Viet Academics

Demystifying the Computer Science PhD Admission in US Universities, Part I

A guide for Vietnamese and International Students

Having been involved in PhD admissions for many years, I’ve realized that many international students, especially those from smaller countries such as Vietnam, lack a clear understanding of the Computer Science PhD admission process at US universities. This confusion not only discourages students from applying but also creates the perception that getting admitted is difficult compared to CS PhD programs in other countries.

So I want to share some details about the PhD admission process and advice for those who are interested in applying for a PhD in Computer Science in the US. While this document is primarily intended for Vietnamese interested in computer science, it may be relevant to students from various countries and disciplines. Moreover, although many examples are specifics for George Mason University, the information should be generalizable to other R1 institutions in the US (though very top schools could be very selective, e.g., see the admission process at CMU).

In addition, this document can help US faculty and admission committee gain a better understanding of international students and their cultural differences. By appreciating and leveraging these differences, CS programs in the US can attract larger and more competitive application pools from international students.

I wish you the best of luck. And if you follow this guideline, you will at least have a good chance at GMU (see why you want to study at GMU). Happy school hunting!

This document is available on Github. If you have questions or comments, feel free to create a GitHub issue for discussion.

“Don’t make fun of graduate students. They just made a terrible life choice”, quoted from the Simpsons.

1. Should you apply?

First, I want to emphasize that PhD students in Computer Science do not need to worry about funding, especially at good R1 universities in the US. If you are admitted, you will almost certainly receive full funding to support your study, including tuition, health insurance, and stipend (monthly salary). Moreover, depending on the university, you may even receive additional benefits such as summer pay, laptops, (conference/workshop) traveling. Section 5 provides more details on funding.

Second, I believe that applying to a good US university should not be any harder than at schools in other countries. If you think you have a chance in other countries, e.g., South Korea, Singapore, Germany, UK, Japan and Australia, then you will surely have a chance in the US as well.

Vu: One of the reasons I create this document is that my colleagues at GMU are interested in recruiting Vietnamese students and are surprised when seeing very few applications in Vietnam (e.g., each year our CS program receives more than 350 PhD applications, most of which are international but only 3–4 are from Vietnam). In general the number of PhD applications from Vietnam to US universities is few and more would be very welcomed.

2. How is Your Application Evaluated?

After you submit your PhD application (usually in December), it will be first screened for general requirements, e.g., Did you submit your transcripts and standard scores? Did your reference writers submit their letters?

Then applications will be reviewed by a PhD admission committee that consists of faculty members in CS (in some cases the committee can involve affiliated faculty from different disciplines). These faculty have a wide-range of expertise and background to ensure diverse perspectives in the evaluation process. The size and the review load of the committee depends on the department size. At GMU, the PhD committee typically has 15–20 faculty, and each committee member is assigned from 25–30 applications. Note that at GMU (and probably many other schools) we have a separate committee for the MS program.

Hung: At UMass Amherst, we also have a separate committee for MS applicants.

Each application is then assigned to about three faculty members, who will evaluate your profile and reach a consensus. Note that while the assigned reviewers are likely the main ones deciding your application, every faculty in the department will have access to your application and can provide inputs and opinions on your profile.

In many cases, the admission committee involves assistant professors in the department. This provides junior faculty the opportunities to recruit students. The chair of the committee will be a senior professor, but they likely will not review individual applications and instead assign them to committee members. The chair will look at various factors such as research interests or mentioning faculty names to assign the applications to appropriate faculty.

Vu: We usually decide that a full-time PhD candidate is either (i) admit with funding (TA or RA) or (ii) rejected. In other words, in most cases, we either admit you with full funding, or we don’t. In some rare cases, we admit without funding because you have funding on your own (e.g., supported by your government or having external grants). We also justify our decision with a summary about your application, where we list strengths (e.g., well-known school) and weaknesses (e.g., weak LORs).

Hakan: At GMU, for full consideration, students should make sure to submit ALL required documents by the application deadline, and should never assume that some required documents (such as official TOEFL scores or official diplomas/transcripts) will be waived by the admissions office. If something is listed and not marked as “optional”, it is mandatory and they should plan for submitting all those.

Why we do not waive application fee? This is typically a requirement of the university. Individual departments and programs do not have the flexibility to waive the application fee, even if they want to.

In my opinion, requiring applicants to pay the fee helps ensure their seriousness, as it filters out non-serious candidates. Also, if the application process were free for everyone, we would receive an overwhelming number of applications to review.

Hung: The application fee varies between schools, and is about 70$-100$. I found this number pretty ridiculous. In most cases, there is no proper justification for that amount. Yes, the application fee could act as a filter, but a nominal amount would do the job. While a blanket waiver might not be feasible, universities should extend the eligibility of fee waiver applications to more countries.

3. Application

The primary focus of the admissions committee is to evaluate your background and interest in research since a PhD in Computer Science and determine if you would fit into the program. To evaluate your profile, we consider the following key indicators, listed in order of importance.

3.1. Research Ability

The most effective evidence of research ability is having published papers in reputable international journals or conferences. Having published good papers is a sign that the applicant has successfully involved in research.

It’s important to aim for the top venues in your field. You can find top CS conferences at places such as CSRankings, which ranks CS programs based on how their faculty publish at top conferences. Local conferences and non-English journals or conferences do not carry as much weight since their quality is often unknown to US faculty.

However, I understand that many international students do not have the opportunities to publish in top places, so general confs/journals would suffice. But be sure to upload your papers with your application and talk about them in your statement (see section 3.3).

Vu: Many international students mention Scopus Q1, which consists of various journals from IEEE, Elsevier, and many other publishers. I don’t know/recognize many of journals listed in Scopus Q1. This might be something to be mindful of, as CS faculty might not be too familiar with Scopus or journals listed there, so devote sometime in your statement to discuss the significance of your papers.

Craig: GMU and many other universities allow you to upload your published papers and other writing samples. In many cases, even if the papers were not published at top places, we can still determine their quality by simply skimming over the paper.

Additionally, work experiences at renown research laboratories, such as Microsoft Research, can significantly strengthen your application. Unfortunately, many good research places in your countries, e.g., VinAI in Vietnam, remain relatively unknown to most universities in the US. So you should explicit say something about them in your statement.

Hung: The reputation of VinAI has been increasing steadily over the past few years; many of my colleagues heard about VinAI.

Finally, participating internationally recognized competitions can also demonstrate your research potential. For example, participating in Math Olympiads if you want to do theory or winning ACM programming contests if you want to “build” stuff, e.g., software analysis.

Thanh: Due to academic culture, professors in Vietnam usually aim for (international) journals instead of conferences. Could you give some tips on how to know whether a journal is good (CSrankings, unfortunately, only consider conferences)?.

Vu: One way is looking at what well-known researchers publish at. For example, if you are interested in a field X, you can use CSRankings to look at active faculty in X, and then look at their websites to see what journals they publish at.

3.2. Letter of Recommendation (LOR)

CS PhD programs will require at least two LORs. Having a letter from an internationally recognized researcher can greatly strengthen your application. However, obtaining such letters can be challenging for international students, who might not have much interactions with such experts. So it is acceptable to have a letter from professors that know you well enough to talk about your specific research experience and capabilities.

Many students have letters written by themselves and signed by their professors. These have little values. We can recognize them and will consider them weakness. Similarly, many professors write generic letters for students, e.g., the students didn’t do any research or make any impression for the professor to write about. These letters are also not useful and considered weak.

Many students get letters from supervisors from company where they did internship or are working at. It is OK as long as it is a research-based personalized letter (once again, we are talking about PhD applications, not MS).

Vu: Some letter writers ask the applicants to write their own letters for them to sign. As mentioned, this will hurt the applicants (admission committee members are actually quite good at determining this)! I was also told that some people did this so that the applicants do not go abroad or only go to places where they want the students to go to. Sometimes students would go through great length just to get letters from well-known professors in their school, but the letters are generic and carry little value, in fact, red flags. Moreover, a top professor in Vietnam might not be well-known to US faculty (see more details in section 3.4). So save the trouble and just get letters from any professors/supervisors who knows you well and can write a good letter about your research ability. It’s better to have a good personalized letter about your own research ability from someone who is less well-known than a generic/weak letter from a well-known person.

Didier: Should letter writers have PhDs? In Rwanda, a lot of students interact more with teaching faculty who might not have PhD.

Vu: In general, at least one writer should have PhD to properly evaluate your research ability. However, I think this is an interesting and useful detail that US faculty might not be aware of and students should mention about this in their statements.

Hung: A sad reality is that most professors in Vietnam DO NOT know how to write a good letter, or are lazy in writing letters hence delegate the writing to the students. Unfortunately, there is no easy solution to this problem.

3.3. Research or Personal Statement

While you might not have control over LORs and where you go to school, you do over your statement! So write it well because we take it seriously. In your statement, you have the opportunity to make make your application stand out and unique, which can make you fit the CS program you’re applying to, even if you don’t have very strong research experience. A well-written statement also shows that you can communicate, which is very important in research, and that you can effectively teach and communicate with students, which is important for TA funding (see section 5).

There are various guides on writing statement, and many example statements are available. So I will not talk too much about statements. In short, discuss about your research vision and convince us that you can achieve it through your experience, e.g., published papers, or if you work on some projects by yourself, talk about it. Also, use the statement to talk about stuff that admission committee members might not know about, e.g., your Github project with 1K+ stars or your regular contributions to well-known open-source projects.

Hung: I think asking for a “research vision” from a Ph.D. applicant is too much. Even people graduated with a Ph.D. have a great difficulty in explicating their researh vision. In my opinion, you should focus on showing (not telling) your research passion (if you do have).

