Meaningful research in PhD with no publications

Question

I have been doing my PhD in Computer Science for one year and a half now (Europe-based) and I feel my research is not going much far. I choose to do this after my master studies because I actually like to do it: I like reading about new solutions, new problems and try to solve them myself. I also really liked taking university courses.

Lately, I started to notice more and more how many of my peers managed to publish in some conferences, while I still have nothing close to submission. Also, I had a talk with my supervisor few months ago, where he hinted that my approach to the PhD was not going to be productive, and that a change was needed. He mentioned the need for more independence on my side, when finding and developing research ideas.

To make it (hopefully) more explicit, it may be that I like doing research for my own fulfillment and curiosity, but I cannot spot questions or problems that needs to be addressed or can have a chance during the review process. I don't know how to explain the current situation in another way. In the meantime I also got a bit discouraged, and asked myself if, career-wise, it is worth to pursue a path where I am happy with my day-to-day activities, but ultimately don't get any "public" acknowledgement. I think I learned a lot in this year (also, my level was maybe not so high as my peers'), so I am not unsatisfied with myself, but the issue here is that progress is measured on papers and not self-achievements. Working hard is also okay for me, so I would say motivation is not the issue here.

Should I consider quitting my PhD, or moving to a lower-rank institute? I'm fine with my work at the moment, but I'm worried it may not get better -- and I don't like the idea of finding myself at the fourth+ year of PhD, still with no publication. I don't think I want to keep on with research afterwards, and I fear those years will not be meaningful career-wise.

Answer 1

Your situation hints at a mismatch between your supervisor's supervision style (which is fairly hands-off) and the style you would need to be productive (which requires more guidance and input in the beginning, with the goal of ultimately having you become more independent).

My recommendation is: Find a co-advisor with a more hands-on supervision style. That could be somebody in your group who is fairly productive (most likely a postdoc, assistant professor or young associate professor), with a visible publication track record. Most people who are that way have more problems to work on that they can solve on their own. You should have your main advisor on board with this, which might be easy to achieve, since advisors have an active interest in seeing their students graduate.

Answer 2

I don't want to step on the excellent answer of lighthouse keeper but an additional idea would be to join or form a discussion group in which a few students, perhaps in a similar situation as yourself, get together to discuss issues and search for open questions. Read a few papers (jointly) and discuss them. What is left unsaid in the papers.

The synergy of a group, if like-minded, can be helpful.

And a group can transform into a collaboration group over time. This can be valuable in its own right.

Answer 3

I like reading about new solutions, new problems and try to solve them myself.

Do you find any flaws in the solutions that you read about? Any space for improvement? Could a different approach be taken?

Are the problems you read about framed appropriately? Maybe they are too theoretical, and making more realistic assumptions would yield different results? Maybe the problems are defined in a too specific manner? Could they be generalized for some benefit?

Start thinking critically and soon you will find plenty of ideas which are publish-worthy.

You can start by nit-picking, finding even small issues in the published works, looking for a hole. Then, you can try a higher level approach, like asking yourself questions why nobody has tried to do something in a specific way, or why some problem has not been tackled in a rigorous manner.

Do not treat published material as some form of revealed truth. You need to believe in yourself. Believe that you can do just as good as your peers (not the other students, but the other researchers which publish), or even better. Basically think of yourself as a researcher, not much as a student. Doing PhD is a (hopefully paid) job, not a school!

Answer 4

On your point about how to spot "research questions": spotting unsolved problems is a challenge in itself because usually one approaches a subject from the student point of view. Practical advice: read papers and pay attention to the conclusions. Often there are paragraphs along the lines of "will be explored in future work" or "remains to be shown". This is where your open questions are. Contact the authors and find out what they are up to.

Answer 5

You should probably quit. Unlike master/bachelors, getting a PhD is all about doing research and publishing the results. In the formal sense, research implies working on novel problems, at the edge of human understanding. Hence, publishable research has to be novel, in the sense that no one has done it before, and significant, in the sense it answers questions other people actually care about.

You should be exposed to a bunch of others research during your PhD, by reading research articles or attending talks (which seem to be important in Comp-Sci). If you watch, you'll notice that almost all research talks end with an 'ideas for further research' or 'limitations to the research'. Doing something that extends previous research, or overcomes the limitations of published research will get the job done.

When I asked about what I wanted to research, I gabbled out some plausible sounding stuff. Years later, I learned that academic research is hard, and that you have to work hard at a very small part of a problem to be able to make meaningful progress. Dangerously, you may pick a small problem no one cares about. I seem to recall books/articles/blogs in compsci about how to pick a good problem. (Professors who ask about your 'research topic' are wankers--a researchable problem is a very small sub-area within a topic area).

