Desi_Mama

Anyone?

I consider myself a very mathematically bent person (theorem-proof basis) and my primary interest is in convex optimization. An exciting application of the same which fascinates me is machine learning and equivalent names in other fields. I would like to pursue a PhD in the same and would greatly appreciate suggestions regarding which department is best suited for an optimization+data work. I have a bachelors in mechanical engineering with industrial engineering focus. Hence all the above departments are new to me, barring OR. I have of course done courses on optimization, algorithms, stochastic processes, introductory machine learning, networks etc. In CS, the only areas that interest me are artificial intelligence and algorithm design (combinatorial optimization). In EE, I am interested in signals and systems. In OR, I would look to focus on mathematical optimization with applications in data. After PhD, my intention is to work in industrial research labs (like Microsoft, Google, IBM, Yahoo etc). I know that majority of candidates here are CS and EE grads. How OR and Stats fare in job market? Would I be eligible for the same jobs regardless of what I do my PhD in (among the departments listed). I am personally most comfortable with OR, followed closely by EE due to large overlap with my undergrad coursework. I may find it hard to get into the *top* CS programs, whereas I might stand a much better chance for their IEOR or EE programs. Do share your thoughts

I completely agree with you that things should be sorted out on a case-to-case basis. I just wrote the previous post to articulate my experiences if someone else stumbles across this thread in the future. My objective was to simply point out that there is a world of a difference between the two scenarios: deferral with guaranteed funding and deferral without guarantee of funding - whatever be the terms different programs use. (I gave one example that I know) In the first case, its quite likely that the student should make a strong commitment and everything that follows from that commitment. In the second case however, I don't think there is any commitment involved. I don't agree that a student should commit to a program, while the program does not commit to funding. When the program does guarantee funding (likely only after reviewing next batch of applications), the student can make the choice. But till that commitment comes from the program, it is unreasonable for the program to expect that the student should not take any other initiative (i.e. apply to other programs or scholarships). I just wanted to point out this distinction. But I agree with everything else you said. In particular: Deferring offer just for sake of a safety school is wrong. Deferring multiple offers is wrong (unless of course both programs don't commit to funding, which basically reduces to my argument in this post).

@juilletmercredi: I by and large agree with what you say and my plan is to not take up the deferral. I want to keep my options open, but will very likely keep apply again this December. However, there are some points on which I hold different opinions, and I'll put them here for anyone who stumbles across this post in the future. IMO, deferral is not a commitment to attend next year, but is akin to asking for additional time to decide. Let me point out that there seem to be two different types of deferrals. The terms the university used were: "deferring enrollment" and "deferring admission". I believe "deferring enrollment" means that you accept the universities offer and funding package, but will enroll at the university 1 year later. I believe this was what you referred to in your post. However, from what I gathered in this forum and elsewhere, this is very rarely given to a student. This is precisely for the reasons you mentioned - student in most cases is not sure if he/she can attend next year, and the program is not sure about the funding situation next year. In most cases, what is granted is deferred admission which is basically a guaranteed admission. You need not re-apply next year, and the admission is guaranteed. However, funding package is not guaranteed (likely, but not guaranteed). This was what happened with me. I don't think this is a commitment to attend next year at all, since the program did not even ask me to give anything in writing. Basically, my admission is guaranteed, and my application will be rolled over for funding assignment around Jan/Feb (I have to compete with Fall 2016 applicants). I was told that I will get an acceptance letter with funding package (if allotted) sometime before April of 2016, and I can accept it at that time. In the above scenario, I think the student has every right to apply to more programs. Firstly because funding is not guaranteed. Secondly, the student never agrees to attend the university in the first place - even the university wants me to accept only around April. Note that this gives an out option to the university too - if I don't compare well with Fall 2016 applicants, they are not obligated to give me funding, which is basically a rejection. Just to reiterate, I will be declining the offer. However, I don't agree that one can defer only if he is willing to commit to the program. This is on a case by case basis between student and the program. In most cases I have seen, the university doesn't commit to funding before they see the next batch of applicants (I agree that this is due to bureaucracy associated with funding allotment process, but it doesn't take away my point). In such cases, its perfectly reasonable and shrewd to apply to more programs. Even if they guarantee funding, grad programs give time till April of next year to decide - to attend, not to attend, where to attend. Also, I don't agree that deferring takes away a spot. If someone decides to attend another university, they will let the first program know before April, so the spot will go to someone else. So, I don't really think there is any "ethics" involved here as some people think. In reality, when you ask for deferral, there is no commitment - neither from the student side to attend, nor from the program side to fund you. However, I agree that it doesn't make much sense to defer an offer unless you really like the program. I am not too fond of the program, and hence my reason for not deferring. There is no moral or ethical conflict at all.

Thanks TakeruK, I agree with you. It doesn't make much sense to go through all this hassle. I'll let the program chair know soon that I will just reapply, but I need to do so very politely. They have put up with all my nonsense for sometime, so I am really indebted. It appears both myself and the program want to keep the options open. I can basically say that I applied to more programs since funding wasn't guaranteed, and after talking to faculty, found another school to be a better research fit. If they don't want me, they can simply say funding is not available, and I'll certainly not attend in that case. This problem is basically set up for a disastrous Nash equilibrium. I totally agree that it's best for all parties to simply reapply. Thanks for your inputs. Also, just another quick question. Do you know the difference between the following terms: "deferring enrollment" vs "deferring admission"? Basically the program chair said I can't defer enrollment, but I can defer admission. I assume the former means guaranteed financial support. Is that a fair assessment? If so, then it basically reinforces my previous point about the program wanting to keep their options open as well. Cheers

Funding is not guaranteed now. Basically, I will get information sometime before April 2016 if funding is available or not. (I will be compared against class of Fall 2016 for funding). What I meant was, suppose if I ask the program whether it would be OK if I apply to a few more programs as back-up option (the reason I would give them is lack of guaranteed funding); the program chair after hearing that request may actually agree to defer + guaranteed funding, which is not ideal for me. I don't know if this scenario is actually possible, but just saying. Do you think I should go ahead and ask anyway - something like: "since funding is not guaranteed, and I also want to keep my options open, is it OK if I apply to a few more programs as alternatives?" Also, does one need to explicitly ask for permission before doing so? I mean, they are not guaranteeing funding, so is it not reasonable for a student to explore alternate options? If someone isn't active in looking at other opportunities, it may be possible that he/she may end up with no other offers and no financial support from the deferred program, which will be quite sad. Though the probability of such a scenario might be small as TakeruK noted, but it is certainly not negligible.

I agree with you that asking if they are OK with me applying to more programs is a double edged sword. On one hand, it is possible that they may be OK with it, in which case its almost close to an ideal scenario for me. However, when I ask them, if they say my funding will be guaranteed, then it will look very bad to pull out. I guess I'll just tell the program chair that I have an internship lined up, which added with some family and financial circumstances, seems a better option. I'll also let them know that if I feel like I can start by Fall 2016, I'll apply by the December deadline. Thanks for your inputs

Taking advice from here, I decided not to defer both the options. I am planning to defer only the program which I have currently accepted. I inquired about the process (not with the program chair, but with an admissions assistant/secretary), and they told me that it is possible to defer the offer, but financial support will not be guaranteed in such a case. Basically, as TakeruK pointed out, I will be compared with the Fall 2016 application pool - based on which funding will be allocated (fellowship, TA, RA, or none). I currently have a TA/RA, so I am not confident of an "upgrade" to fellowship. Not sure about the possibility of funding either - they say it is not guaranteed, but I would likely be given funding. Any suggestions? If funding is not guaranteed, should I go ahead and ask them if I can apply to more programs (which would be ideal in my case )? Or should I drop it altogether, and say I will apply again next year?

