2

I am looking to apply for a PhD in the CS realm in the next cycle, particularly something with an AI flavor like Machine Learning or NLP. I am a bit of an anomaly in CS as I have undergraduate majors in a very different area (one was history...). I am finishing up a Masters focusing in signal processing at a well-renowned school and have a good GPA and top GREs. I was a CS minor undergrad, but started too late into college so I couldn't have gotten the major, although I took the core courses.

Although I finished undergrad with plenty of accolades (and I TAed in CS for 3 semesters), and I have excelled in the engineering courses at the Masters level, I am really concerned about my chances to get into a great PhD program. (There are a number of reasons why I want to pursue PhD, but at the same time, *for me*, the time commitment doesn't feel warranted if I'm not working with great faculty at a great institution.)

I have some research experience (an undergrad honor thesis in my majors--I know it's not exactly related--and I'm a co-author on a couple non-CS papers published in IEEE journals from some summer work), but nothing really in depth I've done on my own. My Masters is a non-thesis program so I've had trouble finding an adviser who'll take me on (I want to do a thesis anyway). I've also had trouble finding (and being accepted) to worthwhile summer research opportunities (academic settings related to my interests). I have actually really enjoyed my previous research opportunities, and I know that for PhD admissions it's research, research, research (and some recs). So here are my questions:

  1. How do I find my way into substantive research endeavors?
  2. With my eclectic background, how can I rise above the thousands of CS undergrads with plenty of relevant research during the admission process?
  3. Any suggestions how to sell my academic background as a positive?

NOTE: Of course I think I'm qualified (every applicant does or else they wouldn't apply). I also think that my unique background is a bonus. I'm concerned that those making hiring/admission decisions will feel differently. (edited since first posting)

Thanks in advance for the advice!

Improve this question
4
  • Do you want to do theoretical work in AI/ML or applied? The reason I ask is if its the latter, then you are better off applying to alternative departments like information science departments (like mine). We love people who have interdisciplinary backgrounds but usually, with a computing flavor.
    – Shion
    Feb 11, 2014 at 5:40
  • Thanks for the tip! I didn't think about that... I'm a bit more interested in applied. For example, when I hear of new ideas or approaches, my first thought is always "what could I do with that?". I'd rather be an Edison than an Einstein (although let's be honest I'd be happy with either ;-)). I always felt my biggest strength is thinking creatively. Part of my desire to pursue a PhD is to develop the technical (and procedural) understanding in order to substantiate some of my crazy ideas.
    – marcman
    Feb 11, 2014 at 5:46
  • Then, I think that a HCI/ischool/information science department maybe what you are looking for. At most places, such departments are equal mixes of social science and computing science and your unique background would be welcomed in most programs, especially if you add a dash of qualitative work in your dissertation as well.
    – Shion
    Feb 11, 2014 at 15:03
  • The most important criteria though, is to have a good vision of your potential research question i.e. it is not enough to say that you are interested in AI/ML - but how you are going to use this approach in what research questions.
    – Shion
    Feb 11, 2014 at 15:03

2 Answers 2

Reset to default
2

Unless I am mistaken, "next cycle" is half a year from now. There are many ways to enhance one's PhD prospects, but many big ones are no longer available to you due to timing.

Robert Peters in Getting What You Came For cites an ETS study about how important various parts of the PhD application are considered by committees, on a scale from 1 to 5:

  • 3.9 Undergrad GPA in major field
  • 3.8 Recommendations from faculty known by committee
  • 3.7 Undegrad GPA in last 2 years
  • 3.6 GRE verbal score
  • 3.6 whether undergrad major is related
  • 3.5 Undergrad GPA
  • 3.0 Educational or career aspirations
  • 3.0 Recommendations from faculty unknown by committee
  • 3.0 Whether applicant is known to the committee
  • 2.9 Academic achievements (papers, projects)
  • 2.9 Quality of undergraduate school
  • 2.7 Personal statement
  • 2.7 Interview
  • 2.6 Work experience
  • 2.6 GRE analytical score
  • 2.5 Non-faculty recommendations
  • 2.5 GRE Subject score (related to program)
  • 1.9 Other test scores
  • 1.9 GRE Subject score (related to undergrad major)
  • 1.6 Particular subscores on GRE Subject

You have half a year. Some of these things, like your GPA, obviously cannot be changed. What you can do is:

  • Try to publish or present at a conference
  • Make sure you don't get low grades from any classes you are taking
  • Make sure you get strong recommendations
  • Study for the GRE, especially the verbal part
  • Make contact with faculty at the programs you think of applying to
  • Research thoroughly the programs you are interested in
  • Start working on your Statement of Purpose so you have time to edit it
Improve this answer
2
  • I disagree with many of these criteria as I can easily point out a counterexample i.e. my department. Having said that, I think that the second part of your answer is much more meaningful in this context and should be elaborated upon in detail (and maybe you should consider not having the first part in your answer)
    – Shion
    Feb 11, 2014 at 15:05
  • @shion I think the survey is a good explanation of why I suggest the things I do and don't suggest others. Also, the book discusses OP's exact situation at length, and is worth a mention.
    – Superbest
    Feb 11, 2014 at 16:17
1

This depends in which country or "world" you live in.

PhD programs and similar are very different around the world.

You have:
- U.S.
- Uk
- Countries like Sweden
- Switzerland
- Central Europe like Germany/Austria
- India
- etc.

Each country has a very different culture in general and a different university culture in turn, and also different resources.

Just one example: in India PhD positions are hard to get, in Germany/Austria there are sometimes 5 applications for a single job, so just by chance it is easy to get a position.

In research in technical disciplines (computer science, electrical engineering), my option is that the most important thing is mathematics. It is the hardest and most gerneral applicable topic. This is also, in my eyes, the most important topic for applied research.

If you don't know mathematics it is hard to do functional programming. You need Category theory to use Monads, Functors, etc., to write programs in functional languages like haskell. There are state of the art programming languages based on Higher Order Logic and others that are based on Martin Löw type theory. Without an excellent mathematical background one is helpless.

If you want to analyze and predict signals in electrical engineering, this is comparably easy if you know how to do harmonic analysis and state of the art statistics.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .