Go back

Top PI: Nigel Shadbolt, guru of open data

Now an adviser to the prime minister on open data, Nigel Shadbolt is also a veteran research council grantee who has raised £26 million in a 20-year career in artificial intelligence. The first grant, he tells Rebecca Hill, is always the hardest.

How did you begin your research career?

I did my PhD in AI in Edinburgh and [three years after that] I moved to Nottingham, in the psychology department. I got Engineering and Physical Sciences Research Council funding there to set up an AI department. Although my first grant was a Nuffield Foundation fellowship for early-career researchers, I’ve had a fairly sustained career getting money out of the EPSRC. I moved to Southampton about 12 years ago, and the first thing I got was an EPSRC grant to set up [a £7.5m] Interdisciplinary Research Collaboration in Advanced Knowledge Technologies. My most recent grant is fundamentally about asking this question: do we have either the theory or the understanding of how the most important systems are emerging in the world today? I think this is an area everybody recognises is hugely important to understand, but we don’t understand it!

Why do you think you’ve succeeded in getting high-value grants?

I think it’s reflected in the fact the grants have often involved a variety of industry and been large and collaborative in nature. I’ve done a lot of research in areas that are on the boundaries between psychology, decision support and computer science—this is where the interesting challenges are, and you can create and exploit different areas of research.

How do you form and maintain your collaborations?

If you look back historically, I’ve tended to collaborate with the same core group of people. You recognise your working style and research style, and though the science is crucial and fundamental it does help if you’ve got like-minded individuals and people with complementary skills. But you can get too used to existing structures and collaborations, and there is a theory you should shake them up a bit. It can also be very easy to get locked into an area and not give yourself a chance to look around or change direction. Networking at conferences and workshops is really important because you’re exposed to a whole range of potential collaborators. Or it might just convince you you’re quite content with the people you’re working with!

How do you apply for a Programme grant?

You become eligible to apply if you’ve a record of having managed substantially large grants, as the research council invests a degree of trust in you. They basically tell you to go away and get on with it—it gives you a huge ability to change in the face of circumstances, which is essential in emerging systems like the web. There’s a detailed programme of work laid out at the start, but what you’re able to do at various key points is to take stock, and if you need to change direction you can. 

What’s the application process like?

The big trick is to write the abstract and get that through; when you get to the point of writing a detailed scientific case you’re in more familiar territory. The programme managers who review the abstract could be anyone from chemists and physicists to IT scientists, so you have to make the challenge and benefits very clear. The great thing is that the EPSRC is allowed to provide a certain level of feedback—for instance to remind us of the key attributes the panel is going to look for—and you do have a few goes at getting that proposition clearly characterised. But of course they ensure it’s a fair process for everybody competing at the time.

What should you make sure you do, and is it a challenge?

Oh it was gruelling! You have the window of opportunity and you know you have one go at this, so it’s important you give it your best shot. You’ve got to have a strong set of statements—they’re not looking for incremental research; they’re looking for something that is going to be significant to the field and shift it in some way. You’ve also got to line up the very best industrial components and potential collaborators—in this case it was people I’d worked with before; I think for programme grants you need to have a high level of confidence that you can deliver what you say.

What sort of preparation did you do before the interview stage?

When you’re called up to an interview panel it’s likely that there are going to be a lot of strongly ranked proposals alongside your own, so it’s a key stage. We had a number of mock panels to work out what questions we might be asked, and spoke to people who’d been through the process before. The general properties people are looking for are whether you work together, that you understand what you’re trying to do, and that you know how to disseminate your results—this idea of national impact is an important element for any modern research programme.

How does a programme grant compare with getting other grants?

Thinking back, the hardest grant is always your first. The basic characteristics of grants—from whatever research council, going through to whatever level—are essentially the same. It’s a mix of ambition, great science and engineering, and a strong sense of how it will make a difference. It’s also about having the right team, and whether you make a sensible set of statements about what resources you need and how you’ll manage them. I’d like to think that it isn’t so very different [to any other grant]. Of course the odds will sometimes be a bit different. It really does pay to do that little bit of background work to ask what the expectations are in this area—what have they funded in the past and why should they fund mine above any others?

Which areas do you think will hold the best funding opportunities in the future?

I think, from the engineering and computing and web angle, and also intersecting into human psychology, one area is to look at the questions that web science can ask itself. For instance, how are people being changed or augmented by technology? I was involved in a report the Royal Society just published on human enhancement, and there are lots of challenges in it; it’s quite a good blueprint for people to look at. Another area would be around the explosion of personal information; the kind of services and challenges we face in managing it, and understanding how to interact with it, will just get larger and larger. It’s going to provide huge research challenges at a level, scale and immediacy we’ve never seen before.