Thank you. I will be straight up honest with the program, and ask what the possibilities are. If they ask me to reapply, as you said, I have no other option but to go that route. I only hope that they don't "punish" me by not taking me again - though this seems likely since they would figure out that the program is sort of my "safety" option. I just have to live with it I guess. Wish me luck for the next year

Thanks for the inputs. Given the circumstances, the intern is better than going to a university and try to transfer. That introduces too many new variables and a lot of anxiety. Also, the intern is reasonably well paid, and I get to work with some scientists whose research profile can match many professors in top 20 universities. Also, I can get a recommendation letter from them, unlike in the case of transferring. Also, I feel that a major reason for why I got rejected by many programs (I applied to 12, got into 2 PhDs and 1 MS ) is no recommendation letters from people with PhD in CS. All my recommenders either had a degree in MechE (2 of them) or Industrial Engineering (the third). I can maybe use this year to rectify some deficiency too - do a discrete mathematics and operating systems course online etc. So I am pretty set on the intern position. I am either looking to defer offers and apply to a few more programs (this would be ideal), or just drop the offers and apply afresh (this scares me a bit, but if it comes to that, I have no other option I guess).

We are friends. I borrowed his computer, and I guess he had automatic sign-in. Forgot to log-in to my account My last post in a nutshell: I am switching majors from MechE to CS. Got into a top ranked MS program but not funded, so couldn't afford (some vague promise of partial funding was given, but I wasn't convinced). Got into a PhD program in top 25, but I feel I can do better since I performed well in my BS (high GPA, one paper). I have an RA job in an R&D lined up (temporary paid intern) which I think will improve my odds. Want to apply again next year, but don't want to throw away the offer I have (what if I get nothing next year?). Essentially I want a safety net and I am not sure how to approach it. How do I go about tell a university that I want to keep my options open? If I tell them that I want to make a decision by say March next year, will they not agree for funding? I understand that it won't be *guaranteed*, but are there good odds that I will get funding if I indicate interest at that time?

Can you suggest any alternatives? I don't want to commit to a program at this stage, but at the same time I don't want to let go of the offer. Would the program be OK with it if I applied to other places? Do I need to explicitly ask permission for the same if I defer?

Hello everyone, I was wondering if someone here could suggest strong universities in the area of machine learning. This will help all future applicants as well. To put my background in context, I did not do by BS in computer science, but rather in mechanical engineering. A better description of my coursework could be "operations research" since all my courses and projects were heavy on the optimization and networks side. I found the area of machine learning and developed my interest in the same at a late stage, so wasn't able to change majors to CS or to take courses in CS to cover breadth and apply for PhD programs in CS. I did manage to get into a few mid ranked PhD programs and the MSCS program at Stanford. After some deliberation, I have decided not to take up the latter (too costly). The PhD programs are not particularly to my liking, and I feel I can do better with one additional year of CS related research. I will be joining a very well reputed R&D lab (along lines of MSR, IBM etc.) as a research assistant (intern) for this year for a project related to convex optimization and machine learning. I am planning to either defer my PhD offers and apply to other programs, or decline the PhD offers and simply apply again this December. Can you suggest some universities which are very strong in machine learning? I am more inclined towards the math part (closer to optimization) as opposed to applications, though I would be OK with that too. I fear the top programs like Stanford, UCB, and CMU might be out of my reach (I will try none the less), but are there any other universities particularly strong in ML? Please let me know. Also, if someone is aware, can you suggest some Operations Research programs strong in the area of Machine Learning. So far, I have been able to pin down only Columbia University. Following is based on my research thus far: Tier 1: Stanford, CMU, MIT, UC Berkeley Tier 2: U Washington, UIUC, Univ. British Columbia etc. It would be great if more people can help fill the list. Thanks