Doing independent research requires preparation, and plenty of professors are happy to have someone just doing scut-work. What you have to realize is that scut-work is training to be able to do it on your own, and what you need to be thinking about is why a professor is telling you to do those things.

As for why you aren't/weren't told these things: research is a craft, and academia (surreally) remains much like a medieval craft guild. There is also the awkward fact that many PhD students are second generation academics, with one or more parents capable of advising them on the 'mysteries' of the guild. Some of the rest have been initiated into the norms through a variety of means.

Answer 6

Should I consider quitting my PhD, or moving to a lower-rank institute?

Why is your situation frustrating to you? Overall, reading between the lines, I see a discouragement at a growing awareness of your limitations versus what you believe about the expectations being put on you.

My foremost recommendation is this: You should have a discussion with your current advisor.

Your intent should be to determine whether the expectations that you are starting to believe about your lack of performance (no publications) and lack of independence (as per your advisor) are real or simply mis-perceptions. A concrete example is to ask your advisor: Should I really have published by now? A broader example is to ask your advisor what specific actions you can do to demonstrate that you are taking on a greater level of independence.

I imagine that, accepting that you must hold this discussion is not easy. The consequences of avoiding it can however be damaging, for example by causing continued stress worrying about what you think your advisor is thinking or worse, discovering that your advisor has essentially given up on you showing sincere interest in learning by pro-activity participating in the advisor-to-student experience. Advisors are not ignorant about the shortfalls of their advisees, nor are they necessarily ignorant of their responsibility to find ways to help when asked.

I would have one specific answer to the second, broader question about possible ways to show your independence. Step forward at the discussion with your advisor with your plan to prepare a PROPOSAL for your research to defend to a dissertation committee by a specific, hard-set deadline. A proposal is NOT a defense of work done. It is a defense of work to be done. It states why the work is important in context of current issues. It states your hypotheses or problems. It defines how and why you will go about doing the work needed to valid the hypotheses or solve the problems.

This type of approach is one that is common in some PhD institutes. Instead of asking for a qualifying exam on didactic courses to be granted permission to continue with a PhD, the department requires that students prepare and defend a dissertation proposal. They require this even before a student is to have submitted any publications. Indeed, in some cases, they set hard deadlines for this step, with the consequence that the student who does not do it is not allowed to continue further.

This action could be one way for you to start to show independence (something that your advisor has already said seems to be lacking). This approach is likely also to result in greater "public acknowledgement", even if only at a local (university) scale to start. Finally, this approach is likely to focus your attention better to prepare a publication on your research topic, indeed even help streamline the writing and acceptance process because you have done the background work for the introduction.

Taking this step will also help you to discover other issues that may be behind your frustration. For example, do you have appropriate levels of self-disciplin, skills, and tools to self-administer the timelines that are required to meet concrete milestones for your dissertation? If not, start spending more time to learn how to set concrete, realistic, and relevant milestones; start finding the tools that you need to do the administration on your own (e.g. calendars and task-manager apps and citation management apps); and start setting concrete milestones to complete at hard-set deadlines. I might suggest here that you "work the puzzle backwards". When you do want to graduate? When must you submit your dissertation? When must you give your oral presentation? When must you submit the final draft to your advisor? When must you stop doing research and start writing? And so on.

Continuing on the above, the remaining question to ask is whether you have a true, deep-seated motivation to even want carry out your current research. Does your motivation to do a PhD come simply from a nebulous desire to explore things further, or does it come specifically from a burning curiosity about or desire to fix a specific problem in a specific context of a specific research field. As you move from the being in the former situation (nebulous curiosity) to the latter situation (actions driven by the need to meet specific desires), your frustrations will change from "I don't even really know why I am here" to "This is not going to be as easy as I thought (but I will get it done)".

Answer 7

(This is more an idea about a different approach than an answer. I'm curious to see what folks currently active in academia (ideally in comp sci) think about it. Meant to complement the other strong answers.)

Seek collaboration.
You have strengths and weaknesses. Make a list of them. I'm guessing you're great at execution, and unusually analytical, based on what you wrote. This is a VERY important skill set, and likely worthy of publication credit as a product intellectual achievement1. People who are great at coming up with good ideas are often bad at execution, and have more good ideas than they're able to make use of. Find out about potential collaborators' strengths and weaknesses including your advisor's. (You CAN ask!) Mars noted your likely gender; likely you have superior people skills. Sharing your list would help your advisor help you, or he's lousy. Odds are good you know colleagues with great ideas willing to collaborate. Also, in CS, PhD students are often paid, so find out what sort of timeline he things is acceptable. So consider having a broader conversation with your advisor to, among other things see what he thinks about complementary skills. Maybe he and/or the department or school hew to the publish or perish maxim.

1:The #1 author criterion: "Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work" per ICMJE anyway (Medical). More ideas