Research taste

Learning to do research is not just about learning to use the essential techniques in the field. It also involves developing a taste of what directions are worth exploring and what directions are not.

I am still in the middle of my PhD, but looking back, the past few years have already changed me in the way I look at different research directions.

Recognizing good starting points

I used to have many reasons to refuse to start some projects:

  • Problem is pointless: Even if the problem is solved, its solution does not seem relevant to the real world.
  • Problem is too difficult: The problem looks impossible to solve completely, so I’m not willing to risk the time and effort.
  • Problem is already solved: At a glance, this problem is already theoretically solved and seems to only involve engineering to make it happen.

Thinking like this might help to criticize an existing project in order to make it more meaningful. However, thinking like this before starting a project is counter-productive. It can prevent you from starting any project.

Before starting a project, look at the bright sides.

  • If the problem looks pointless in itself, try to think of it as a prototype. Does it have details that look like an analogy of a more important problem? Does it have details that inspire imagination in other scenarios?
  • If the problem looks too difficult to solve, try to think of it as a grand idealistic goal. Are there downgraded versions of this problem that look solvable? Are there potential viewpoints that have been overlooked by previous researchers?
  • If the problem is already solved theoretically, try to think of it as a dream for a new world. Can you imagine encountering interesting problems if you are to realize this solution? Is it meaningful to make the dream come true?

If you have good answers to these questions, it’s probably worth starting the project. Almost all research projects have limitations and are imperfect in one way or the other. So, at the start, don’t aim at choosing the perfect thing to do. Instead, just aim for something meaningful to do. In other words, aim for the “ugly baby” [1].

Recognizing dead ends

Since doing research is about exploring the unknown, it is ok to give up a direction and turn around if you recognize a dead end ahead. I confess that I am not good at judging dead ends, but the following hints are consistent with my (limited) experience.

  • If you have no idea how to solve it, think about what assumptions you can make about the problem to make it easier.
  • If your solution looks contrived or not generalizable, think about a more realistic assumption and try to solve the new problem.

The most convincing way to learn these lessons might be to get first-hand research experience. It might also help to read/hear stories of great researchers and focus on how they made decisions.

[1] Ed Catmull and Amy Wallace, Creativity, Inc.: Overcoming the Unseen Forces That Stand in the Way of True Inspiration

Added the “ugly baby” reference.