Over the past few weeks, I’ve had some discussions which touched on the “meta” topic of the role of academic research compared to industrial/commercial R&D (you can also think of “basic” vs “applied” research).   I guess the basic question is how can academics “add value” in a context (such as AI in 2018) where companies are spending large amounts of money on R&D, far beyond what academics can spend.

My view is that academics should be “explorers” or “scouts”, who go into unknown “research territory” to look for new insights, technologies, etc.  Like all good scouts and explorers, academics should “fan out” so they can explore as much territory as possible, they shouldnt “bunch together” and all go in the same direction.   From this perspective, industrial (applied) researchers are more colonisers.  Once the scouts have identified promising territory, the colonisers move in to settle and exploit the territory, and map it out in detail.

If you agree with me that this is sensible, it has some implications for how academics should behave

1. Academics, like scouts, should spread out so they can explore as much virgin territory as possible, instead of bunching up and all exploring the same area.  Ie, academics should not all jump on the latest “whats hot” bandwagon, whether its tree-adjoining grammars (the “hot” NLP area 25 years ago), or deep learning (the hot area in 2018).
2. Academics, like scouts, should honestly report what they have discovered, including unpromising findings and territories.  Ie, research should be rigorous and carefully evaluated, and negative results should be honestly reported.   Some academics are better at this than others.
3. Academics, like scouts, should not stick around after the “colonisers” have moved in to exploit an area.   Ie, once the well-funded industrial R&D types have moved in, academics should move on to new territory, instead of trying to compete with industrial labs.

Of course, the academics-as-scouts analogy is imperfect.  For example, PhD students may insist on studying whatever is trendy and “hot”, especially if this is where the high-paying jobs are; and academic funders may also prioritise whatever is trendy.  I have some sympathy for the students, especially since most of them will probably end up following the “coloniser” (applied/commercial) track rather than the “scout” (basic/academic) track. I have less sympathy for the academic funders, although again I appreciate that when govt research funding agencies make their pitch to politicians for money to support academic research, the politicians will probably be more receptive if they hear whatever “buzz words” are floating around in the popular media.

Anyways, the “academics as scouts and explorers” vision is probably idealistic, and its not for everyone.  However, I do think the world is a better place if at least some academics act like scouts and explorers.