1
$\begingroup$

I recently came across a new PhD student who works in a materials modeling group which mainly focuses on applications (doing DFT-based studies on different solid/electrolyte interfaces). To my surprise, he actually did his master thesis in a well-known group which mainly focuses on theory and method development, where he was involved in the development of some deep learning models for representing the DFT Hamiltonian. He claimed that he had the opportunity to stay in that group to do a PhD but eventually he decided to switch to application-focused groups. He also said that he made the decision not because he found method development too difficult for him. His reasoning for the switch is that he is determined to go to industry after PhD. In his view, an application-focused group makes it easier to publish on high-impact journals due to the higher chance of collaborating with experimental groups. Industry would favor people with publications that provide practical values rather than theories and methods.

Personally, it's hard for me to understand his decision. I think it's not a bold statement to say that it requires more skills to work on theory/methodology than applications, and most people who work on theory/methodology are adequate to do applications as well. Is it true that the industry weighs the "practical values" more than individual skills? Is it true that theory/methodology-focused groups are more suitable for people who want to stay in academia?

Improve this question
$\endgroup$
3
  • 6
    $\begingroup$ I think this question would be more suited for Academia SE. I am not voting to close this question, but the question that you ask does not strictly belong here. $\endgroup$
    – Hemanth Haridas
    Jul 29 at 9:35
  • 3
    $\begingroup$ "I think it's not a bold statement to say that it requires more skills to work on theory/methodology than applications" It requires more skills to work in theory/methodology while still understanding the challenges posed by the applications. The hardest is to successfully connect the two worlds. If launching a DFT calculation is 'simpler' than coding a DFT code, solving real world challenges using DFT is no simple task. As for your friend point of view, I would say it depends what you want to do later, if you aim a CS career than the theory/methodology might be a better option etc $\endgroup$
    – Okano
    Jul 29 at 11:21
  • $\begingroup$ "I would like to know your thoughts on the pros and cons for joining a group focusing on applications vs. theory/methodology, in the long-term perspective of both industry and academia." --> Since there's already a lot to address in your post, and this part of your post can lead to enormous answers on its own, it's best left for a new post. $\endgroup$
    – Nike Dattani
    Jul 29 at 17:32

1 Answer 1

Reset to default
2
$\begingroup$

Reasons why he switched groups for his PhD

"His reasoning for the switch is that he is determined to go to industry after PhD. In his view, an application-focused group makes it easier to publish on high-impact journals due to the higher chance of collaborating with experimental groups. Personally, it's hard for me to understand his decision."

This is what he told you was his reason for the switch, and it was what he was capable of putting into words at the time. The decision to switch groups, and the decision of where to embark on a PhD program (a program which usually constitutes more than 5-10% of a person's entire years of life, and can have significant implications for the remainder of their life) almost always involves more than one factor. Dozens or hundreds of micro-factors play a role in where people land, and these are almost never articulated (he's not even aware of them all). You mentioned his desire to publish in high-impact journals, but surely there's other groups that publish in high-impact journals just as often or more often: Why didn't he go to those places? I can think of several possible reasons.

What requires more skills: theory or applications?

"I think it's not a bold statement to say that it requires more skills to work on theory/methodology than applications, and most people who work on theory/methodology are adequate to do applications as well."

I think that's an extremely bold statement. Okano's comment touches on this. You probably know far more than us about what his former group (which focuses on theory/methodology) does, and what his new group (which focuses on applications) does: The rest of us don't know the names of those groups, and therefore don't know why one would require more skills than the other, but in general there's theory/methodology projects being done successfully at undergrad-only colleges, and there's application projects whose success involves decades of post-PhD expertise (and there's vice versa examples).

What skills do industry groups value?

"Is it true that the industry weighs the "practical values" more than individual skills? Is it true that theory/methodology-focused groups are more suitable for people who want to stay in academia?"

This will depend strongly on the group in industry doing the hiring. You mentioned "high-impact journals", but the vast majority of people working in industry (or academia) do not have papers in high-impact journals like Science, Nature or even PRL or JACS. Industry (and academia) both value "practical skills" and the ability to apply science to achieve practical outcomes. In countries that have enough money, far more academic positions are available for experimentalists than theorists, and publishing application-based papers in high-impact journals is extremely valued in academia (as well as in industry).

People focusing on pure theory/methodology, rarely secure academic positions without making a strong case in their research proposal and interview about the potential practical applications (and it's the papers by the application-focused groups that serve as "validation" that the theorist is doing useful work; if the pure theorist has some such papers in their portfolio, it would be to their benefit). You said "he actually did his master thesis in a well-known group", and it's hard to comment without knowing the group, but academia and industry have both changed a lot since most such groups became well-known.

However, there's groups in industry that may specifically want someone that can help develop theory/methodology. Such job availability might in some cases be even more prevalent in industry than in academia (where money is tight, and mostly funded by tax-payers, and so there's more and more pressure to do things that directly benefit the tax-payers; whereas in industry a billion-dollar company might have the freedom to spend more of their money as they want, without the pressures that exist in academia).

Improve this answer
$\endgroup$
4
  • $\begingroup$ "There's theory/methodology projects being done successfully at undergrad-only colleges, and there's application projects whose success involves decades of post-PhD expertise (and there's vice versa examples)." I do believe these scenarios exist, but I'm afraid there are way more "vice versa examples". It is true that it requires skills to "solve real world challenges using DFT" (Okano), but most DFT application papers are not solving any real world problems. IMO the main difference is that there is often no ground truth for an application result, but it's either right or wrong for a theory. $\endgroup$
    – Shaun Han
    Jul 29 at 19:28
  • $\begingroup$ Maybe it is a bit bold to say applications require less skill. If we are talking about the highest-level research, nothing is easy. However, I do believe at entry/medium level, the skill requirement for applications is much lower than theory/methodology. $\endgroup$
    – Shaun Han
    Jul 29 at 19:46
  • $\begingroup$ To improve applications so that they can better solve real world problems, the most important thing is to get the physics right, which still relies on theoretical groups. I would like to hear what exactly are the "skills" that application people could have but theory/methodology people may not. $\endgroup$
    – Shaun Han
    Jul 29 at 19:57
  • 2
    $\begingroup$ Reading your last three comments, I will allow myself to say that your opinion seems to be slightly biased. I can also throw a random biased opinion: People doing applications must read the theory and understand it well enough such that they can solve a problem with it. In the meantime most theory/methodology groups just have to read the theory and paste the equation in Fortran. $\endgroup$
    – Okano
    Jul 29 at 21:36

You must log in to answer this question.

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