PhD Opportunity in Deep Learning and Scientific Modelling
We welcome applications for a four-year D.Phil in Mathematics at the University of Oxford on the topic of using deep learning to develop reduced-order partial differential equation (PDE) models of computationally-challenging physics. The research project will be supervised by Professor Justin Sirignano. The PhD studentship is open to all applicants, irrespective of nationality/residence, and the start date is 01 October 2022. The successful candidate will be based in the Mathematical Institute at the University of Oxford. The PhD position is fully-funded for four years, including a stipend (approximately £16,000 per year) and university fees (e.g., tuition).
Accurate, efficient simulation of non-equilibrium fluid mechanics remains a major challenge for high-speed flows and hypersonic flight. Although the governing equations from physics are known, their numerical solution is computationally expensive or intractable for real-world applications. Boltzmann equation solutions using direct-simulation Monte Carlo (DSMC) methods are prohibitively expensive at transitional Knudsen numbers and for realistic chemistry, while Navier-Stokes solutions at these conditions are known to inaccurately predict shock thickness and shock-boundary layer interaction due to the breakdown of the continuum assumption.
This research project will use machine learning to develop reduced-order PDE models of non-equilibrium effects in high-speed gas flows. Our goal is to substantially advance the state-of-the-art by developing a fast, accurate model which uses deep learning to augment the computationally resolvable physics. The research will be highly interdisciplinary, involving deep learning, applied mathematics, physics, and high-performance computing (HPC).
The project has a multi-university research team, consisting of experts in machine learning, computational fluid dynamics, and hypersonics. There will be close collaboration between the principal investigators and PhD students. Students will have access to HPC/super-computing GPU resources for the development and simulation of models. Students will also have the opportunity to interact with the larger engineering and applied science communities.
APPLICATIONS SHOULD BE MADE ONLINE TO THE MATHEMATICAL INSTITUTE at https://evision.ox.ac.uk/urd/sits.urd/run/siw_ipp_lgn.login?process=siw_ipp_app_crs and should include a CV, cover letter (1 page), 2-3 reference letters, and transcript(s) of all previous degrees. In the section of the application form “Departmental Studentship Applications” applicants will be asked whether they are applying for an advertised studentship. In this section please state “Yes” followed by “22MATH01WEB”.
Applications will open on 1 June 2022 and must arrive by Noon on 22 June, 2022.
This studentship is linked to Somerville College.
For further information please contact @email.