Agent-based models are increasingly used throughout industry and academia, in areas ranging from financial modelling to logistics and supply chain management, where they are used to model complex systems down at the level individual actors/decision-makers. Agent-based models allow us to capture aspects of systems (such as emergent properties, which arise from the interaction of many agents) that conventional modelling does not permit. Agent-based modelling came to international prominence when an agent-based epidemiological model of COVID-19 was revealed as one of the key drivers behind the UK government's decision to enter a lockdown in March 2020 . Although they are widely used, as an engineering discipline agent-based modelling remains in its infancy, and subsequent criticisms of the COVID-19 model highlighted many difficulties currently associated with agent-based modelling. First, current agent-based modelling environments force us to embed key assumptions directly within code, thereby obfuscating such assumptions and making it hard to understand them (clearly essential for situations such as the COVID model). Second, we need better ways of populating such models with realistic agent behaviours. Third, such models are limited in the extent to which we can rely on their predictions: we do not know how to calibrate such models (crudely: how can we be confident that $1 in a simulation corresponds to $1 in the real world?) Third, we have no available methodology for validating such models: existing techniques (e.g., model checking, used for formally verifying that systems satisfy their requirements) are unsuitable in their present form for agent-based models. The overarching technical goal of this project is to effect a step change in our ability to develop and deploy robust large scale agent-based simulations. Using state of the art techniques in AI and machine learning, we will carry out the fundamental research to develop the scientific & engineering methodology necessary to transform our capability in each of the areas identified above: allowing us to develop, populate, calibrate, and validate agent-based models at scale. Working with major industrial partners, we will test and refine our techniques on a range of real-world case studies. If successful, then this project will transform agent-based modelling from an ad hoc, trial and error process into a robust engineering discipline with a rigorous methodological foundation. It will establish Oxford as the world leader in the applications and analysis of multi-agent systems, and consolidate the UK's existing strengths in this area. Given our previous experience with agent-based financial modelling, we expect our results will be of considerable scientific interest and will have direct commercial value.

The CDT will train the next generation of leaders in statistics and statistical machine learning, who will be able to develop widely-applicable novel methodology and theory, as well as create application-specific methods, leading to breakthroughs in real-world problems in government, medicine, industry and science. The research will focus on the development of applicable modern statistical theory and methods as well as on the underpinnings of statistical machine learning. The research will be strongly linked to applications. There is an urgent national need for graduates from this CDT. Large volumes of complicated data are now routinely collected in all sectors of society, encompassing electronic health records, massive scientific datasets, governmental data, and data collected through the advent of the digital economy. The underpinning techniques for exploiting these data come from statistics and machine learning. Exploiting such data is crucial for future UK prosperity. However, several reports from government and learned societies have identified a lack of individuals able to exploit this data. In many situations, existing methodology is insufficient. Off-the-shelf approaches may be misleading due to a lack of reproducibility or sampling biases which they do not correct. Furthermore, understanding the underlying mechanisms is often desired: scientifically valid, interpretable and reproducible results are needed to understand scientific phenomena and to justify decisions, particularly those affecting individuals. Bespoke, model-based statistical methods are needed, that may need to be blended with statistical machine learning approaches to deal with large data. Individuals that can fulfill these more sophisticated demands are doctoral level graduates in statistics who are well versed in the foundations of machine learning. Yet the UK only graduates a small number of statistics PhDs per year, and many of these graduates will not have been exposed to machine learning. The Centre will bring together Imperial and Oxford, two top statistics groups, as equal partners, offering an exceptional training environment and the direct involvement of absolute research leaders in their fields. The supervisor pool will include outstanding researchers in statistical methodology and theory as well as in statistical machine learning. We will use innovative and student-led teaching, focussing on PhD-level training. Teaching cuts across years and thus creates strong cohort cohesion not just within a year group but also between year groups. We will link theoretical advances to application areas through partner interactions as well as through a placement of students with users of statistics. The CDT has a large number of high profile partners that helped shape our application priority areas (digital economy, medicine, engineering, public health, science) and that will co-fund and co-supervise PhD students, as well as co-deliver teaching elements.