starmaker

There's a lot of machine learning folks here, though we seem to be into different aspects of it. I'm into neuro/bio-inspired computing, robotics, and computational perception. yamamotomiechan, I think you're right that most people in the US don't know about your qualifications, unless they have ties to Japanese industry. I honestly don't think your industry engineering experience will be that useful for getting into PhD programs - it's not research-related enough. It might help with MS programs, though. The fact that you have research experience and publications will help will PhD programs. The teaching experience might help with both MS and PhD programs, because it might help you get a TAship. Your verbal GRE score is very low, but some programs will not care as much about that as others.

Here is the fundamental problem with having a grad student be your recommender for a PhD program: The point of recommendations is for your recommenders to vouch that you can succeed in and complete a PhD program. A grad student, no matter how great a grad student, hasn't completed his or her PhD. He or she still has part of the process left. Why would anybody take his or her assurance that you can complete your PhD seriously? For people (like me) who have worked in industry between undergrad and doctoral study, this problem comes up with our bosses. Depending on who my boss is over the next year, I could run into this. However, at least a boss in an industry research position has the experience of being a PI, even if he or she doesn't have a doctorate, and can attest to your ability to become a PI. It's something, at least. If the recommender were a postdoc, it would be different. A postdoc might not be a famous scholar or anything yet (and might not have had the experience of being a PI - the opposite problem from the industry boss), but a postdoc has completed his or her doctorate.

Yes. I can't imagine any way in which it could hurt, and it will probably help. Having submitted a paper is good, but having had your work vetted and accepted by qualified scholars - which is what the peer review process is about - demonstrates your ability to do quality scholarship in a way that just having submitted it for such vetting doesn't.

The information contained in your signature is basically useless. It's just GPA and GRE scores. Those alone won't get you in anywhere. Your verbal score seems awfully low. If you want to see how CMU evaluates PhD applicants, read this (by a CMU CS prof): http://www.cs.cmu.edu/~harchol/gradschooltalk.pdf

I was also a field-switcher (though I did it right after undergrad), so I can relate. To some extent, it might depend on what your undergrad curriculum was like. Computer science curricula vary so widely in terms of requirements! If you didn't take, for example, intro physics and chemistry classes, or calculus and differential equations, you'd probably need to take those. Different programs will have different prerequisites, but NukeE's the sort of field that gets a lot of grad students who didn't major in it as undergrads (though they are more likely to have come from something like physics than from CS), and that will help you. If all else fails, and you are eligible for US military service, you might consider trying for Nuke School in the Navy, whether as an officer or as an enlisted sailor. This is a very academically rigorous program, and the Navy will actually pay some of the people who work in their nuclear-related positions to get graduate degrees in NukeE.

If I get in somewhere for fall 2012, I'll be 27 when I start. I got my bachelor's at 22, went to work in industry, have been getting my master's part-time while working and building my credentials. In addition to boosting my GPA, understanding of my own research interests, resume, and CV, this approach has done wonders for my bank account.

Does the community college have some sort of committee on academic affairs that you could petition? It's rather rough that he would drop your grade so much over a misunderstanding, especially when he implied before that he wouldn't. And there being nothing he can do seems like nonsense to me. If he really wanted to help you, he could support you with the bureaucracy in the effort to get your grade changed. He gave the grade in the first place! If he doesn't think you deserve to have your grade changed, that's his prerogative, but he could at least be honest instead of pretending that he's powerless. If the school actually does allow grades for retakes to replace old grades, rather than simply showing both grades, then for all practical purposes you can erase the past. I realize that you'd rather get in this time, though, rather than next time.

So, there's plenty of detailed advice out there for NSF GRFP applicants. But I haven't been able to find much for other major STEM fellowships. There's Philip Guo's fellowships advice page (which covers NSF GRFP, NDSEG, and Hertz, but mostly focuses on the NSF) and the SMART Forum (for SMART fellowship applicants). But I haven't found anything specific to the NDSEG or Hertz, and I haven't found anything at all for various other major fellowships - the NPSC, the DOE CSGF, the various NASA fellowships, etc. These are big-deal fellowships that get lots of applicants, so it surprises me that there's nothing out there. Maybe my Google-fu is lacking. Does anyone know of any advice resources for the other, non-NSF, STEM fellowships?

I actually think Northeastern would not be a bad fit. It offers funding to some MS students - scholarships and assistantships. It is fairly easy to gear your curriculum, as an MS student, toward practice-oriented classes. It has a first-rate careers office, and allows MS students to do co-ops (where you work full-time at a company for a semester while maintaining your student status), which are very helpful for job placement. It is easier to get into than places like Stanford, and they like people with industry experience. Also, it is in Boston, and there are many good research universities in the Boston area, to which you could apply (making it more likely that you could be together).

I worked in my school's admissions office for food money as an undergrad, so I understand what you're saying about college admissions. There are two things to keep in mind, though. 1) Some of those perfect stats high-schoolers get rejected for a reason - they inflated that GPA by taking too-easy classes, or they come across as an arrogant little jerk in their essays, or they're obvious grade grubbers who don't appear to have an intellectual or personal life outside of doing well in classes and on tests, or they seem like they'll crumble the first time they get a B on a homework, or they chose their recommenders, poorly or they're just not a good fit with the university that they're applying to. They aren't all perfect applicants who got crowded out by other perfect applicants. Similar things can happen in grad admissions. An applicant with top numbers might get rejected because they had a bad recommendation, or the professor that they want to work with isn't taking new students, or whatever. So don't read horror stories of applicants with top numbers and lots of awards getting rejected and say "Clearly this means I have no chance!" 2) Grad applicants aren't judged by the same criteria as undergrad applicants, so it's apples and oranges. I did notice something about the quoted paragraph that I wanted to mention. You said that you're worrying out loud about how you're not going to get in, and that your family and friends are tired of hearing it. Are your friends also applying to grad programs? Are some of them, perhaps, less exceptionally well qualified than you? If so, then listening to you talk about how you won't get in could be really upsetting and demoralizing for them. And I'm sure that this isn't your intention at all! Most people are anxious during this process, no matter how well-qualified, and of course they want to vocalize that! I am just pointing out how it might look from their point of view, because I would hate for you to get through this process only to find out that you've accidentally alienated your friends and perhaps damaged the friendships.

If you are including MS students, I got into a solid MS program, with a Dean's Scholarship (merit scholarship for which only the top third of acceptees are considered) having had a 2.5 in undergrad. We will see how I do next year when I apply for PhD programs.

I don't know how top school PhDs are doing, but as a current job-hunter in, among other things, robotics, I can assure you that there are job opportunities in robotics right now for people with advanced degrees.

From the article: "Yet, when you ask graduate students and postdoctoral scholars what their individual experiences are, a science career is a very tough road with low pay and few career prospects." I can't speak for any other field. But right now, I am job-hunting for computer science research jobs at the sub-PhD level. And I see loads of jobs for CS PhDs (heck, I've even applied to - and been granted interviews for - a few, in the hope that the rest of my app is strong enough for them to overlook my lack of doctorate). And they aren't low-paying, either. Heck, the fact that I've gotten interviews for a couple of these jobs, despite not even having finished my master's, suggests to me that there's not a huge pool of people applying for them. If I didn't think the PhD was going to open up more job opportunities at higher levels for me, I wouldn't bother trying to get into a program. This article is way too focused on faculty positions, given that it's talking about sci/eng. Many research scientists/engineers are not in academia. There's national labs, DoD/DHS/DOT labs, FFRDCs, private government contractors, the big corporate labs (e.g. IBM Research, Microsoft Research), non-profit research foundations (e.g. Fraunhofer), cutting-edge startups, and normal private industry R&D. None of these are academia, and all of them have doctorate-level research positions in sci/eng.

This, or a variation of it, is not so uncommon. When I applied for MS programs, Northeastern had me do this for all computer science classes that I'd ever taken. I always assumed that programs do it because they want to see if your preparation was actually adequate for the program to which you're applying, and they don't trust course names alone.

Oh yeah, definitely this. I am constantly amazed at how many MS and PhD students in my current department have absolutely no clue of how to triage. Were their undergrad programs really that un-challenging? I wouldn't have made it through undergrad without developing serious triage skills.

