Discovery of new materials for applications on glass using Mathematical Optimisation and Machine Learning methodologies (Ref NSGPVCS2023)

Description

The functionalisation of flat glass is a multi-billion-dollar industry spanning applications including displays technology, energy saving windows and energy generation through photovoltaics. To maintain the rate of progress in these and other emerging fields new materials are required with performance beyond that of known materials.

The interplay of many considerations including structure, bonding, and defect chemistry makes for a challenging opportunity to develop a new material that is stable and has interesting functional properties that can be applied as a thin film on glass. Mathematical modelling and optimisation which are ubiquitous in machine learning and artificial intelligence, are also in the core of a multitude of complex problems in chemistry and material science, such as prediction of new functional materials with desirable properties, statistical characterisation of data, experimental design, production control and planning.

This PhD project will study and apply existing optimisation methods and propose methods and mechanisms for novel ones. In particular, focus will be given to black-box optimisation methods capable of handling difficult complex problems that are difficult or impossible to model directly. Examples include grid, coordinate and pattern searches, metaheuristics, model-based methods, surrogate models, evolutionary optimisation methods such as genetic algorithms or ones utilising natural gradient and information geometry, and Bayesian optimisation. Many of them relay heavily on the use of various machine learning and statistical mechanisms to adapt to the optimisation landscape and automatically collect data where applicable.

The student will work closely with computer scientists, inorganic chemists, physicists, and material scientists to develop tools to predict new materials that may exhibit desirable properties. The developed or improved optimisation algorithms will go far beyond classical approaches deployed by physical computational science researchers thus far in the literature.

Qualifications: Applications are welcomed from students with a preferably 1^st or 2.1 class BSc or equivalent in Mathematics, Computer Science or Physics. Previous experience with optimisation and machine learning (or any knowledge in chemistry) is not a requirement, though successful candidates must have strong mathematical and programming skills.

Please apply by completing the online postgraduate research application form here: How to apply for a PhD - University of Liverpool  

Please ensure you quote the following reference on your application: Discovery of new materials for applications on glass using Mathematical Optimisation and Machine Learning methodologies (Reference NSGPVCS2023) and make the application to Computer Science