Why is academic research so difficult? — The problem of not having a problem

A closer look into the perils of academic research

Research is hard, no doubt about that. And several reasons go behind it. But is it intimidating only to newbies or is it difficult even for the rock-stars? In this article, we are going to analyse the various facets of research hardness, and give you a feel of where (and how) things go crazy in the entire research pipeline. BTW, this is the first issue of a series on the subject. The attempt is to dissect, diagnose, and with all our researcher readers, collectively brainstorm ways in which many of the core pain-points can be eliminated. So please share your views and comments :-).

The problem of not having a problem

We begin with the genesis problem where many of us (especially early-stage academic researchers) spend most of our time “exploring” and “hunting”. A problem or subject-of-study must have the following characteristics for qualifying as a research project:

1. It has to be non-trivial in nature and demand rigorous, in-depth analysis.
2. It should have a strong motivation — i.e. is it important enough (or has enough contemporary implication) to be solved or investigated?
3. It has to be open enough — i.e. is the state-of-the-art solution or analysis good enough, and any contribution is just going to be a small incremental step that cannot be held to be a valuable contribution?

Now, there are usually three ways in which academic researchers hunt for a problem:

1. Look for a “hot” area and then pick up a problem that is the in-thing.
2. Look for a “hot” approach/technique/methodology and then pick up a problem solvable by that.
3. Look for theme-based grant opportunities and then frame a problem accordingly, so as to maximize the chances of winning the grant.

Although such ways may invoke a certain amount of competition within the community, and that itself can bring forward some amazing solutions, yet they have their own perils. Like the first way of hunting can lead to a negative congestion effect, thereby thwarting growth of other areas that are “potentially hot” — i.e. stuffs that are ahead of time (I still remember people used to joke about me when I was a Ph.D. student back in 2006, working on some seemingly bizarre computationally expensive A.I. ideas!). The second way can actually slow down a generation of newer paradigms and approaches. The third way of hunting problems can tempt us away from solving problems that are connected to ground-reality, and, more importantly, from those problems that we are passionate about. And there is yet another concern, many of the existing exciting areas suddenly come to a standstill, with no significant attention from the community.

Traditional problem-hunting strategies need to be revisited

Many researchers carry forward the traditional custom of remaining in their own disciplinary silos and are totally disconnected from what is happening in other disciplines and, very importantly, how that is related to government policies and corporate investments. This can drive us towards problems that are either super-saturated or leave a very little possibility of significant contribution. It’s quite amazing to see how often researchers commit themselves exclusively to the study of latest published research papers for picking up problems that are, in the best case, three-years-old.

I can give you at least three arguments to support my claim:

1. The usual tendency is to look up papers that are within the time period of 3–5 years. This, btw, has a cascading backward chain effect.
2. The usual tendency is to look up papers that have high citations. It takes roughly around 3 years to get some sort of citation, much of that can be credited to my first argument.
3. Finally, just think about the long cycle of submission, rejection review, re-submission, acceptance, and final print, and the time of start of the research work on the problem. If I get a paper rejected then definitely I am not going to move on to another problem. I will persist, and that can go to publication even after 2–3 years!

After talking to many of my colleagues and Ph.D. students, I was quite surprised to see how few of them actually read domain-specific professional blogs, or watch videos of tier-1 conferences/workshops/tutorials/meetups across the world, or read up daily news articles related to their areas of research.

Just imagine the number of great open problems that go unnoticed because of this age-old, discipline-specific, paper-only behavior.

As researchers, we should be able to develop the skill of smelling what may be the future — and not just what is the status quo reported by the papers. And for that, we need to be aware of all the ideas getting baked now, and be able to collect and connect dots from across disciplines/sub-disciplines so as to come up with exciting new yet meaningful problems to work on.

The article has been first published in LinkedIn.