3

I debated between asking this question here and in the Academia stack exchange, but decided Computational Science was the most capable of answering the question.

I'm a new graduate student (~1 semester now) and would really like to make a push to be the best I can in my field. My undergraduate was a top 5 engineering school. However, my graduate school is top 50. So I'll need some really good research to get noticed. I have an excellent advisor who is very supportive of whatever I'm working on, so I can somewhat guide the "kind" of research I'll be doing further down the road. And I also have excellent access to any kind of computational resource I need.

Right now I'm doing molecular dynamics simulations and am spending time learning about quantum monte carlo on my own. I also collaborate with an experimental researcher, thus my work has a tie-in to experimental results.

I've been reading journals in this field, but I'm not sure whether "huh, this is an interesting article" matches up at all with what other researchers consider a top-tier contribution. Again, I'm still new at this.

My question: So, as of early 2013, what are the cutting-edge, hot topic, highly-publishable research areas in computational science? I really just need a starting point so I don't go off in a rabbit-hole of research that is interesting to me but that nobody else cares about.

The Hiary
  • 251
  • 1
  • 9
Nick
  • 359
  • 1
  • 8
  • 2
    Nick, welcome to SciComp! Your question is a really good one...that I'm not sure is a good fit for this site, because it's broad (highly-publishable research in computational science spans a lot of disciplines that intersect with computational science), and subjective in a way that may invite short answers (exascale!) when your question really demands long answers. That said, subjective questions can be really good (see good subjective, bad subjective), so I would like to hear from the community about this question. – Geoff Oxberry Mar 12 '13 at 06:31
  • 9
    I don't think there is a meaningful answer to this question. – David Ketcheson Mar 12 '13 at 09:23
  • @David, what do you mean exactly? Any sort of input would be useful for me. Even if it's suggestions of things to not research. – Nick Mar 12 '13 at 13:46
  • 2
    I think the issue is that computational science is utilized in such a wide variety of fields, that there isn't one cutting edge, there are dozens (if not hundreds). – Godric Seer Mar 12 '13 at 14:45
  • If you want to become a top scientist then you will actually on day create (or at least 'reinvent') a new field of research yourself. Thus this field is unknown at present. If you are pragmatic and want to enjoy quick success, reading research journals to identify the hot fields and publishing related work will be what you do. If your first preference is to become famous then science is probably not for you. – highsciguy Mar 12 '13 at 15:39
  • A list of hot topics in comp sci might be useful nevertheless. Though it will be very diverse, if it is good, and very short lived at the same time. – highsciguy Mar 12 '13 at 15:46
  • 3
    On a side note, I'd like to mention that I've become a little bit frustrated with Stack Exchange lately and what now qualifies as a "question". This is a recent trend. I believe my last 5 or so questions on SE (very roughly) have been deemed "not a good fit" or some other strange term. Honestly, there's no where else to ask the questions I have to the type of audience I want to ask them too. I wouldn't ask these questions if the answers weren't going to be of great use. I will of course seek out many other avenues for advice, but the collective online community – Nick Mar 12 '13 at 15:54
  • of experts in their field is a great resource, and it's unfortunate that I am unable to learn things from these people because of some anonymous person's downvote-happy habit or OCD moderation trends. – Nick Mar 12 '13 at 15:55
  • It is generally advisable to go to some research seminars in one of the Top 10 institutions and simply listen to what is being discussed... Mind, this is the kind of question fit for Academia SE, not Scicomp. – Deer Hunter Mar 12 '13 at 18:57
  • 1
    @Nick: From a moderator's perspective, what can happen with some of these types of questions is what happens with this question and this question. Voting becomes a useless way to determine answer quality; so many people answer, answers start to repeat. Everyone jumps in with any old answer they feel like. It becomes frustrating from a mod perspective as well. – Geoff Oxberry Mar 12 '13 at 22:13
  • 4
    @Nick: In any case, I have left the question open because I do think there should be some place for some subjective questions on the site, for precisely the reasons you mentioned. I agree with DavidKetcheson here and question whether there is a meaningful answer to this question (at least as far as where computational science is concerned) because it is broad and highly dependent on opinion. Even funding agencies don't necessarily give clear answers to this sort of question (replace "hot topic" with "research the agency wants to fund"). – Geoff Oxberry Mar 12 '13 at 22:18
  • 1
    @Nick: The underlying question here ("How do I picke the best topic of research") takes a book to properly answer. What you're looking for are the best references for help, not short answers from the members of this site (everybody's experience is different, and we don't know enough about you or the future to give you useful information beyond what others have already taken the time to usefully write down). From my own personal experience, I'd recommend that you find a group of high-calibre people that you can work well with, and work with them, but that's pretty general, isn't it :) – Aron Ahmadia Mar 13 '13 at 13:02
  • I'd like to suggest turning this question into a community wiki. – Paul Mar 13 '13 at 14:30
  • The fact that this question is tagged "molecular dynamics" is pretty funny. – David Ketcheson Mar 14 '13 at 17:37

2 Answers2

10

This is not an answer to your question. But if a student came to me asking this question, I would ask him or her to read these articles:

And a warning: jumping into a "hot" subfield as a student is very dangerous if your advisor is not at the front of that subfield. You're very likely to get scooped or end up doing something irrelevant.

David Ketcheson
  • 16,522
  • 4
  • 54
  • 105
7

First, an attempt to just answer the question. My (personal and subjective) feeling is that many jobs and funding opportunities right now are tied to applications in biology and energy, and that the hot research topics are in "multi-scale" and "multi-physics" phenomena, as well as uncertainty quantification. However, in your case you need to know what will be important in 5 years or so, which is much more difficult.

Second, some ideas of where to go to find your own answers to the question. You could browse some listings at mathjobs.org for an idea of what kind of specializations employers are looking for, and nsf.gov to get an idea of the grants and funding opportunities that are being offered. You could also browse some recent issues of SIAM Review to try to see what research is attracting attention. Big conferences (for example the 2012 SIAM annual meeting) often have two type of invited talks: established researchers surveying a career worth of contributions, and young rising stars talking about their exciting new discoveries. The latter are working in areas that are "hot".

Finally, some advice not directly answering the question. We do not know what the hot topics will be in 5 years, and we don't know what the funding landscape and job market will look like, and most people who expect a research-heavy academic job do not get one. I recommend working hard on some research that is interesting to you, building teaching experience, and at least considering what will happen if you need to look for a job outside academia. My recent experiences on the job market depended much more on my teaching experience and my programming experience than it did on whether I was in a hot research area.

Andrew T. Barker
  • 677
  • 3
  • 10