Guidance with publishing a conferences paper alone

Question

I'm going to try and make this as concise as possible. I need to recap my education quickly. I finished my bachelor's in EE 2012. Between 2012-2016 I was a journalist and learned to code and started a software company. I finished a one-year CS conversion master UK 2017 and now I'm doing a CS masters in NYC. I'm trying to do research in NLP and focus on low resource languages.

I'm trying to publish a paper or two in a conference but no idea how. I tried contacting professors in my department and they don't seem to be interested in working with master's students. I'm trying to apply for research internships in the States but it is very competitive and just got rejected because my background is very different and lacks research experience as I didn't have any chance to do any research in my undergraduate because that just doesn't happen where I come from.

That is why I think my only option is to publish things by myself. I can come up with solutions to problems that are interesting to me after reading a paper and I can implement some of those solutions. The way I do this is I read a few papers about a specific problem see what they recommend for future implementation and try to see if I can do it or if I have another idea.

What I would like to get advice on is: How would you go about writing the conference paper? It seems to me that there are many different requirements for each different conference. Should I try to publish a conference or journal? Is there a way to quantify a significant finding? Would be easier for me to do a review paper or build a new dataset? Is there another way you think I should approach writing a paper as in how do you usually find a problem and solve it? Would you advise against doing this and try and find a supervisor somehow?

My goal is to research in NLP whether alone (not preferred), in academia, or to do it my Ph.D. but it seems like my profile is lacking big tech or research experience in comparison to students from Europe, States, China, and India.

I hope that there is someone who can give me some sort of guidance here.

Answer 1

I think it's almost impossible, if not truly so, to do your first publications without assistance (assuming those publications are to be in legitimate, reputable venues). You say your only option is to publish by yourself without guidance; my advice is strongly to say that path is not likely to be fruitful and suggest that you reconsider your other options.

I don't know what you have to do or can do to get an advisor at this point, but do that. If you were an undergraduate looking for a graduate program, I would strongly advise that you choose a program that shares your interest in doing research, and make sure you will be supported in that research. Since that choice is in the past, your options are limited, and you'll have to consider what makes most sense for you...

It might mean working through your program's staff to help connect you with someone if people aren't responding on their own.

It might mean reaching out to professors who teach courses you take and asking them for advice, if not on your research directly then on finding an advisor.

It might mean rethinking the specificity of your research interests and instead working on problems someone else finds interesting. Only tenured professors have the most opportunity to choose their own problems.

It might mean finding a collaborator outside your institution.

It might mean changing programs.

Things in most places are pretty messed up right now and these do not qualify as ordinary times, especially in NYC; your best options might include some patience and waiting for things to return to a more stable setting.

---

Too many possibilities and too many unknowns/factors specific to you and your current institution to give any specific answer, but the process of research, not just the contents, are difficult enough with help. This is not a good time in your career to venture out alone prematurely.

Answer 2

There is a path, but it isn't short or easy. You have three main problems. The first is to know what has already been done. There is a vast literature. What you study in a masters is mostly what is well known, though a bit advanced. But to get started with a journey to the edge, you might pick out an annual conference that appeals to you and select within that conference one of the specialized tracks. Read all of the papers from the previous couple of years in that track and go through their bibliography, reading more, until you get a sense of things. Doctoral students do something like this (and more) for their literature search. It isn't good to try to publish something that is known.

The second problem is learning to write your results in a comprehensible way. The literature search can help you with this if you get a sense about how things are presented in that field. What level of detail, etc. Perhaps your masters work has given you some sense of this already.

The third problem is having something that is both new and significant. That is hard to judge until you are deep into the study of a specialized area and can take experience. But, if you think it might be good enough and you've done the other two things, then you just submit to the conference, usually through their online submission system. You will get feedback and for a first attempt it may be disappointing feedback. If you have really missed the mark then you won't get much help, just a quick reject. But if you are close to measuring up then reviewers of your paper will give you some detailed feedback that can help you even if you ultimately get rejected. But whether your work based on readings and musings measures up to the required quality will, in the end, be measured by others.

If you have the funds, you can also attend a conference. You don't normally need to be a presenter to attend and you will have the opportunity to hear about a lot of research. You can get a sense from this whether your own ideas are of a similar quality. Can you follow the presentations? Is what they present something you might have thought of?

Another option is to take an advanced course at one of the universities in your village (NYC, I think you indicate). Many universities will let you take courses without being in a degree program (in the US, anyway). Take the most advanced course you are prepared to handle. And use it as an opportunity to meet some faculty, though waiting until we see the end of the pandemic would probably be necessary for any real contact. Your masters may provide this opportunity now, in fact.

As Bryan Krause noted in a comment, it is very hard to get into the game without a guide.

And, what I've described is more or less what a doctoral research student will do, though with the help of an advisor. But some people are self directed enough to work with minimal (occasionally no) help from the advisor.