels and you'll figure it out.

Edited August 4, 2010 by timuralp

[explorer-c]

They also pointed out some schools that were ranked highly but should not be considered for application, because of their admission reputation.

Can you explain this statement? I can understand not applying to a highly ranked place because no one works on your area. But I think I've never heard "because of their admission reputation" being used as a reason before.

Now, to highlight that it's a crapshoot, at the time I applied to: #2, #4, #5, #7, #11, #13, #14, #15, if I recall correctly, and got into, with full funding, #5, #7, #11, and #14. The problem is that every year, even if you're a great candidate, the professors in your area may or may not be looking for students. Further, if you're not as great of a candidate but somehow few people applied in your field and there is a new faculty or a lot of students graduating, you may find yourself with much better chances at an offer.

I also attest to this crapshoot-ness of application. I applied to several top 10 schools and some of the results seem a bit random.

[timuralp]

Can you explain this statement? I can understand not applying to a highly ranked place because no one works on your area. But I think I've never heard "because of their admission reputation" being used as a reason before.

"Some places, for instance some unnamed top schools, will admit lots of students with no money. Do not be lured by a big name school that has no money for you. That is a huge mistake." That kind of thing. I didn't apply to one school only because of things related to that and their policies on research, etc.

[BKMD]

I would say, pick an area of interest and stick with it through the application process. You can always change your interests as you work on your degree, but it's good to be focused in your essays. You should have specific goals and specific professors you want to work with. If you want to be able to apply your skills to real software after you graduate (a good goal), areas like HCI/NLP are good choices because these are very applied, and there are plenty of industry jobs and startup opportunities here. I don't know that much about jobs in theory, but something tells me that theory researchers will be less hands-on with the software (though I could be wrong about this).

I agree with the advice that you should talk to your professors about this. It will be easier for them if you have some specific ideas already, but they can hopefully tell you which schools are good in these areas and also which schools you're competitive for.

It takes a lot of time to research the schools you want to apply to and to put together a solid application, so start as soon as you can, since you'll be applying this December (and you'll already be busy with the upcoming semester).

I think everyone is correct that admissions can be a crapshoot, so apply to a number of places. And keep in mind that there are schools ranked below the 'top' schools that still have very good programs and faculty, but often only in certain fields (especially smaller schools), so the key is to figure out which schools are good for you. If you're not too confident about which area you want to go into, make sure whatever schools you pick are big enough that they do research in all of the areas you're interested in.