What do I do if I really only figured out what I want to research just after applying

Question

I recently finished a round of PhD applications.

I always knew I wanted to do research, but when I applied I didn't know what sub-field of Computer Science I wanted to do my research in. I was already out of school a year when I decided that I may never figure it out just by sitting around, so I just bit the bullet and applied to a PhD program with what I knew.

Naturally, a month or two after I applied, I realized that what I really wanted to go into was robotics, which I have very little experience in at the moment and which I didn't mention in my personal statement.

I'm pretty good at teaching myself things. Give me another few months and I'll probably be able to learn quite a bit about the subject, and I'd bet anything I could put together some decent projects that could make me impressive enough to get into a good robotics research track, but I've already applied to a bunch of schools and it's at the point where if I'm getting in, I'm not getting in for anything robotics related.

So, here are the options I see in front of me:

1. Turn down the offers I get, try again next year when I can be more focused and well-researched on which programs/professors I want. This feels like giving up, and I'm always nervous about putting things off, especially for a year.

2. One of the schools I got into happens to have a good program for robotics, but I doubt my application was on any of those professor's radar. I could contact professors from that school and ask if they'd want to work with me, but I don't know what I'd say with so little robotics knowledge. Another school I applied to but haven't heard back from has an even better program for it, so if I am emailing professors, there's still time to email professors from that school as well.

3. Go to a school I got into, and just hope I'll be able to switch programs.

I know I ultimately have to decide for myself what to do, but any advice is greatly appreciated.

Answer 1

I saw this sort of thinking/discussion all the time. Something like, I'm a second year Ph.D. student but I want to start another one, because X is big and I'm excited about it. This is a bad way of thinking about a Ph.D. program.

Here is the thing:

One of the main pitfall of PhD students is that they can't get their mind into one problem.

Here is another fact:

During Ph.D. you learn how to think and read/write as a researcher, regardless what you are working on.

Advice: I say choose one among all the offers you get. Drop all, unless you have a pretty good damn reason for it (e.g., got a job at NASA).

Answering 'what I want to do for research' always changes! Thats the nature of research. You can not keep up with all trends during Ph.D. First, you need to think, read, and write as a researcher; and you learn them during a Ph.D. Then, you can follow the trends as a postdoc or an academic.

Answer 2

You've probably already made your decision, but in case anyone else stumbles upon this:

I always knew I wanted to go into academia. I finally found something I loved (physical geography, specifically climatology), and jumped in. Some time in my 2nd year (& as an MS student, I only should've had 2 years), I realized that I'd been focusing on academia, but not what I actually enjoy. I got through; I learned a lot about my subject, critical thinking, networking, writing and myself. I'm so glad I went, but I could've done a better job if I was more aware of the options within my discipline (I'd only had 2 courses when I applied), and especially if I'd known the culture of my subject.

What I'm trying to say is, if you know you can teach yourself pretty well and make some impressive projects, you're not "just sitting around." You are exploring your field. In the meantime, you should research faculty, read some papers, and figure out what you want to dedicate your life to for 4+ years. Because a PhD program isn't just about building robots: You're learning the culture, writing papers, creating posters, networking, writing grants, applying for fellowships, teaching in the classroom and mentoring/teaching in your lab...If you love your subject and your projects, it will be worthwhile, though sometimes you will not think so. If you discover, once you really start digging in, that you don't actually want to sleep, eat, and breathe robotics, all the rest of being a PhD student can be miserable.

That's just grad school in general. Specifically about your question:

There are more things to consider than which school and which program. Grad programs usually accept only 5-20 students per year. In my experience, more than 10 is a big cohort. You may be one of the top applicants this year, but next year you may be outshone. So there is definitely some merit to accepting whatever you get. You never know - you may find something else in CS that you get really excited about. If you're certain you're interested in robotics, see if there is someone in engineering or physics or wherever the robotics people are, who will be your outside faculty member. They may even take on more of your project than your official PI/chair, they could even be a co-PI.

Grad school isn't like undergrad (pretty much in every way except that you're on a campus) - if you want to change departments you have to reapply. If you have to change depts, go to the other dept's journal clubs, get into a lab & be their best student, make yourself invaluable. In your new letter, emphasize your interdisciplinary background and how that will benefit your studies and the department in general. Impress the PI of the lab you've been working in & they'll write a great letter of rec, appeal to the selection committee on your behalf, go to bat for you against admin. It's possible to change, but there are significant obstacles.

Good luck! I'm curious to know what you went with, when you make your decision.

Answer 3

Unlike Dave Rose, I absolutely think you should pursue your interest in robotics if you find you truly enjoy it. You can (and should) figure that out right now by spending a long weekend with an arduino.

However, I really don't see much need to reconsider your applications, unless any offers you receive are incredibly specific. Many CS PhD programs don't require (or allow) you to work in a particular group or with a particular prof until you've completed the first year or so of coursework, but if a program accepts you with the explicit understanding that you'll work with X on NOT_ROBOTICS, then you might want to decline.