In this article, we shine the spotlight on Dr. James Kermode, Associate Professor at the Warwick Centre for Predictive Modelling and School of Engineering, University of Warwick, focusing on developing multiscale materials modelling algorithms, and the role of error and uncertainty in calculations.
What is your role?
I am an Associate Professor at the Warwick Centre for Predictive Modelling and School of Engineering at the University of Warwick. I’ve been at Warwick for four years, it’s a traditional academic role with a split of academic and research duties, so there’s quite a nice balance across those areas. My area of specialism is developing multiscale materials modelling algorithms and the software that implements them.
What do you enjoy most about your role?
The variety really, moving between teaching and working on developing new methodology within the research aspect. But also the applications focus, I try and maintain a balance between methodology and applications which invigorate each other quite nicely.
What’s your background?
I started off in condensed matter physics. In 2004, I did a PhD in the Theory of Condensed Matter (TCM) Physics at Cambridge University that finished in 2007. I kind of specialised within this field and in 2009, undertook a post doctoral position at King’s College London, focusing on developing software and methodology for multi scale materials modelling, with the aim of doing accurate simulations at large scale. Since moving to Warwick in 2014, I have tried to address the broader role of error and uncertainty in calculations, which is really one of the focuses of the Predictive Modelling Centre here at Warwick. It’s trying to put estimates on the errors you are making on the outputs of simulations. I think this is one of the important future areas of research in this field, to help increase the uptake of modelling in simulation outside of academia.
How did you get into this area? Was there anything in particular you were interested in at school?
I was always interested in science at school but experiments weren’t my strong point, so from there onwards I picked all the theoretical options. I have always enjoyed using computer systems as well, so it’s the perfect balance between doing a little bit of work and then experimenting through the computer, where you don’t have so much chance of the equipment letting you down.
In terms of your career path, going down the academic route - why was that?
It was more of a natural evolution, than a definite choice, and a way to keep doing the things I enjoyed the most. As opportunities came up I took them and developed within each role.
Are you working on any projects?
For a number of years I have been involved in working on multiscale quantum mechanical and classic simulations, particularly for things that break. We have put a lot of effort into fracture and plasticity in materials, trying to extend these techniques to work for rare events for things like motion of dislocations or processes that happen quite slowly, as well as big things that need multiscale models but are quite slow. This has been quite challenging, and we have had to do a lot of new methodological developments to apply those techniques.
Your area of focus is quite topical within the industry at the moment. Are there any particular sectors where you are seeing quite a strong interest?
Yes, I think the aerospace sector is becoming more and more interested on the nanoscale details of the materials being used to construct aerospace components. For example, super-alloy turbine blades are dominated by the atomic scale response of the material, so there is a lot of interest in what’s going on at the atomic scale. Another area I have been working on the edge of and is really picking up interest is machine learning. This is something we have been using as a tool for the last few years within the research group at Warwick, but is increasing across all areas of science and engineering, I think this is an important trend for the future.
Within your particular area of expertise, what are you looking forward to the most?
One of the focuses of Warwick Centre for Predictive Modelling is to predict errors on the output of simulations. This will be really important in increasing the uptake of modelling and simulation in the industrial sector, to try and reduce the number of prototypes built, lowering the cost of bringing new materials and processes to market. It is really important to have a confidence limit on the simulations that you use, so you know when to trust them, when not to and when you can’t. I think feeding this all the way through from the atomic scale, up through things like finite element models and into engineering design products over the next 10 or 20 years, could really have quite a dramatic impact, leading to a step change in the way that engineering and design of complex materials is carried out.
If you were stranded on a desert island, and you were granted three items, what would they be?
My Kindle so I have lots to read and no heavy books to carry! If I could take people, I would take my wife and baby son, and then the Kindle!
● A. P. Bartok, S. De, C. Poelking, N. Bernstein, J. R. Kermode, G. Csányi and M. Ceriotti, Machine learning unifies the modeling of materials and molecules. Science Advances3, e1701816 (2017). [arXiv] [Open Access]
● T. D. Swinburne and J. R. Kermode, Computing energy barriers for rare events from hybrid quantum/classical simulations through the virtual work principle, Phys. Rev. B96, 144102 (2017). [arXiv] [Open Access]
Click here to view more published research from James.