FACES OF MATHEMATICS

Gareth Roberts works in applied probability. His research focusses on stochastic processes and their application in computational statistics. He has established crucial results on the Markov chain Monte Carlo technique, an essential part of modern Bayesian statistical modelling. Roberts' work involves explorations of the mathematical foundations of computational algorithms, and it combines both a deep theoretical understanding with a keen sense of the way such results can be used to guide and enhance practical applications.

"My research really splits into three areas: one is theoretical Markov chain research, there is Markov chain Monte Carlo methodology work, interfacing theory with the methodology of algorithms, and then there are the applications themselves. I find myself stretched in both directions at the same time, and doing some quite applied problems while still doing the theoretical Markov chain work.

"Collaborating is one of the most pleasurable things in research. Most of the people I collaborate with end up being quite close friends, and I think that is one of the most privileged things about being a researcher of any type. Working with somebody is really the catalyst for getting things done.

"In some sense, in the areas that I work in, the important things are not the theorems. The important thing is finding a way to interpret the theorem in a way that is practically useful. There are many very, very good theorems in mathematics that are not practically useful, but if you're working in an area that is guided by algorithms, then you have to find a way... "