Interview with Dr. Derek Bingham

Professor, Department of Statistics & Actuarial Science

Industrial Statistics

Dr. Bingham’s research focuses on the statistical aspects of uncertainty quantification. He develops statistical methodology for new types of data or new applications in fields ranging from cosmology to glaciology. This applied work is complemented by a fundamental research stream that looks at the broader and more theoretical side of what you can and cannot do with these computational models.

How has your research program evolved since becoming an independent researcher?
I started off working on the mathematical design of experiments. Although motivated by applications, the work was theoretical and technical in nature. Later, I became more involved with applications-based research. I am a methodologist rather than an applied statistician; I’m on the lookout for projects where the technical data or the problem arising in their application isn’t in a book because it hasn't been solved.

Your research impacts a wide variety of fields. What are some of your ongoing projects?
A lot of applied mathematicians, physicists, computer scientists, etc. use computational models to describe and investigate physical systems. One such collaboration I have is with cosmologists who are programming supercomputers to develop calculations that describe the Universe. The models are generally faulty, but they represent the best physics known. This work is complemented by physical observations made from real physical systems (e.g., measurements obtained using satellites). So, you have observations of reality, which are noisy and have limits due to the observational system, and then there is this imperfect computer model built using the best knowledge available. The question is how to put the imperfect computer model together with the noisy observations of reality to make predictions of what will happen in the future.

I work on a range of other projects that use computational models and observations, with my role being to use the computational models to make inferences. For example, in a glaciology project we are designing experiments and using computational models to make inferences about glacial evolution. Another team project, funded by the Center for Exascale Radiation Transport (CERT), is with nuclear engineers at Texas A&M University who are developing exascale computational models for very large supercomputers to understand thermal radiation transport.

Having had two terms as a Tier 2 Canada Research Chair (CRC), what did it allow you to do that you couldn't have done otherwise?
With the help of my Department and the University I built a lab, which is not common in statistics, because typically what we need is computing resources. Through that excellent resource, I recruited students and created a rich environment for many visiting researchers. As a CRC I received protected time for research; this allowed me to reach out to scientists outside my university, to travel to their labs and to have them visit my lab to create the research enterprise.

What personal research experience was the most exciting for you?
As a PhD student with Randy Sitter at SFU, I wrote an algorithm to generate all kinds of experimental designs so that I could investigate which was the best. My supervisor eventually asked me where I got the program, and when I told him that I wrote it, he looked at me and said, “Well, that is a paper!” It was exciting to realize that some of my ideas were not obvious to everybody else and that I could actually do research. I didn't know what a significant thought was, so this first taste of research was a real thrill.

As a student I first assumed that research was magic. The truth is, research is incremental; you have to understand what others have done to make progress yourself. Step zero is to read everything, to become an expert, and then you can apply your creative ideas and discover something new. Many people can do this if they work hard; there is a method in place, it is not magic.

Over your career, how has the funding landscape changed for Canadian researchers?
In Canada the main funding agency for natural sciences and engineering research is NSERC. NSERC funding has changed a great deal since I started. There's a greater focus now on training highly qualified personnel, which is a good thing in general. But it's become more difficult for new people to get started because they don't have a track record.

For the most part, it used to be that you could go to any Canadian university and be better off than a new researcher in the U.S., because you would have an NSERC operating grant; however, funding is more difficult to get now. If you're at a smaller university and aren’t funded initially, it's difficult to start building your career in a way that will allow you to continue to be funded. The previous approach was a strength of the Canadian system, the current model is a weakness.

What approach to funding do you take?
In addition to NSERC, sources of funding are also available to me through maths institutes like the Pacific Institute for the Mathematical Sciences and the Canadian Statistical Sciences Institute. I have also found success through being part of research teams in the U.S. because the U.S. funds ‘big science’, something we don't do in Canada. These U.S. projects offer alternate sources of funding and great science to be part of. But building those relationships is hard work and means time away from work and family.

What other roles do you play in your research community? Why are these contributions important?
As your career progresses you have opportunities to have a positive influence on the researchers in your field. I have served as an NSERC Evaluation Group member and most recently as the NSERC Evaluation Group Chair for Statistics.  Being part of the system that evaluates peoples’ grants matters because it affects their careers, especially those of early career researchers.

What have you learned from serving on the NSERC Evaluation Group?
I am impressed by the large number of good researchers in the country and by the way others write and frame their work. Overall, I have a deepened appreciation for the research that goes on in Canada.  I also noticed what other people’s students graduate with in terms of publications, awards and so on; and as a result, I make sure that the bar gets raised for my own students.

Derek Bingham and colleague Tom Loughin

What contemporary scientific issue concerns you the most and needs more attention?
As a scientific community we need to better communicate our results to the general public, especially when it comes to understanding numbers. Explaining what we do, the value of what we do and why it's worth doing is important.

Public debates about things like climate change are political, not scientific. If people were better at communicating the science behind an issue (e.g., through schools or through writing) then these issues would cease to be political, they would just be facts. If more people with science backgrounds became journalists, then a politician talking about a scientific endeavour could be asked informed questions and challenged if what they say is not supported by evidence.

Do you have any final thoughts to share?
This is the golden age to be a statistician. Statistics is like the new calculus: every student in all areas of science should have statistics training. Now is a good time to be a student – the opportunities are just so cool!


Read more: Dr. Bingham’s profiles on the Department of Statistics & Actuarial Science website, the Grad Studies website and the Featured Researchers page

Interview by Jacqueline Watson with Theresa Kitos