How do professors change research interests?

Question

My question is along the lines of How do professors usually discover "new" research interests? but more to do with "HOW"? The above question deals with how professors discover new areas, my question is once a professor or a grad student has found an interesting lead, how do they pursue it?

Consider Prof. X who is a mathematician and just talked to his old MS student who is now working on Machine Learning/Artificial Intelligence. He speaks to the student for 20 minutes about how is applying math to AI. Where does he go from here?

Does he:

1. Go to Wikipedia/Encyclopaedia to get a rough overview of the subject.
2. Search for papers.
3. Search for a good book to read.
4. Speak to someone in his institution who works on AI and ask them for reading references.
5. Something else?

How does a professor make sure he gets the right resource considering his expertise in some subject? For instance, the professor would definitely want a book which starts from a primer on Linear Algebra

Answer 1

"All of the above"... but maybe not wanting a book that started from a primer on Linear Algebra. (Probably the latter remark in the question is a result of a misunderstanding of some sort.) In short, all possible sources are used. I have little interest in connections to textbooks (which typically feel an obligation to include busywork exercises, etc. and express reverence to the founders of the subject by recapitulating "troubles of the old times" (which have been solved, often).)

The operational point is that one skims all these resources, looking for key points, rather than reading ploddingly, slavishly. I look to be persuaded... or, equally interestingly, to be persuaded that the people doing the thing don't adequately understand it so as to be able to explain it simply. Quasi-ironically, the latter affords more possibilities for making genuine progress. :)

Answer 2

I've actually done this 2 times. First going from pure theory to 50/50 theory/experiment, and then switching departments. I'm a physicist but I think the following observations may hold generally:

1) don't quit your "day job". I.e., work gradually into the new field. I wouldn't take on any grad students in the new field until you were really sure.

2) start going to meetings in the new field. talk to the experts, get to know them.

3) if you have a sabbatical coming up, try to do it at a place that is a real focus of your new field. In my case I did a sabbatical at a place that allowed me to do theory and experiment.

4) some people switch fields to go after funding. I don't recommend this. If you have a real passion for something, find some way to fund it. who knows what the fashion will be in 10 years.

Finally, an observation. I found that colleagues tend to put you in a box and 20 years from now they'll think your doing the same old thing. I know very few academics who have worked in the same area their whole careers. No one expects this and it would be pretty boring if you ask me. Switching departments is harder, but switching fields I think is pretty easy if you're motivated.

Answer 3

The answer can best be understood in terms of the intended goal of the researcher making the switch. When moving to a new field, one is typically doing so because something they had heard/read/seen about that field was professionally interesting to them—i.e., they heard of an interesting research field or specific research problem—and they are considering trying to work on that problem themselves.

Given that framework, our researcher will learn everything they can about the research question that they can, using whatever tools make the most sense in a given field of research. In neuroscience, I'd start with recent review articles and gradually move towards more and more specific journal papers. If the field is entirely new to me, I'd probably try to find a good overview online somewhere before reading the review papers. I might try books, but they take too long to read, and will likely contain far too much information that I don't care about and too little that I do. If I had a colleague in the field I might ask them to point me to some good resources, and maybe discuss the field with me over coffee.

In the end, you want to be versed in the field enough to (1) understand the problem that interested you and (2) be able to formulate your own related problem for you to work on. Any resources that can help you achieve that goal can and should be consulted.