anacron

Just accepted the NSF, so good luck to those still waiting for NDSEG.

As someone who is currently on a department fellowship and filed taxes, they do not withhold taxes on fellowship stipend because the university does not consider it payment for work (i.e. they do not generate a W2). If you read the IRS publication about fellowships, they are non-taxable if the stipend went towards educational expenses (tuition, fees, books, etc). Housing/living costs are taxable and should be reported as income.
 
Your university should generate a yearly tuition expenses/fellowships form (I think its a 1098 or something) which will report your fellowship stipend. After reading up on the IRS material it wasn't too hard to file my taxes.
 
The usual disclaimer applies: I'm not a tax advisor blah blah blah =D

I'm in the same situation as OP but I decided to accept the NSF GRF funding.
 
Here are my reasons:
 
1) Funding is approximately the same (NDSEG pays slightly more)
2) Even though the NSF GRF doesn't cover full tuition, my department will be waiving the rest.
3) Much more flexibility on when to take funding (CS PhDs in my department have been successful in winning one-year industry fellowships from Google, Microsoft, etc.)
4) NSF GROW - I'm really interested in doing research abroad so NSF GROW is a great opportunity.
 
The clincher was the flexibility; I like having the option to defer funding so that I can apply for industry/national lab fellowships during my graduate study (I'd hate to be locked in for 3 years and miss some of the eligibility criteria).

I just called the office. Notifications are going out either today or tomorrow.

I'd advise against it, I got similar ratings last year and just worked harder on my research the following semester.
I ended up leading a very cool project that got lots of people excited and finally won the award with an improved application.
 
Point is, you'll have 2 more chances to apply so take the reviewer comments seriously and use this as extra motivation to improve!

I'd advise against it, I got similar ratings last year and just worked harder on my research the following semester.
I ended up leading a very cool project that got lots of people excited and finally won the award with an improved application.
 
Point is, you'll have 2 more chances to apply so take the reviewer comments seriously and use this as extra motivation to improve!

I'd advise against it, I got similar ratings last year and just worked harder on my research the following semester.
I ended up leading a very cool project that got lots of people excited and finally won the award with an improved application.
 
Point is, you'll have 2 more chances to apply so take the reviewer comments seriously and use this as extra motivation to improve!

Hi, I am a Computer Science and Mathematics double major and Economics minor at a top 10 university and want to apply for a PhD in Computer Science (in Artificial Intelligence, Machine Learning, or Algorithms). I have a 3.8 overall GPA and a 4.0 in Computer Science. I've had 2 Microsoft software engineering internships. Since I am a junior I don't have my GRE score yet (but I got a 35 on the ACT in high school so I expect it to be decent). LORs should be great and the profs who will write them have all written be recommendations before. I hope to TA next semester in/close to my research area, and I have many other CS/math extracurriculars. 
 
I have worked on various research projects. One dealt with computer science education and resulted in conference publication (one of several authors). I am working on math research with a professor that might end up getting published but I'm not sure. I'm writing my thesis on game theory algorithms to graduate with distinction in CS, and hopefully I can try to get a publication from that before I apply for grad school. I have done other research which naturally did not lead to publication—a paid research assistantship in the business school, and some CS/optimization work for a committee that has real-world applications for my uni rather than a theoretical basis for a paper.
 
Although I have been involved in all this research and have loved it, I am nervous about my chances of getting into top PhD programs if the only publication is the one on computer science education since it's not in my CS interest area. I'm also worried that if this happens my research will seem scattered. If I only end up with that one publication, would it be better for me to apply for Masters first if my aim is to attend a top PhD program?
 
Do you think I could be competitive in the PhD programs below, and if not what range should I aim for?
Carnegie Mellon (Computer Science- Algorithms, Combinatorics and Optimization)
Stanford, Harvard, MIT, etc.

Part of the game is to be a concise as possible since the readers only have a limited amount of time to go through apps. You do yourself no favors by rambling on for 5 sentences when 1 would do. In any case, I doubt they'd throw out an app just because you added a few of extra sentences.

If you really have that much research experience, or that much to say then you could potentially put up a website with all the details and add a pointer to the site in your statement.

