Enhancing Machine Learning with Physical Constraints to Predict Microstructure Evolution

De-mixing is one of the most ubiquitous examples of self-assembly, occurring frequently in complex fluids and living systems. It has enabled the development of multi-phase polymer alloys and composites for use in sophisticated applications including structural aerospace components, flexible solar cells and filtration membranes. In each case, superior functionality is derived from the microstructure, the prediction of which has failed to maintain pace with synthetic and formulation advances. The interplay of non-equilibrium statistical physics, diffusion and rheology causes multiple processes with overlapping time and length scales, which has stalled the discovery of an overarching theoretical framework. Consequently, we continue to rely heavily on trial and error in the search for new materials.

Our aim is to introduce a powerful new approach to modelling non-equilibrium soft matter, combining the observation based empiricism of machine learning with the fundamental based conceptualism of physics. We will develop new methods in machine learning by addressing the broader challenge of incorporating prior knowledge of physical systems into probabilistic learning rules, transforming our capacity to control and tailor microstructure through the use of predictive tools. Our goal is to create empirical learning engines, constrained by the laws of physics, that will be trained using microscopy, tomography and scattering data. In this feasibility study, we will focus on proof-of-concept, exploring the temperature / composition parameter space for a model blend, building the foundations for our ambition of using physics informed machine learning to automate and accelerate experimental materials discovery for next generation applications.

From an economic perspective, the most immediate beneficiaries are those companies that focus on multiphase materials discovery, development and optimisation with applications ranging from personal care products to aerospace and energy. The UK has a strong tradition of hosting the research and development activities for such knowledge-based companies, many of which have their own in-house modelling groups. Our program of research will open up new approaches to materials discovery by enabling traditional materials modelling techniques to be enhanced with the power of machine learning. By improving and refining models with learning from prior experimental data, we will enable acceleration of materials development, thus contributing to the need for industry to innovate quickly in order to remain competitive.

From the perspective of public engagement, the beneficiaries will be physical scientists who may not have previously considered a career in artificial intelligence as a viable option. We will focus activity, through the Sheffield University Diversity in the Cultures of Physics - International Summer School, which is aimed at women. This is a group that is under represented in the rapidly growing AI branch of the tech industry. We will use the School as an opportunity to introduce women undergraduates to machine learning and enable them to explore, through involvement with the research, how a background in physical sciences provides a legitimate background for a future career in AI. We also aim to use our research as a springboard to engage in conversation with the public about the future role of AI within society. Given that the media is naturally focussed on the highest profile, and often controversial, developments in AI, we are keen to broaden the discussion to include the impact on society of research developments that are traditionally off the radar of non-specialists