I was in a similar situation thinking about MS CS from Stanford (which means loan of ~$50K) and PhD from a lower ranked (albeit top 15) school. I was wrestling with the decision for a very long time (well into May) and finally went with the latter. One reason was the money (which in my case wasn't a big problem since my undergrad was free ride). However, there are much bigger reasons you need to worry about: You will not get funding for MS. I talked to many people who went this paid MS followed by PhD route who were totally disgruntled. The logic is simple, if they had money, they would have taken you for PhD! Don't expect to get any funding. If you are super lucky, then maybe yes. But its wise not to count on it. You may get a TA for a semester but good luck making time for research with courses and teaching duties. You may not get a good adviser - This to me was the clincher. As an MS student, its very hard to find a good professor for RA. You may want to work with a star faculty. Being an MS student you walk up to him, and ask for RA. There will be a queue of PhD students with fellowship waiting to work with him (meaning he doesn't have to pay). Why should he/she pick you over them? You may feel that you are a better fit or a better student, then why didn't you get into the PhD program? Such questions will come up, and it is very unlikely that you will work with the top 2-3 faculty of your choice. In short, you spend $50K, do sub par research with a not-so-good faculty, take courses, and do teaching while trying very hard to make time for research. In short, this is a classic recipe for a horrible grad school experience. My recommendation: If you are interested in PhD, don't go for an unfunded MS, even if it is from MIT/Stanford/Harvard etc. (whichever is the top school in your field). You are likely to be much more succesful and happy going to a PhD program that wants you (i.e. funds you). I highly recommend that you apply again next year, to possible a better mix of universities. In the meantime try to gain some research experience in a university or a setting appropriate to your field. Without any courses and other distractions, you may be very productive in the next 6 months or so before applications are due which can improve your chances.

Hello, The short version of my question is: can I defer an offer and still keep my options open? My scenario: I have offers from two programs, but I would like to delay start of grad school to Fall 2016 due to some financial and family circumstances. I was wondering if I need to commit to a program if I am asking for a deferral. What I expect: Defer offers from both programs and make a decision by April 2016, just as a fall 2016 applicant would do! My reasons: Professor's funding scenario might change - In both the places, I am interested in working with only 3-4 faculty members. I am not comfortable with committing to start in Fall 2016 without knowing if they would have openings in their group, and whether they are as interested in working with me, as I am with them. I just had some passing email conversations with them which went like "come here, we can talk, and then decide about positions in my lab". I guess if I ask them for Fall 2016 information now, their reply would be something to the effect of check back at a later stage (say December, January). My opinion is that it is not wise to commit under such uncertain circumstances. My interests might change - I will be taking up a job at a reputed research lab (along lines of Microsoft, Google, Bell, IBM etc.) as a research assistant (intern). The work would entail doing research and possibilities of publishing in top conferences or journals. Since I would be working there for at least 6 months before April '16, my interests could change a bit (I am certain about an area which is quite broad - but I may sway more mathematical, more computational, practical, applications in biology etc.). In such a case, one program may be more appealing than the other, and I would prefer to make that choice towards April. There are other universities too - Though I am happy and content with the offers I have, there are a world of other options too. This in some sense relates to my previous point. If I develop a keen interest in some niche, there may be other universities with great expertise in that area which may be worth looking into. Committing to a program now will psychologically limit me from exploring. A direct extension of my previous point - if I commit to a program, I would be reluctant to explore topics outside the universities' expertise. This may not be feasible since the problem I get to work on would most likely be picked by my employer. Is there any way around this? I don't want to let go of all my options and apply again next year. Also, if I turn down these universities, I am not sure if they will admit me again. The universities under question are very good research fits for me without a doubt, but I wish to keep my options open, not be confined, and be allowed to explore. If someone has good suggestions or have encountered a similar situation before, please let me know. Thank you.

Thanks for the heads up. Just to expand on it a bit further, based on your experience, what is the likelihood of an MS student getting RA/TA support at Stanford? I am not greedy, expecting support for the entire program (though I would absolutely love that!), but what is the likelihood for say 2 or 3 quarters of funding? Have you personally seen MS students funded for a good portion of the program? Thanks a ton for your inputs.

