Machine learning molecular simulations with physical constraints

Molecular simulations have become a powerful approach, highly complementary to experiments, to interrogate the structure and dynamics of macromolecular and colloidal systems. The resulting datasets are however typically highly dimensional, complex and noisy, making the extraction of useful information difficult. Artificial intelligence techniques such as deep learning are remarkable in their capacity of extracting and presenting useful information even in situations where humans would struggle. This project aims at the development of artificial intelligence-based methods for the analysis of molecular simulations and generation of models that are faithful to available experimental data, and consistent with physical laws. We will investigate the usage of generative models to enhance the sampling of molecular conformational space. These models will be exploited to discover new states and identify the specific transition path between two known states. Potential applications of the developed computational tools range from protein folding and colloidal self-assembly to formulation science.

1. PEOPLE. The SOFI2 CDT will have varied economic and societal impacts, the greatest of which will be the students themselves. They will graduate with a broad and deep scientific education as well as an entrepreneurial mind-set combined with business awareness and communication skills. The training programme reflects the knowledge and skills identified by industry partners, the EPSRC, recent graduates and national strategies. Partners will facilitate impact through their engagement in the extensive training programme and through the co-supervision of PhD projects. Responsible Innovation is embedded throughout the training programme to instil an attitude towards research and innovation in which societal concerns and environmental impact are always to the fore. The team-working and leadership skills developed in SOFI2 (including an appreciation of the benefits that diversity brings to an organisation and how to foster an atmosphere of equality and inclusion) will enable our graduates to take on leadership roles in industry where they can, in turn, influence the thinking of their teams.

2. PROJECTS. The PhD research projects themselves are impact pathways. Approximately half the projects will be co-sponsored by external partners and will be aligned to scientific challenges faced by the partner. Even projects funded entirely by the EPSRC/Universities will have an industrial co-supervisor who can provide advice on development of impact. The impact workshops and Entrepreneur in Residence will additionally help students to develop impact from their research, while at the same time developing the mind-set that sees innovation in invention.

3. PUBLIC. The public benefits from innovation that comes from the research in the CDT. It also benefits from the training of a generation of researchers trained in RI who seek out the input of stakeholders in the development of products and processes. The public benefits from the outreach activities that enable them to understand better the science behind contemporary technological developments - and hence to make more informed decisions about how they lead their lives. The younger generations benefit from the excitement of science that might attract them to higher education and careers in STEM subjects.

4. PARTNERSHIPS. SOFI2 involves collaborative research with >25 external partners from large multinationals to small start-ups. In addition to the results of sponsored projects and the possibility of recruiting SOFI2 students, companies benefit from access to training resources, sharing of best practice in RI and EDI, access to the knowledge of the SOFI2 academics and sharing of expertise with other partners in the SOFI2 network. This networking is of particular benefit to SMEs and we have an SME strategy to facilitate engagement of SMEs with SOFI2. SME representation on the Management and Strategic Advisory Boards will support the SME strategy.

CPI/NFC is a key partner both for delivery of training and to connect SOFI2 research, students and staff to a wide network of companies in the formulated products sector.

The unusual partnership with the Leverhulme Research Centre on Forensic Science may lead to a stronger scientific underpinning of forensic evidence with positive impacts on the legal process and the pursuit of justice.

5. PRODUCTS. Partner companies identify areas of fundamental and applied science of interest to them with the knowledge that advances in these areas will help them to overcome technological challenges that will lead to better products or new markets. It is an expectation that scientific discoveries made within the CDT will drive new products, new markets and potentially new companies. SOFI2 CDT seeks also to develop innovative training materials, for example, in RI and in data analytics and AI (in collaboration with the Alan Turing Institute), from which other CDTs and training organisations can benefit.