2

I'm studying Master in a school where it's required to publish 1-2 papers in conferences. My lab focus on computer vision/machine learning but I'm not so good at mathematics. I tried to learn but I'm quite overburdened because English is not my native language.

I'm quite good at engineering, coding but not in comprehen math works (the more symbolics & less natural language interpretation --> more scary for me, I don't understand). I don't think I like to be a researcher much because I recognize that every paper is just a node of a huge knowledge tree. Sometimes to understand one node (one paper) I have to also read a lot of dependency papers too. The more I read, the more I get confused and messed up. (Please not advise me to drop out, my parents and siblings always push me for that).

So I have 2 questions. Is there any chance to get published with only engineering work in machine learning? (Not something too academic/scientist much with math work) And can you recommend me some conference in CS with some more focus on engineering/application using existing techniques ?

Thank you very much.

4
  • Find a conference / journal that has published related work to that in your paper and submit.
    – Solar Mike
    Commented May 19, 2019 at 5:57
  • Submit to the same venues that you cite.
    – JeffE
    Commented May 19, 2019 at 17:52
  • @JeffE is that increase the chance to get published ? Commented May 20, 2019 at 4:57
  • 2
    Every research community has its own social norms/expectations about novelty, intellectual rigor, mathematical sophistication, citation patterns, authorship, notation, terminology, coverage of past results, assumed expertise of the audience, preference for theoretical versus experimental work, and so on. I assume you are reading and citing the types of papers that you want to publish. If your paper follows the same norms, it is more likely to be accepted in the same venues.
    – JeffE
    Commented May 20, 2019 at 7:19

1 Answer 1

0

The three primary ways of finding a community and venues that suits you are:

  • talk with people familiar with communities that might be of interest to you (will be heavily based on luck and talking with a lot of different people about their work)
  • find publications that make the sort of contributions that you want to make (requires broad reading of the literature in your field, and possibly beyond it)
  • find venues that sound like they are on-topic for you and read papers from it to see what variety of papers seem popular there recently

In short, there is a three-way link here: People <-> Publication <-> Venue

Find one and you can work your way to the others. Find a publication that fits what you want to do, and you can find the venue it appears in, as well as the people involved with it. Given people, you can find what other papers they publish and where they publish them. Given a venue, you can find what sort of papers are published there, and who is responsible for those papers.

All of these are in a constant state of change, so what contributions are valued, what work people do, and what are top/good/ok/bad places to publish changes at least yearly. Ideally you would have local contacts (advisor, other professors, fellow researchers) who can help you, as the search space is so large, but sometimes you are (or want to be) in the minority and so what works for most people near you won't work for you.

Specific to engineering, you should be aware (and look for) the fact that not all engineering contributions are the same, either. Some focus on novelty, technical difficulty, real-world application, industrial applicability, actual deployment, etc. You'll discover that as you read through papers.

I should also clarify that math is not a short-hand for academic or scientific work. While math is often a part of papers, there are many types of paper where you could remove the math completely and the contribution would be the same. Sometimes math is part of the key contribution, sometimes it is sprinkled on for effect (like a garnish), sometimes it is just for clarity, and sometimes it is just because it is standard to include some definitions/equations but they aren't otherwise important. Learning to distinguish them is just part of the process of reading through them - though I suspect you are getting hung up on the fine details before you get to the general big picture, which will actually make it harder to figure out what is going on.

I would suggest you try to read the papers in a less detailed way on the first pass or two, and basically replace the math with this: input -> (magic happens...) -> output. First understand the general gist of what purpose the math is to serve, and then you can go back and look at the math more closely when it seems relevant - but most papers won't be relevant enough to the task at hand to bother with that level of detail. It can also help to join a reading group so you can get the experience of seeing how other people talk about the papers and work through them. This is especially important with unfamiliar math, as there are many natural language ways of talking about the math that makes it make sense, when you would never get that from only reading the paper itself.

Eventually you will be able to find papers and say, "hey, this is the kind of paper I want to write and the work I want to do, only instead of this detail here I would instead like to do this other thing..." - and there you go. Since you have less time to do this than a full PhD, I suggest you be a bit less picky and instead focus on the lower-hanging fruit of being closer to the work your group does, or at least focus on papers you feel are more intelligible to you and settle for an OK fit.

Finally, don't be afraid to ask for help from those around you, that is part of the process and point of being in a program. Well, you can be afraid, that's fine - just ask for help anyway. You will not understand every last bit of what you do, much less what other people do, and that's part of the work - don't worry about it. Much of the goal of the field you are in is to try to make advancements even though the level of complexity has long gone beyond what any one person can understand fully, even with many more years of experience than you have. Do what you can, muddle through, and accept that you will only understand a small percentage of what there is to know.

2
  • You said that "instead focus on the lower-hanging fruit of being closer to the work your group does", for the case the existing researcher are leaving after graduation, how do I know which direction is still have some weakness (some hole, gap) to exploit (improve) (of course base on that reason to) publish? Commented May 23, 2019 at 8:46
  • @chickensoup Basically by talking to them and your advisor and asking them about their experience and for advice. Ideally they would offer some useful, informed next direction steps, preferably in more detail than you could find in any paper, especially if that's not something they will be working on themselves - sort of "pick up where they left off". Won't know until you talk with them, ideally you would be able to come up with multiple options so you can pick one that seems better for your goals and skills. It doesn't always work, but in many groups its the best strategy.
    – BrianH
    Commented May 23, 2019 at 12:50

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .