General Advice -- Questions Welcomed

[fdhkjal]

I'm a first year PhD at a top 4 school; I remember the (horrible) process of applying last year, so I figured as deadlines start approaching I might give advice (particularly to all those, "will I get in").

(1). Recognizability is the key.

When you're applying, make yourself as recognizable as possible. From that I mean, if you have a paper at a conference, the more promininent the better (obviously this goes without saying). Also, your letters of recommendation are key. If you can get even a single letter of recommendation from a faculty member who is known by one of the POI at the school you are applying to (and it says good things), your chances increase dramatically. I have met more than a few people who didn't go to prominent schools, didn't have research experience or the best grades, etc, but managed to work with a leader in the field and get in.

For foriegn students, this can be very bad news. I would say that if you are a foreign student, and don't go to a top school in your country (China - I think its the top 7? Tsinghua, Peking, etc; India - a top IIT; Iran - Sharif; and so on), the chances of getting into a top 10 or maybe even 20 program are negligible. The best route to play if you are foreign and want to go to a top school is to first do a masters in the US/Canada/Europe(to a lesser extent) at a mid-tier school. This has two advantages: (1) mid-tier schools tend to be the ones that will provide support and a stipend for master's students, whereas top-tier schools see these programs as cash cows. (2). Arguably, your research opportunities at a mid-tier school at the master's level will be significantly higher; at top-tier programs, professors already have quite a few PhD students (who get preference over master's students), whereas mid-tier programs, professors have a harder time finding good people to work on the interesting projects they have funding for.

(2). GRE isn't that important.

In general, the GRE is taken to fulfill requirements. This is both true of the applicant, and the university. If anything, the GRE is to weed out people who REALLY don't belong there. Its hard to believe someone who has trouble with basic math (600- on quant) will be able to do hardcore theoretical computer science. Also, the Computer Science GRE is even more useless. The CS GRE tests mostly theoretical computer science knowledge, which is great if you're applying for theory, but doesn't test your knowledge in a field such as computer vision in the slightest. If any department would use the CS GRE, it would almost certainly only be in the theory subfield.

(3). When to contact.

Contacting a professor in the top 10 is simply a waste of time. First, most of these programs accept students to the department, and not to a specific advisor, and guarentee them funding. However, the further down the list you go, the more professors want to hear from you and know that you are interested in working with them. If you can contact a professor and get to know about their projects and show them the benefit of having you as their grad student, then I would say it greatly increases your chances. Lower-tiered schools tend to accept students to a specific advisor; at least more-so than high-tiered ones.

That's all I can think of for now. If anyone has any questions or anything they think they should add, I'd love to know.

[smmmu]

Hi fdhkjal,

many thanks for the advice, they're really good points.

I do have one question: Most of the top-ranked CS programmes / departments share two common features I think. Firstly, they are generally quite large in size. Secondly, they all do committee-based admissions to the programme (rather than to specific areas / faculty members).

For such programmes, how do you argue your fit to the department on your statement of purpose?

Sure, I can write something along the lines of 'I'm interested in Stanford because Prof. X and Y are doing very interesting work on Z, which fits well with my background in this area.', but then so are Profs. A and B at Berkeley and C and D at MIT, and so on, so I feel that this isn't a convincing argument why I should be at Stanford specifically. Also, obviously a programme like Stanford has outstanding faculty in just about any area you can think of, so I'm not sure how much an argument as above would make me stand out from the crowd.

At the other end of the spectrum, I could try to argue that I feel I fit very well with the programme overall due to the department's strength in several key areas that I might be interested in. But then this would essentially amount to not much more than 'I want to go to Stanford because it's Stanford', again not a very convincing argument.

Any thoughts?

[fdhkjal]

Hi fdhkjal,

many thanks for the advice, they're really good points.

I do have one question: Most of the top-ranked CS programmes / departments share two common features I think. Firstly, they are generally quite large in size. Secondly, they all do committee-based admissions to the programme (rather than to specific areas / faculty members).

For such programmes, how do you argue your fit to the department on your statement of purpose?

Sure, I can write something along the lines of 'I'm interested in Stanford because Prof. X and Y are doing very interesting work on Z, which fits well with my background in this area.', but then so are Profs. A and B at Berkeley and C and D at MIT, and so on, so I feel that this isn't a convincing argument why I should be at Stanford specifically. Also, obviously a programme like Stanford has outstanding faculty in just about any area you can think of, so I'm not sure how much an argument as above would make me stand out from the crowd.