Chances are unlikely at the top schools like U. Waterloo or U. Toronto since you've been out of undergrad and unfortunately because of the unemployment. One of the big factors, like you mentioned, are the reference letters. It might prove difficult for profs. to write strong letters for someone they haven't seen for a few years - of course, this depends entirely on how much interaction you've had with them in undergrad and post-graduation.

Why did you pick computational science specifically? Most of these jobs require a PhD to even be considered for employment or at least want an MS + several years of experience. It seems like your interests are geared more towards research fields than industry so I'd tread carefully because industry jobs might be hard to come by especially since you're not interested in a PhD.

For a Master's only program I'd try to maximize employment outlook by picking relatively broad and established specializations such a software engineering or systems. AI/machine learning aren't bad choices especially given the whole big data and cloud movement. If computational science is absolutely what you want to do then you might want to look at PhD programs instead.

I think you should focus your SOP on research experiences that you've had and how those experiences (in whatever area) would help you contribute to other areas - diversity in research is a good thing. If you can link the various experiences under a common theme then you would have a strong SOP.

Personally, I would refrain from name-dropping too many research topics (i.e. being too broad) unless you have something meaningful to say and if you can link those areas to your undergraduate research. Be concrete and give examples of potential projects that you could work on in the areas that you mention. Don't just write down a laundry list of research areas.

Keep in mind that you can always change your mind once you get admitted..

Don't limit yourself, research shouldn't be just about one sub-field or another. You should really aim for what interests you and where your research takes you. I'm actually in the HPC field but I'm moving towards applications of machine learning algorithms and molecular dynamics algorithms.

So, I wouldn't go into grad. school thinking that its one or the other. It should really be an opportunity for you to explore whatever interests you.

Don't limit yourself, research shouldn't be just about one sub-field or another. You should really aim for what interests you and where your research takes you. I'm actually in the HPC field but I'm moving towards applications of machine learning algorithms and molecular dynamics algorithms.

So, I wouldn't go into grad. school thinking that its one or the other. It should really be an opportunity for you to explore whatever interests you.

Based on my admissions experience, no research experience will kill your chances at top schools. They can overlook a bad GPA if there was lots of (stellar) research experience to back it up. However, the sheer number of applicants makes it hard to stand out without any research experience and a bad GPA. Now, theoretically, waiting a year and getting some research experience might sound good but will you be able to find someone to work with, do publishable work and publish that work within a year? Even then there is no guarantee...

As far as bringing your own funding is concerned, I don't think the big universities would really care about that. They bring in more than enough funding through grants and industry sponsors to be able to fund every PhD student they take in with RAships. This was certainly the case at UIUC and Berkeley.

A Ph.D from UPenn is nothing to frown upon. Sure, it may not be your dream school but in the end your Ph.D is defined by what you do not where you do it. If you can take the rejection from your dream school and turn it into motivation for a great thesis, then - who knows - you may just end up at your dream school doing a postdoc.

Don't give up a sure thing just because its not the best of the pack.

I agree with most of everyone else. Although it is somewhat surprising to me that someone with an MS from a top-20 school AND their own funding get's accepted into only one school, your overall application must have just been okay and didn't impress anyone. A professor from UC Berekely told me that someone with an MS and no research experience is much much worse than someone with a BS with no research experience. At the MS level you've had a year or two to prove your research skills. I think overall it was a lack of research skills, lack of GPA, and possibly a lack of LOR that just made your application decent. I believe a Professor from UPENN saw that you were given funding and decided to take a chance on you. Honestly, if you've applied twice and only get accepted once, I wouldn't take that risk again.

This is true, but there are many applicants who have both stellar GPAs and research experience. I am talking about the top universities where there is no shortage of well-qualified applicants. In order to be competitive there you need to be good at everything...

I say this through my experience and through conversations with Profs. at UIUC and Berkeley. By the way, a 3.05 is fairly low. Berkeley's minimum requirement is 3.0 - so being close to this means you need to make up for it with research experience.

If you want a good shot at these schools - don't take GPA too lightly. That said DJLamar's advice stands if you're worried about optimizing your GPA (ex. you have a 3.7+ GPA and you're trying to decide between research and improving GPA).

Every CS PhD applicant should read this 'paper' written by a CS Prof at CMU: http://www.cs.cmu.ed...dschooltalk.pdf

Based on my admissions experience, no research experience will kill your chances at top schools. They can overlook a bad GPA if there was lots of (stellar) research experience to back it up. However, the sheer number of applicants makes it hard to stand out without any research experience and a bad GPA. Now, theoretically, waiting a year and getting some research experience might sound good but will you be able to find someone to work with, do publishable work and publish that work within a year? Even then there is no guarantee...