This was for my MS applications - we'll see what I think next year when I'm applying to PhD programs. 1. Deciding where to go. I was choosing between a more prestigious department, with good facilities, that gave me a scholarship and was much more logisitically friendly (I'm a part-time student, so this means "Offered lots of classes after work"), whose environment I knew nothing about, and the friendly department where I did my post-bac work, that is less prestigious, and much less logistically friendly, but where I already knew that I could succeed and the professors liked me and had helped me along. I chose the former. I cried when I declined the offer from the latter, because I felt like I was disappointing them when they had done so much to help me. 2. Sending in the application. Dealing with transcripts was annoying, one of my schools didn't receive my GRE scores until three months after I ordered them, and the application websites were not the best-designed websites that I have ever come across (to put it mildly). 3. Preparing to apply. I had already done some of this (e.g. taking GREs) a couple of years earlier while I was still an undergrad. I already had a resume from when I had applied for jobs, and just had to update it. I was only applying to two programs, so it wasn't that hard to produce two SOPs. I knew who I wanted to write letters of recommendation, and all of them agreed and got their letters in quickly. 4. Waiting for replies. I had other things to do during that time, like working full-time, taking grad-level classes as a non-degree student, and getting repeatedly tested for a chronic illness. Both programs that I applied for, had relatively late application deadlines, so my waiting time was relatively short. 5. Moving. I didn't have to move, so this is not applicable.

Usually I've seen it listed as its own subfield. My undergrad alma mater, my current MS university, and the US News CS subfield rankings all consider it to be its own subfield. It can overlap with other subfields, of course. For example, in my current department, parallel programming is listed as both systems and theory.

Well, no. It does, just not necessarily in the way that you think. PhDs.org will let you rank programs according to what percentage of their PhD recipients have an academic job offer at graduation. That seems like a ranking that you would care about a lot, because that is the sort of job that you want. If one program only places 10% of its grads into academic jobs, and another places 40%, you would probably rather study in the latter program than the former program, even if the former is a bigger name. There are many reasons that one program might place more people in academic jobs than another. Some of them are linked with prestige (famous profs with good connections to write your job recommendations, more/better resources to enable your productivity) and some are not (focus of the department on preparing students for academia vs industry, availability of departmental or university-wide job-search training). And clearly, even the ones that are linked with prestige are not entirely tied to prestige (a less-prestigious department can still have really good funding, or good facilities, or a few famous profs). But really, it is all about the outcomes. Some programs have much better outcomes, from your wanting-an-academic-job perspective, than others, and you want to identify which are which. It's just that outcomes are not solely about which program has the bigger name.

Don't buy into the myth that in order to be a serious scholar you're not allowed to have a life outside of school.

It's good to be focused, but you do run the risk of shooting yourself in the foot if you're too specific. I don't think it's wrong to talk about ideas, but I'd state them as possibilities, emphasizing that you are open to other possibilities, rather than what you definitely plan to do. You can demonstrate your potential by talking about your track record - what other research you have done, what useful skills you have learned, etc - as well as what you want to do and how this fits with the department. Also, make sure that you don't come off as only wanting to work with one prof in the department. What if that prof is about to leave the program, or go on sabbatical, or just isn't taking any new students for a while?