At the other end of the spectrum, I could try to argue that I feel I fit very well with the programme overall due to the department's strength in several key areas that I might be interested in. But then this would essentially amount to not much more than 'I want to go to Stanford because it's Stanford', again not a very convincing argument.

Any thoughts?

I would stick to (this is what I did at least) "I'm interested in Prof X_1, X_2 at University Y because they are doing very interesting work in Z." Although it is a committee-based admissions process, unless you are the top applicant coming from harvard and have a hertz fellowship, you want a professor to want you. Think of what goes on in the committe-based admissions process if you write your SOP in that way: Professor X_1 (as you mentioned in your SOP) say "I think applicant A should be accepted," Professors X_{2, ..., 99} "I don't see anything wrong with that." On the other hand, if you don't have at least one professor fighting in your favor, it's much tougher.

One of the things I regret doing during the admissions process was being too general during a phone interview at one of the schools I applied to. I knew I was going to have it days in advance, so I looked through all the previous publications of the faculty in the area and took note of all the things I found interesting. When the professor conducting the interview asked what/who I'd like to work on/with I essentially ran through the entire area and all the professors in it, citing papers and all that. But in the end I didn't get accept to that school. Looking back I think the reason was because I was being too general during the interview when I should have been more specific, say, cite one or two papers, why I found them interesting, how my previous background relates to that kind of work, why I think I would be an asset in taking that research forward, and why the professor/group in charge of the cited work is the only place I would be able to do this research.

It's also worth noting that schools want people with a passion for something. Its more common for a person to turn a passion from Z to Z', than for a person to have passion come out of no where for Z'. The latter (people who come in wanting to get a degree from University Y because its University Y), tend to be the people who drop out more than the people who are less smart, but have a passion for something.

Edited November 30, 2012 by fdhkjal

[Azazel]

Hi fdhkjal,

many thanks for the advice, they're really good points.

I do have one question: Most of the top-ranked CS programmes / departments share two common features I think. Firstly, they are generally quite large in size. Secondly, they all do committee-based admissions to the programme (rather than to specific areas / faculty members).

For such programmes, how do you argue your fit to the department on your statement of purpose?

Sure, I can write something along the lines of 'I'm interested in Stanford because Prof. X and Y are doing very interesting work on Z, which fits well with my background in this area.', but then so are Profs. A and B at Berkeley and C and D at MIT, and so on, so I feel that this isn't a convincing argument why I should be at Stanford specifically. Also, obviously a programme like Stanford has outstanding faculty in just about any area you can think of, so I'm not sure how much an argument as above would make me stand out from the crowd.

At the other end of the spectrum, I could try to argue that I feel I fit very well with the programme overall due to the department's strength in several key areas that I might be interested in. But then this would essentially amount to not much more than 'I want to go to Stanford because it's Stanford', again not a very convincing argument.

Any thoughts?

Here are two examples of successful personal statements which, to varying degrees, address this question: Jean Yang's (now at MIT) and Ethan Fast's (now at Stanford).

From Yang's (for CMU, where she got in):

My career goal is to remain in academia as a professor so that I can pursue my research interests while training and recruiting others to solve relevant problems in programming languages. I want to continue my studies at Carnegie Mellon because it has a large programming languages group with many professors with whom I would like to work, including Professors Harper and Aldrich. Because I find it important to talk to with different areas of focus than my own, I also like the diversity of CMU's computer science research. Because I look forward to teaching as a graduate student, I like that CMU has a reputation for having motivated undergraduates who can challenge my understanding of the materials. For these reasons, I would like to spend my next six years at Carnegie Mellon.

From Fast's:

My research explores the boundaries of automated software development, investigating to what extent computers can assist humans in writing code. In particular, I am concerned with how techniques from machine learning might be applied to applications in programming languages. Along these lines, I am especially interested in working with Professor Engler on automatic bug detection, Professor Aiken on statistical debugging techniques, or Professor Koller on applications of machine learning. By combining statistical learning methods with elements of semantic and program analyses, I hope to construct algorithms that reduce the cost of developing reliable code.

[And later, addressing fit:]

In working with Professors Weimer and Forrest, I have learned much about what it means to produce high quality research, and I have come to realize that I will need a PhD to work on the cutting-edge problems with which I am interested. Moreover, I expect that I would thrive at Stanford. My contributions to APR occurred at many levels of the research process, from high-level brainstorming and hypothesis formation, to the logistics of experimental design and academic writing. Further, I have published work contributing directly to the state of knowledge in programming languages and machine learning. In short, I have discovered a passion for that deep level of understanding enabled uniquely by research, and as a graduate student at Stanford, I would have the resources and mentorship necessary to pursue this passion fully.

