Next generation scientists learn to apply computer simulations to science and engineering problems
World-leading experts taught data visualisation, code optimisation and machine learning techniques that could help solve some of sciences greatest questions at a summer academy for researchers at the University of Southampton.
The EPSRC Centre for Doctoral Training (CDT) in Next Generation Computational Modelling (NGCM), hosted over 50 students from across the UK at this months week-long event on Boldrewood Innovation Campus.
The summer academy brings together PhD students and early career researchers who work on the computer simulation of science and engineering problems and want to extend their training with courses by key developers on relevant software tools.
Expert courses during the week included a two-day overview of machine learning delivered by Professor Mahesan Niranjan, Professor Adam Prugel-Bennett and Dr Jonathon Hare from Southamptons School of Electronics and Computer Science.
Dr Andreas Jüttner, NGCM director, says: Be it improving the aerodynamics of a Formula 1 car, designing new medicines that better penetrate a cell membrane in the human body or the question about what the universe looked like shortly after the Big Bang - computational modelling of the underlying physical or chemical processes is an indispensable tool that is driving progress at the forefront of all branches of scientific research.
The complexity of the questions to be addressed is constantly increasing, requiring people able to harness ever stronger computing resources, develop new algorithms, data analysis and machine learning techniques, to further our understanding of nature and to improve the world we live in. The NGCM is training a new generation of researchers with the required skillset to address these questions.
This months summer academy provided workshops in data visualisation techniques that can make data accessible and easier to understand, code optimisation that can make computer code run faster, theory and practical examples in the transformative discipline of machine learning and the increasingly important computing platform of coding Graphics Processing Units (GPUs).
International experts Dr Prabhu Ramachandran from the Indian Institute of Technology in Bombay, Adrian Jackson from the Edinburgh Parallel Computing Centre and Jony Castanga from the Science and Technologies Facilities Council delivered courses on VTK and Mayavi, Optimising Scientific Software and GPU programming using CUDA.
The NGCM provides a four-year PhD programme that consists of a year of taught material before students spend three years dedicated to a research topic. The CDT benefits from a dedicated physical space in Boldrewood Innovation Campus and an extended network of industrial and academic partners.