What are your research interests? I am no expert at this, but I was admitted to the USC CS PhD program, and hence researched the school well. I also talked to a few faculty from EE (my interests are at interface of AI and systems theory, so EE faculty were excellent options too), and I really liked the place and department. Among the three, I feel USC might be a better option, if you are also interested in systems/data/optimization and the likes. Columbia is also an excellent school, but I would personally choose a university in California, and would recommend the same for all Indian students

If I may ask, based on your knowledge and experience at Stanford, what is the probability to transfer to the PhD program after the MS program? If I have to re-apply, does the MS degree from Stanford help? In my case, I think I have a much better chance than the norm, since the professor I spoke to seemed very enthusiastic. To paraphrase his words when we talked, "I would be happy to take you, but I don't have enough positions on funded projects at the moment. It is possible that more funds may arrive between now and Fall, but I do not have a good estimate of the probability at this stage. But I would be happy to discuss research possibilities with you when you arrive, so that you can start research right away." The sense I got was that he is definitely interested in taking me as his student, and I would love to work with him too. My B.S. thesis was primarily an extension of his work, and I even managed to get a paper through to a top conference. What worries me is this scenario: I don't get enough financial support (no RA/TA support for even 1 or 2 quarters) which puts me with about $50K in debt, and also not get into the Stanford PhD program, and I have to move to a not-so-prestigious university for PhD. I am not able to estimate the likelihood of such a scenario, and hence my anxiety.

Thanks for the reply. I have not been given anything in writing, but when I contacted a professor at Stanford, he seemed very enthusiastic since our research interests are a great match, and I had a project which can be considered extension of his papers. He said he is open to taking me, but does not have funded positions at the moment, but *might* get more funding by Fall. I have lost negotiating power now since I have already accepted the offer from Stanford. Given the options, it was always a choice between joining Stanford now vs joining Stanford a year later. My only concern is, I don't want to start a PhD with debt. If I decide to join Industry after MS, I think the debt will be easily paid off in 1-2 years. PhD is a totally different ball game

To explain my situation: I am an fresh out of school Indian student from one of the top NITs. (top 15 school in India) I have a B.Tech in Mechanical Engineering, but my heart lies in data analysis, and overlapping areas like AI and signal processing. As for future plans, I see myself doing research intensive work. I did a summer internship at a top Canadian University on image processing, and really liked the experience (even got recos and an offer from there). Keeping these in mind, I think Stanford gives me an excellent launching pad, and getting the offer is a big achievement (at least from my perspective). My only concern is the steep cost. Considering it is $90K, I have to take out a loan for minimum $40K. Though I feel that it is quite possible to get TA/RA for one or two quarters, I expect the total cost to be upwind of $60K. In the event I decide to join industry after MS, how long would it take to repay $60K? What is an average salary I can expect after MS in CS from Stanford? If I decide to do a PhD after my MS (hopefully at Stanford or another top5 university), the interest rate would pile on. Would a good research position (and perks that come with it like job satisfaction etc.) after PhD offset this high initial cost? My alternate option is a reasonably good job in India, which I am happy with (though the pay is just above average), but not very excited. I can save up some money, though not much and attend next year; or can work for 2-3 years and see if I want to directly go for a PhD. The drawback with this is that, if I decide to go for an MS or PhD later on, I am basically wasting 2-3 years, at a mid-level job. I enjoy research and am quite confident of my ability, since I was accepted for PhD at Purdue and USC with RA. The reason I chose Stanford was, well its Stanford, and I think an MS from Stanford would open more doors for me. In my case, I was not sure about a 5 year commitment, primarily because I am switching majors and my interests are not well defined. If I get some chance to do research at Stanford (RA or independent study with prof/postdoc) and really like it, I plan on getting a PhD. Otherwise I would be joining industry either in USA or India. If you know about chances of TA/RA positions at Stanford, kindly share your inputs. If I can get assistantship for even 2 out of the 5 quarters, I would be set to attend (ofcourse, more the better). My father recently made a ton of money by selling some old stocks, and has agreed to fund me. This should quite likely cover 3 quarters, so it becomes a sort of zero sum game if I can get funding for 2 quarters.