I am currently Professor of Statistics at the University of Warwick. I have a research group, which consists of about nine postdoctoral research assistants, and four PhD students, which is quite a large group. They are mainly working on problems related to the theoretical analysis of statistical methods and the development of statistical methods, along with algorithms for the quantification of uncertainty in a large number of application domains, one of them being in engineering. I’m also one of the Executive Directors for the Alan Turing Institute for Data Science, which is the national research institute for data science in the UK.
I think that uncertainty and uncertainty quantification is one of the foundations of statistics and statistical science, so in many ways, I’ve been working on uncertainty from my PhD. I think it's very much been a natural progression in finding the areas of engineering where this is now becoming very important.
I actually studied mechanical engineering at the University of Glasgow, so it’s been quite a journey from mechanical engineering to statistics. But that’s what I initially trained as, a mechanical engineer. It’s been a natural progression from mechanical to statistics over a long period of time. I think that engineering and mechanical engineering very much rely on applied maths and building models of various physical processes and systems. It wasn’t much of a leap from moving to that and actually looking at these systems as stochastic systems.
This came about because there was a competition to select a small number of universities that would form the Alan Turing Institute. After a round of international assessments and reviews, five universities were selected; the Universities of Oxford, Cambridge, Edinburgh, Warwick and University College London. I had been involved in the Warwick bid for the competition, and then it was considered, given my background and expertise, I would probably be the ideal candidate to take a role as an executive director for the Alan Turing Institute.
The IMS selects a small number of these lectures every year or two years, and they usually select someone who has made a contribution, had an impact or created a lot of interest in a particular area of mathematical statistics. I think I was selected based on some of the work I have been doing in uncertainty quantification.
As an academic I very much enjoy the freedom to really explore any scientific problems that I have an interest in. For me, this is the most enjoyable thing that I can do. I have the freedom to work on anything. If I decided I wanted to work on something completely different tomorrow, I could do that but I would need to build up a track record and I would need to obtain funding to support that work. Obtaining funding is a very competitive process, so in many ways it’s a little like running my own business but I very much enjoy being an academic and enjoying that freedom.
After I graduated from the University of Glasgow, I actually spent 10 years working at IBM as an engineer. I left IBM to further my studies and study for a PhD and then I ended up working where I am now, at the University of Warwick. I have had the experience of working within industry at the start of my career and I do still work with industry on a number of projects.
There are certain methods in computational statistics that I am known for developing. Without going into technical detail, it allows someone to quantify uncertainty based on probability distributions for very complex mathematical models. My work is used, ranging from engineers who model oil flow in subsurface aqua flows to applied mathematicians who are studying the heart by developing these large scale multiphysics multilevel models of the heart, and biologists who are modelling protein transport in the cell. All of these areas use the techniques and methods that I have developed which is one of the main things I am known for. Basically, wherever there is uncertainty on how that uncertainty should be quantified, propagated and used by the decision maker or policy maker or for investment decisions, in that respect, the scope of its application is very wide. What we’re trying to do is to provide tools for industry to help identify and quantify the sources of uncertainty in whatever is being studied.
For me, one of the biggest things that I really enjoy is to actually see work that I have done and research transferred either to a commercial application or something that’s picked up and used. I’ve had experience before of doing work that has ultimately led to products and services in the financial sector. I look forward to continuing that productivity. I am generally not able to talk about the majority of projects due to their sensitive nature but there is one I can mention. I did some research work for National Cash Registers (NCR), who manufacture the automatic teller machines (ATMs), assessing the validity of bank notes deposited into machines. I looked at some statistical approaches to try to solve this problem, and the solutions that I suggested actually went onto form technologies that were patented which were developed, and are now part of some of the ATMs that we use today.
Regarding the future, I think the most important thing is that for policy making, the understanding and communication of uncertainty is probably one of the biggest things we need to get right. The engineering community at the moment is trying to figure out ways to capture and quantify uncertainty. The really important point that will emerge from this is what we do with this. How does the engineering community optimally use this information?
The first item would be a very high powered laser so when an aircraft flies by, I could point it at them and they would see me and get me off the island. The second would be a good supply of Nutella, to keep me going until I get saved. The third item would be my iPad, as it has all my books on it but I guess the problem would be there would likely be no wireless access!