As far as bringing your own funding is concerned, I don't think the big universities would really care about that. They bring in more than enough funding through grants and industry sponsors to be able to fund every PhD student they take in with RAships. This was certainly the case at UIUC and Berkeley.

A Ph.D from UPenn is nothing to frown upon. Sure, it may not be your dream school but in the end your Ph.D is defined by what you do not where you do it. If you can take the rejection from your dream school and turn it into motivation for a great thesis, then - who knows - you may just end up at your dream school doing a postdoc.

Don't give up a sure thing just because its not the best of the pack.

I think the numerical scale would have some of the same problems E/VG does -- it would still confuse people to have higher numbers than others who won. Even if z-scores were given, we'd have the same confusion due to 1) re-ranking and 2) differences in winning z-scores across panels -- even in the same major field of study since many of them have multiple panels. This leaves the actual rankings within the panel groups, but there's a problem there too: geographical bonuses that come into play AFTER re-ranking if you're in Q2.

I think what might be useful is a modified, coarser grain releasing of the panel rankings:
1) Don't give out ranks to those offered awards. Yes, this means those offered awards don't get as much information, but I don't think it's a good idea to point out who likely got an award due to geography. Furthermore, it might even lead to things like encouraging people to list their rank if they won (since high ranking people might) or high ranking HMs comparing themselves to winners.
2) Divide the remaining applicants into rank-deciles, covering the whole range of remaining ranks. Fudge this a little so that the Q3 and Q4 don't overlap a decile (because having an HM and a non-HM in the same decile might cause problems). Let them know their deciles "e.g. Top 30% of non-awardees." Since panels review 250-300 applications, assuming ~15% get awards, this leaves at least 200 applications, so the deciles would cover about 20 people each. (At this point they could also just take the remaining 200 and renumber them 1-200, but I think fine-grained rankings like that would be misleading because the process to come up with ranks isn't that precise.)
3) Give everyone their reviewer comments of course!
4) I think everyone (award or no) should also get their public K-12 textbook scores (E/VG/G/F/P) too. It's something closer to what the individual reviewer is thinking and I think that's important for everyone's future proposals. I think having positive comments with an E and positive comments with a G do tell us something, even if we're not quite sure what. Reviewers care a lot about science and students and are working hard all weekend to get through the applications while trying to be fair, but they aren't infallible and time is short -- that's why we're now mining comments to make more sense of the feedback, but this is something that's going to continue to vex us for the rest of our publishing and grant writing lives. The decile-rankings and the E/VG ratings will tell different parts of the whole story to the non-awardees.

If you were admitted to UCB, please don't hesitate to contact your peer advisor (or faculty members who've contacted you) for any questions like this.

But the short answer is: yes, you've understood the course requirements approximately correctly [1]. 24 credits is approximately 8 courses. Most people take about 2 courses per semester, so you can finish off your coursework in about 4 semesters. Keep in mind, your primary focus in the PhD program is research (rather than coursework) -- this is a significant change from an undergraduate program.

No, Berkeley is not unusual in this regard. If anything, Berkeley PhD students take more courses than at many other PhD programs. I think you'll learn a lot in the courses, and they'll support your research nicely. Many Berkeley courses have a large project component, where you do a research project of your choosing -- which is a great way to learn about how to do good research and try out new research ideas.

Again, if you have questions, please don't hesitate to contact your contacts at Berkeley! They'll be glad to explain how it works, if this is a concern for you -- but let me reassure you that the requirements are quite reasonable and are nothing to be afraid of.

[1] OK, OK, there are some details. We'll tell you about them if you come, or if you ask. At this point, you really don't need to worry about them. (But if you want to know, technically, the 24 credits is divided as follows: at least 12 credits in your major area, at least 6 in an inside minor area, and at least 6 credits in an outside minor. The major area and inside minor are two areas within EECS. The outside minor must be graduate courses in some department outside EECS.)

I emailed one of the program coordinators and got the following news:

Dear anacron,
We will be notifying applicants early to mid April and will let the
awardees know when we need their acceptance decision.

