What could academia or industry could do (short or long term) to promote more collaboration?

I did a little poll of friends and colleagues about this question. Here are some of the answers which I found quite thought provoking:

I’m a recovering academic from many years ago.  I feel like I have some perspective on graduate/research departments in computer science, even though I am sure things have changed a little since I was in grad school.

One problem I saw is that a ton of the research done in Universities in computer science (outside areas like quantum computing, etc) lags behind industry.   A lot of graduate students in Software Engineering worked on projects that capable companies had already solved or that a senior industry developer could solve in a few weeks.

I also see a lot of graduate student project where they end up “building a tool” except the tool ends up being something nobody would ever use.

Every single one of those kinds of projects destroys the credibility of academics with industry.

A victory for academics seems to be publication or assembling statistical evidence for an assertion.  I get it but nobody in industry cares about those things.  Nobody.  Change your goalposts and align them with industry if you want to collaborate with industry.

I also think there is huge overlap between graduate student research and startups.  Lets say I’m 24 years old, and I think I have an idea to change the world with technology.  Instead of doing it at the University for a M.Sc I can just get some investment and build a startup (even without a business plan sometimes).

If academics want collaboration they need to be brutally honest with themselves and get more focused while facing where they sit today.  The software being written inside Universities often sucks. The research often moves too slowly.  Startups are the innovators.  The kinds of evidence and assertions being “proven” in academia are mostly uninteresting.  The outputs like publications are only read by other academics.

It might hurt but if you want credibility, cancel some of that crap. Work in the future, not in the past, understand your strengths and weaknesses and play to your strengths, change your goals to deliver outputs that are really consumable…

Its a lot to ask, so I don’t see any of that happening…

My company, engages quite a lot with academia, and even runs an Institute partly for this. The following is a bit of a brain-dump.

Within the institute we employ an academic-in-residence (Carlota Perez.) This is to explicitly support and sponsor work that we think is valuable and should be completed. In this case, to help her finish her second book. The institute also runs a fellowship programme. This is broadly defined to attract individuals with ideas and talent to offer them a network and opportunities, supported by a stipend. We explicitly define this quite broadly to allow people who may not want to start businesses to find value.

Obviously we’re interested in finding people who want to start businesses, but we keep that distinct from the fellowship to allow more far-reaching visions space to grow, at least a little. If fellows do want to found a business, and are capable of it, then we draw them into and support them in that.

We’re looking to participate more in academic-industry think-tanks, and other bodies. We individually connect to people in these bodies, and in academia, a lot in workshops we run. Mostly to generate ideas and explore spaces.

Finally, we read a lot of papers.

In our view, this is a start, but not enough. We are doing a little to sponsor the development of ideas within academia, via Carlota Perez, and we’re allowing people to start research projects in the fellowship. But we want to help with more execution and scale. We’ve tried to partner with some universities, but we find that they’re not commercially-focused enough to support us in raising the capital to actually execute with. They want to provide ideas, we provide execution, and capital appears by magic. We need a bit more than that.

I was affiliated with [Top UK University] for a time and here is my top-2 list of difficulties:

– IP: the university makes it really hard to separate the IP between work done during the collaboration vs work done in the day job (industry). The amount of paperwork is typical of a bureaucratic institution. Turn off for many people (why bother).

– IP again: this is slightly tangential to the original question and is more related to a different kind of industry-academia collaboration, one where the prof does a startup while in academia. [Top UK University] for example had a policy that 50% of the equity of the startup belonged to [Top UK University]. That number is huge. Prevents other VCs from investing in the startup. Guarantees that basically no one will do a serious startup. A more comparable number in leading US universities like Stanford is 2-5%. There were creative ways around that, but it was a grey area legally. Again, why would one bother going through the hoops. It’s easier to just not deal with academia at all.

My suggestion would be that industry and academia need to develop more understanding of, and respect for, each other’s needs and incentives. To put it bluntly, the career demands are very different: industry people need to ship products that customers care about, while academics need to publish papers in good venues. With those different incentives come different timelines for working (industry thinks about shipping quickly and long-term maintenance; academia thinks about big ideas for the future, but doesn’t care about the code once the paper is published), different prioritisation of aspects of the work (e.g. testing), etc. Of course those are over-simplified caricatures, but I hope you get the idea.

I don’t think one is better than the other — they are just different, and for a collaboration to be productive, I think there needs to be mutual understanding and empathy for these different needs. People who have only worked in one of the two may get frustrated with people from the other camp, feeling that they just “don’t get what’s important” (because indeed different things are important).

Caveat:  I’m still affiliated with various academic advisory boards so am somewhat biased by the progress we’re making. A few personal comments / observations:

– Although academia has shifted slightly to focus more on “impact” not just papers.

– The points made about  have always been particularly troublesome for working with [Top UK University] due to the[Top UK University] Innovations licensing arrangements but I think as that arrangement expires there’s recognition that companies can’t keep sinking massive grants into Universities unless they’re philanthropic without new creative commercial ways of working.

– Linked to the above two points one of the frustrations for industry is that a low TRL development that appears to be 80% of the commercial offer realised in a Uni can be achieved in 20% of the time but the other “20%” productisation to commercial fruition / TRL7 will be 800% of the industry partners production costs and associated time etc… This should be reflected in the engagement and IP position but isn’t really.

– Academia is only just recognising that it must adjust to collaborate or risk being out competed where “Quantum compute” or “fundamental battery tech”,etc ,etc research groups are appearing in bigger tech companies.

Caveat – my subjective view out of ignorance from the fringes: The EPSRC Industrial Strategy Challenge Fund and Prosperity Partnerships are a massive opportunity and yet the ISCF Waves that have appeared appear to have done so with limited industrial awareness, formal structure and engagement. So those that have been engaged have been at the table more likely through personal relationships, etc. So this needs more publicity and more formality… There also needs to be a clear understanding of Innovate UK, the Catapults’ and Research Councils’ roles.

I’m not sure I have a great answer to this but I think it’s an interesting question. In the distributed systems world academia plays an important role, but there is always a divide. Things that I think might be useful:
– Doing more to reach the audience in industry. The best example of this i’ve seen is https://blog.acolyer.org/.
– Partnering to study why things work well in practice rather than in theory. For example there is much the wider community can learn from the internal design decisions made by key open source components that run in the real world. So in my field the design decisions made building Kafka, Cassandra, Zookeeper, HBase could use further study which would be useful for the next iteration of technologies.
– Making it easier for industrial practitioners to play a role in academia. I know a few people that do this, but i’m not entirely sure how it works, but I feel it could be done more.

Finally some comments on twitter here: https://twitter.com/benstopford/status/917991118058459138

Posted on October 14th, 2017 in Blog