Yeah, cost of living, plus I think our food restrictions really did hurt us. "Oh, you want all of your food to be gluten-free? And it has to be actually gluten-free, including during preparation, not just free of foods with wheat in them, otherwise the groom will get violently ill? And you want such food to be provided and served by a reasonably-priced company? Haha, suckers." We spent 10K on catering alone (including serving staff and rental of tables, chairs, plates, and flatware) for about 100 guests. We looked for ages for a cheaper place that could do what we wanted, thereby annoying the fancy-pants (but really good) caterer that we eventually ended up signing on with, and never found one. Let this be a warning to anyone planning a wedding who requires certain kinds of disability accommodations. Other things that cost us: - Alcohol (open bar) was around $1500, not including the temporary liquor license. This was by far the best deal that we could find and had been recommended to us by everyone. And we got to keep the opened-but-not-finished bottles, which have provided for a number of subsequent bar nights at our apartment. - Venue was $2800 (for eight hours - we had both the ceremony and the reception there). We were restricted to places that let you pick your own catering company, which eliminated a lot of possibilities. We saved $1200 by having the wedding not-on-a-Saturday (they gave us a discount). - The wedding rings cost a couple grand or so. - Tent rental (for the reception - there was no room in the venue that was big enough to fit tables and chairs for 100 people) in case of inclement weather cost somewhere in the one to two grand range. - The various licenses and permits required (event, alcohol, etc) added up to close to a grand, IIRC. - The gluten-free chocolate-and-fresh-raspberries wedding cake cost about $500 (and was worth every. single. penny.). This was a splurge - it wasn't like we actually required a cake to have a wedding - but it was so worth it. We really did try to save where we could. Less than $1K total on clothes, all handmade decorations, free donated flowers, did our own invitations, borrowed audio equipment from friends, music via iPod, had a backyard bbq for the rehearsal. There just wasn't much else we could do, though, without sacrificing the comfort of the guests (we weren't going to invite a bunch of people, many of them out-of-towners, to a wedding and not feed them or give them alcohol). The primary way we could have saved would have been to invite fewer people.

I'm into AI (both theoretical and applied), especially neuro/bio-inspired computing - right now I'm working on a neuroevolutionary algorithms project for school. Actually, I'm into a lot of things, but that's probably the biggest one. I work full-time for a famous government contractor as an imaging scientist, on anti-suicide-bomber tech. Before that I worked as a scientist in machine learning and computer vision at a small government contractor for a few years. My MS thesis (which I'm writing next year - I just found an advisor) is going to involve legged robot locomotion in uneven terrain.

As a current job-hunter (I'm a part-time MS student who works full-time and plans eventually to go for a PhD - and my contract at my current workplace is about to end), I agree that this is pretty big in industry right now.

Very few schools have a separate machine learning department. Your question is so vague that it's hard to answer, honestly. These programs differ in all sorts of ways - number of students, stipend level (and local cost of living), job placement rates, percentage of graduates who end up in academia, [in]formality of the culture, emphasis or lack thereof on interdisciplinary work, requirements to get the PhD, average time to completion, quality of the facilities...I could go on and on. I suggest creating your own customized ranking at phds.org. The person who is going to have the greatest impact on your experience is your advisor. It is a mistake to look only at a bunch of top departments and say "Well, that is where the cutting-edge research is, so that's where I will apply." Cutting-edge research can happen just about everywhere - most programs have a truly outstanding faculty member or two. Look back at papers that you thought were cool, and find out where the people who wrote them are. You are going to want to apply to some non-top departments as well as the top ones, and you might as well figure out what those are. I can tell you that if you have no research experience you have very little chance at the top programs. A lot of them (CMU, for example) use that to screen people. However, there is such a thing as a final project for a class that is also research experience (I got a conference poster presentation out of a term paper a while back). Obviously, since I don't know anything about your final projects, I can't judge them.

This is not how you want to do it. You don't want to know that there exist people who attended this program and ended up with shiny jobs. Well, you do, but that's not the main concern here. You want to know how most people are doing, because unless you have a very good reason to believe that you are unusual, you should never assume that you will be the great shining exception of the department. If the top 5% of the recent graduates end up with absolutely amazing jobs, the next 15% end up with decent jobs, the next 20% end up wtih crappy jobs, and the remaining 60% end up with no job, you probably don't want to attend that program, no matter how fantastic the jobs of that top 5% are. So, what you want to know are things like: - What percentage of graduates have jobs lined up upon graduation? - What percentage of those jobs are in academia? (Even if you don't want to go into academia, this can give you an idea of where the department's focus is.) - What percentage of graduates have jobs within a year of graduation? - What percentage of graduates end up in which sectors, industries, or companies? If all those check out, then you can look at the shiny jobs that the stars get (it might also be worth it to find out what their credentials were when they got those jobs, so that you have something to aim for). But any department that's just giving you anecdotes of the greatest recent success stories is misleading you (whether intentionally or not).