Finally, this is something easy to do, but is missed by many applicants: customize the statement for the school you’re applying to, e.g., why do you apply here? Provide names of professors who you’re interested in (in many cases your application will be forwarded to them for evaluation). This shows that you’re serious and have done homework on places you’re applying to. Admission committee will look for this part at the end of the statement, so do not skip it.

Vu: I always read the research statement first and then LORs. If I am persuaded by then, I would skim over other factors and advocate for admission (unless I see red flags in other parts). If I am not convinced, then I will likely recommend rejection (unless I see something standout in other parts). Do careful research on professors, don’t mention emeritus or adjunct faculty. Also, be careful not to send statements to wrong schools or mixing facts (e.g., talking about school X but mentioned about working with profs. at school Y; and definitely do not talk about George Washington when applying to George Mason). I have seen such statements more time that I should.

3.4. Your School

Graduating from top universities that we recognize helps. For example, if your school is well-known, then it is “top foreign”, which is definitely a plus. However, if committee members do not know much about schools in your country, they will likely treat your school as “unknown foreign”, which can be a minus point because we are uncertain about the quality of your school.

So what can you do about this? several things including asking your CS dept to put itself on CSRankings (it’s the easiest way to get CS people to know about the school) and explaining about your school in your statement (ask your LOR writer to do that too). Of course, if you’re interested in working with Vietnamese professors, consider a PhD program in the US that have them.

Vu: Sometime PhD admission committee in the US will share a document such as this one, which lists the top schools in several countries. In some cases we ask other faculty and students if we think they know about the place. For example, when I was a postdoc at UMD, people in admission committee ask me to evaluate applicants from Vietnam. During my time at UNL and now here at GMU, I have looked at Vietnamese applications (whether they are assigned to me or not) and provide input to their reviewers, e.g., X is the top tech school in Vietnam and so it should be top instead of unknown foreign, which makes a huge difference.

Hung: I personally don’t care about which schools students attended. There are many other ways to evaluate an applicant.

3.5. Grades/GREs

Having good grades is important, but unless your school is well-known, having top grades/ranks usually will not help. This is simply because we cannot evaluate them.

This can be an issue for students in many top international universities where the competition is so high that very good students can still have low GPA relative to others (and be overlooked by Admission committee). So what to do with this? well, same as section 3.4., e.g., put a note about this in your statement and ask your LoR writers to talk about it.

Note that while having good grades at unknown school might not help, having very bad grades will be red flags (unless your LORs or statements give proper explanation). This is especially true if you have bad grades in relevant, e.g., CS and Math, courses.

Thanh: Vietnamese universities typically offer specialized programs, such as the talented engineer program at HUST, that have highly competitive entrance exams and a limited number of available slots (e.g., 30 per year). However, these programs often set higher requirements for students, including more demanding tests and assignments, resulting in lower GPAs and overall rankings. For example, an 3.5 GPA students from such talented programs are typically much better than a 4.0 GPA students not in those programs. Similarly, variations in GPA standards exist among different universities, with technical universities generally having lower GPAs than economical universities. These make gaining admission in the US difficult as US faculty are not familiar with these issues.

Vu: Vietnamese students and even faculty often lament how this competitive/low-grading system hurts Vietnamese students when applying abroad. One way to mitigate this is making these issues known to admission committee in your statement. Schools with Vietnamese profs are probably aware of them, but in general your letter writers/you can explicitly mention these in their letters/ your statement.

GRE Most CS programs in the US no longer require GREs, so you don’t need to take them. However, they might be useful for international students from programs we are not familiar with.

Vu: Personally I do not pay attention to GRE scores as I do not believe they are good indicator for research. However, if you have good GRE scores then you should list them in your application.

English Test Unless your degrees are from certain countries such as these, you will need to take standardized English test. Just do well enough to pass minimum requirement set by the university, which nowadays has many options for you to choose from.

Vu: The minimum for GMU (being above this might not mean much, but below is a red flags.

3.6 CV/Resume

This should be a summary of the accomplishments of the applicant. It should allow the reviewers to quickly scan to identify standout achievements (e.g., Publications, Programming Competition Awards, Teaching Experience).

3.7 Interview

Sometime a faculty wants to interview an applicant to make a decision. This means they are leaning toward admitting you (if we don’t like your application, we will not bother doing the interview).

An interview lasts about 15–30 mins, and one implicit thing you will be evaluated on is whether you can communicate effectively (i.e., speak/understand English). You will also get chance to ask questions about the university so think of something to ask (just the same as interviewing at a company).

Vu: At GMU, we are encouraged to interview candidates. For very strong candidates, the interview is actually to recruit them. In some cases a faculty interviews a candidate that they see potentials and wants to argue for admission, i.e., without the interview, that application is likely rejected. In any case, getting interview means you have a very good chance of being admitted.

(to be continued)

by ThanhVu Nguyen