That was me.

I emailed one of the program coordinators and got the following news:

Dear anacron,
We will be notifying applicants early to mid April and will let the
awardees know when we need their acceptance decision.

As usual, these articles fail to acknowledge opportunities for PhD-level scientists outside the academy. We get DoE national labs, DoD labs, the NIH intramural program, NASA, NOAA, the NSA, FFRDCs, government contractors, nonprofit research institutes (e.g. WHOI), Big Pharma, technology startups, the energy industry (both traditional and alternative), the research labs of large for-profit companies (Intel, IBM, Microsoft), the Public Health Service Commissioned Corps, the military. Not everybody wants or needs to be a professor. PhDs are crucial to large parts of industry and government. The fact that there are more STEM PhDs than there are assistant professorships is a feature, not a bug.

One of my projects at work - a government contractor - has four PhDs working on it. These people didn't fail to get academic jobs. A couple of them were the pick of the litter. They came here by choice. They are happy here, they get to work on cool research, they avoid the race to tenure, and they make more money than they would in academia.

This sentence is just incredible:

"Many young Americans bright enough to do the math therefore conclude that instead of gambling 12 years on the small chance of becoming an assistant professor, they can invest that time in becoming a neurosurgeon, or a quarter of it in becoming a lawyer or a sixth in earning an MBA."

Ha. Hahaha. Do these authors know what the job prospects are these days for lawyers, and how much it costs to go to law school? What kind of oblivious person these days markets law school as a path away from penury and toward job security (of course, if you combine it with a STEM PhD, you can be a patent lawyer and have a relatively easy time getting a cushy job)? Do they realize that in order to be a neurosurgeon, you have to spend four years in med school at $50k/year, completely win the residency matching process - almost nobody's a shoe-in for a surgery residency - and then spend six years working 80+ hour weeks for pay of $40k/year, and maybe a couple more in a fellowship? Have they noticed that an MBA doesn't do much for most people unless it's from a highly competitive, expensive, top school? At least in a STEM PhD program they pay you while you get your PhD, even if not very much.

Also, with all due sympathy for the biomed folks, not every field of science expects you to spend half your life bouncing around postdocs.

"A prime symptom noted by all: a growing aversion of America’s top students — especially the native-born white males who once formed the backbone of the nation’s research and technical community — to enter scientific careers."

Oh noez, not the native-born white males, whatever shall we do...oh wait, in the modern world, the US scientific talent pool includes women, people of color, and immigrants. Now, in fairness, they say that top US students in general are avoiding scientific careers. But aren't they writing this article about how US students shouldn't go into scientific PhD programs? Shouldn't they approve? If they don't, have they ever considered that articles like this one might be playing a role?

Eh, just disregard this. There is apparently too much emphasis on one magazine's rating of a college and this causes them to behave badly. I know UIUC's law school has dropped several rating spots and is now under investigation from the Bar Association for falsely inflating their admitted student's statistics. Hope it was worth it for them. I sent my transcripts to UIUC and I was told "I am eligible" to apply whatever that means. Obviously they want to reject people with high grades and waste their money to make themselves look good. And of course Roy Cambell had to mention he knew the director of the program I am encouraging me to apply. As an alumni to University of Illinois I will remember this and especially remember the bullsit batch email I got.

I know I am probably just bitter after a rejection letter I received from UIUC and job prospect rejects but I am wondering about CS departments at university systems with a campus that has a high CS ranking. I know from what I read on a blog written by a former US News writer that the college rankings cause colleges to behave badly. I am currently working on a MS in Computer Science at University of Illinois Springfield. Even though the director of this program knows the admission director at UIUC and my GPA is 4.0 with two classes to go, which I currently have an "A" average, I was rejected to UIUC. I know that their is a lot of competition to their MS and PHD programs because they can be funded (MCS is not funded and has a better admission rate). It got me thinking though about other campuses of colleges that have a high ranking. Should you bother with those or will you always be in the shadows of another campus? This has no baring on their program, they may have a good program. You know who I am talking about, UMBC versus University of Maryland-College Park, U of Wisonsin-Stout versus Madison, University of Michigan-Dearborn versus Ann Arbor. I personally can see why some may want to attend a different campus like not having to relocate, a city location like University of Illinois-Chicago, or smaller class sizes versus a stadium sized lecture hall.