I come across such research on a regular basis; those look pretty extraordinary during presentations; however, when the same research is presented to an industry expert, that research doesn't look like a research project anymore. That looks like a novice trying to crawl through.

Last Monday, I attended a seminar on machine learning and AI.

A PhD student of bioinformatics presented a talk on how he obtained DNA sequences by systematically applying a series of ML algorithms. He will most probably publish this project in a peer-reviewed journal.

Afterwards, I talked to a professional ML/AI expert to discuss the presentation and the project, and he told me that that could have been done more efficiently in a different way. Then he described the whole procedure.

So, the question becomes: how do I know that a research project that my professor gave me to do is actually novel research?

  • 1
    Unless the student claimed efficiency as the goal of their approach, I don’t see how the statement of the industry professional invalidates it as research. Can you clarify your doubts? Commented May 24 at 5:23
  • 6
    Well, that is not what the question says, but even if we roll with it: Why do you doubt them? Some random industry person at some random conference said something random about some random student - so now you doubt your professor why exactly? Sure this can probably be answered in the most general way - based on just "I have doubts" - but then all but the last sentence of your question is moot. If all those information somehow relates to your doubts, it would be useful to clarify how. Commented May 24 at 6:37
  • 1
    It's also worth pointing out a PhD student is - a student. The way I structure PhD projects is to have a relatively straightforward task/project in the first year that helps develop research skills, etc. The results may be relatively small/incremental/trivial (but still relevant for the PhD), but it is the skill development that is important (inc. presenting at conferences). The more significant (harder) research can be done once those skills have been sufficiently developed.
    – atom44
    Commented May 24 at 9:42
  • 3
    This question confuses the concepts "publishable", "novel", "useful" and seems to treat them as interchangeable when they are not. This all seems like a very roundabout way to express doubts in your research mentor, which you have done many times on this site. If this is how you feel, I think you should quit, you're likely wasting your advisor's time and your own.
    – Bryan Krause
    Commented May 24 at 14:07
  • 1
    How do you know? By knowing the literature well enough to know how your work might fit in, trying some things, and revectoring if progress is not made. That is, you learn how to do research.
    – Jon Custer
    Commented May 24 at 14:21

3 Answers 3


First, I think you may be confusing "publishable" with "valuable."

Publishable research is relatively easy to define, at least in the abstract-- it's a unique contribution to human knowledge.

Valuable is a whole different thing, and it varies a lot from individual to individual, group to group, etc. It's trite, but people value things differently from one another. I am acutely aware of this because I got a PhD while working part time in industry, and I regularly experienced a sort of intellectual whiplash between evaluating things as working engineer, or as a novice academic researcher.

Second, I think you are over-weighting one individual's opinion.

I remember clearly standing in a poster session presenting a negative result. It wasn't a bad experience except for one moment when someone walked up, looked my poster over, asked a perfunctory question, and walked away saying, "Looks like your hypothesis was wrong, then." Like I didn't already know that. But it was still a contribution to human knowledge because a lot of experts thought it would work. I didn't let that one guy shut me down.

In your case, do you know that your industry expert could do it better? Does he claim to have done it, or is he just speculating (even if it expert speculation)? And regardless, you stipulate that he gave a different approach-- this means that more than likely, the presentation's approach was still novel. Novel methods are still advances of human knowledge.

Third, you have some responsibility to guide your own dissertation.

It's not unreasonable to want to do the kind of research you're yearning for, by the way. But this is your (probably) one and only trip through the PhD process. You have a significant responsibility to make the most of that trip for yourself.

A good thing to do here, would be to discuss the sort of research you want to do with your advisor. Maybe they're the right fit for you, maybe they're not. More likely, they fit some aspects but not others... but they probably know people. They may be able to suggest collaborations, either at your institute, another institute, or in industry, that will get you the kind of experience you want. You may be able to tailor your committee to get more of the advice you want.

There are options, here. Assuming your advisor is an idiot until that's been conclusively demonstrated is not a good one.

Fourth, you are novices trying to crawl through

There's not much of a body, here. But you are novice researchers. That's what a PhD student is.

Don't be so hard on yourself. Or, in this case, your advisor.


You know if a research question is novel by doing literature research. You just start reading what others have done in this area. Finding the "research gap" is a standard part of doing research.

Don't think of that as an adversarial process: research is done by humans with usually quite severe restrictions imposed by their budget, ethics, and just the way the world works (e.g. the fact that we still cannot time travel is a severe limitation for historical research). So research is more a collaborative project where different (groups of) researchers try different imperfect ways of answering the same question. No single group will answer the question, but body of research of all the groups together can (though there is no guarantee). So a productive approach would be: "What can I add to this joint project?" rather than "What did they do wrong?".


'Novel research' is what you will deliver by the end of your PhD (question tagged phd). What is given is a research topic and possibly a proposal for how to try to study the topic -- i.e. a pointer towards research that may be novel. Of course there is no threshold to test whether or not something is novel.

You must log in to answer this question.

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