So I actually don't think you need to be particularly convincing (neither of these examples were!)

Edited November 30, 2012 by Azazel

[Snoq]

How much does your undergraduate research areas affects your latter (PhD) research and your Statement Of Purpose (SOP) if you want to change research area for your PhD? How should one write such a SOP (if wants to change areas)?

I am interested in the following areas: Algorithms, Computer Architecture and Machine Learning (even though I haven't taken any course in Machine Learning yet - I will take one next term). But I don't know in which area to focus. And there is a possibility that those interest may change (I also like networks). I can only do research in two of those areas (and I have still 3 semesters and 2 summers to spend).

[fdhkjal]

How much does your undergraduate research areas affects your latter (PhD) research and your Statement Of Purpose (SOP) if you want to change research area for your PhD? How should one write such a SOP (if wants to change areas)?

I am interested in the following areas: Algorithms, Computer Architecture and Machine Learning (even though I haven't taken any course in Machine Learning yet - I will take one next term). But I don't know in which area to focus. And there is a possibility that those interest may change (I also like networks). I can only do research in two of those areas (and I have still 3 semesters and 2 summers to spend).

For a top tier schools, I would say that your undergraduate research area definitely affects your statement of purpose and your admission chances, but not necessarily your PhD research. For example, if you had lots of research in HCI but really wanted to do theory, you'd be better off writing your SOP towards HCI and then switching after you already got in. The best way to prove that you want to do research in a field, is by having done research in that field, kind of like the best way to get a job is by having a job. Yes I know this is somewhat disingenuous and deceitful, but ultimately your best shot of getting in. The last I heard, over 50% of incoming grad students (at my institution) change their area (I mean big changes like Systems -> AI, or AI -> Theory; not small changes like Computer Vision -> NLP or Algorithms -> Computational Complexity), so its not that big of a deal if you change. Although, it definitely helps to have done research in the area before starting your PhD, and it helps even more to know what area you're doing when you start; you simply waste less time doing things that don't ultimately help you with your thesis.

For lower tiered schools, since a professor normally accepts you and your funding options are more limited, you might be stuck with a certain professor/area upon acceptance, so I would be a little less deceitful in your intentions.

In any case, the best way to have freedom for your research area is to get a fellowship; unless a professor has a ridiculous number of PhD students, they're always willing to take on someone who they don't have to fund.

Yes, I know this is sort of a convoluted answer, but I think there's lots of factors involved.

[Snoq]

Here are two examples of successful personal statements which, to varying degrees, address this question: Jean Yang's (now at MIT) and Ethan Fast's (now at Stanford).

 

Are there available any other SOPs from students who got in top universities (not explicitly only those two)? I think i will be having a problem whith my SOP because I will be starting my thesis next year's January so I will be having only a few things to present about my research.

[fdhkjal]

This might not be exactly what you're looking for, but here is a list of some NSF applications: http://www.alexhunterlang.com/nsf-fellowship .  I wrote my NSF application before I wrote my SOP, and essentially condensed it into statement of purpose, getting rid of stuff that's relevant for the NSF but not for grad schools (i.e inner city, high risk, female, racially diverse, under represented minorities in STEM).

 

There's a few statement of purposes I ran across, but chances are they're not relevant to most applicants (see: http://people.csail.mit.edu/mip/docs/phd-application05/statement.pdf), or I guess if they are relevant, then you really shouldn't be worrying about getting in anywhere.  You can also find a bunch of statement of purposes online with some google search trickery (of people who put stuff up accidentally), but I'll leave that as an exercise for the reader.

[chron]

I'm applying for a Master's (thesis-based) and not applying to a top 20 US school or equivalent. Should I be worried about getting stuck with a particular professor / research area upon admission? I had thought if I applied for a Master's rather than PhD I didn't need to have a specific research area in mind to get in. But one school says I should contact profs after I submit my application. Doesn't that mean I need to have a specific research area already in order to contact profs in that area? What should I say to the profs? How many profs can I contact?

 

Even besides this particular school, I am confused about whether or not I should be declaring my areas of interest on my SOP, mentioning profs' names, or contacting profs. Although I think I have a general direction, I don't have a specific area of interest. I have been mentioning a few different areas of interest along with the names of a few profs on my SOPs so far, but I'm unsure if that's the way to go. Or if it's good or bad idea for me to be contacting profs at each grad school.

 

I had thought that I would be able to enroll and spend some time talking to advisors and concentrating on courses at first before I have to pick my topic/supervisor for my thesis. But I'm also not a top candidate, and need all the help I can get to get into a grad school, so if it means I have to pick an area now I guess I should do it.