The Alan Turing Institute and the University of Sydney's Centre for Translational Data Science have signed a memorandum of understanding to collaborate on joint research projects of strategic importance to the Australian economy, including criminology, air quality, and geosciences. The collaboration will be centred around The Alan Turing Institute’s data-centric engineering programme, a major research programme funded by the Lloyd's Register Foundation.
The Centre for Translational Data Science was established in the University of Sydney in 2016 and is one of only a few centres worldwide that has a research focus of modelling complex phenomena and the translation of this research into practical outcomes which benefit society.
"Three areas of collaboration in data science already underway are space-time models for environmental air quality monitoring and improvement, probabilistic modelling for Australia’s natural resources, and Bayesian optimisation for criminology. All of these areas of collaboration will make a real difference in the world," said Professor Sally Cripps, Co-Director of the Centre for Translational Data Science.
Air quality and improvement
The first collaboration focuses on an air quality and improvement project. The data-centric engineering programme is already working with the London Mayor’s Office on a joint project to understand and improve air quality over London by developing advanced spatio-temporal statistical and machine learning methods for estimating and forecasting air pollution levels at a hyper-local scale. These are further linked to critical monitoring stations and policy interventions.
The Centre for Translational Data Science is working with the Nature Conservation Fund to develop statistical machine learning models to assess the impact that ‘greening’ cities has on air quality, and how this improvement in air quality affects health outcomes, vital for sustainable cities of the future.
Geosciences and natural resources
The second project has the two centres working with statisticians, machine learners and earth scientists from the Universities of Sydney and Western Australia to use the latest advances in data science to transform the process by which decisions are made in the management of natural resources.
Discussions are underway with IAG, McKinsey, Rio Tinto, Lloyd's Registry, and government agencies to form a multidisciplinary partnership to drive transformational advances in the earth sciences to build scale and human capacity in the resources and environment industries.
“The world’s economic, societal and environmental future depends upon the balance we place on competing outcomes when making decisions and policy surrounding the use of our natural resources. Despite the importance of this issue, many decisions and policies are based on limited amounts of data and rudimentary data analysis, with little appreciation of the critical role that understanding and quantifying uncertainty plays in the process.
The collaboration will allow data science researchers in Australia to tap into the wealth of experience that exists within the Turing to revolutionise the way decisions are made in this important industry sector,” said Professor Cripps.
“This collaboration opens up huge potential for exciting new projects. We look forward to exchanging researchers with the University of Sydney to strengthen the existing Memorandum of Understanding and jointly-funded collaborations we have,” says Professor Mark Girolami, Programme Director for Data-Centric Engineering, The Alan Turing Institute
“This partnership represents an unrivalled opportunity to transform many domestic and international industries. The combined skill set of the international collaboration between the two institutions will result in outstanding outcomes that neither could achieve alone,” concluded Professor Girolami.
Criminology
The Centre for Translational Data Science criminology research is led by Dr Roman Marchant, who works closely with Turing Fellows, Associate Professor Theo Damoulas and PhD student Louis Ellam.
“Statistical models can help us to understand criminal behaviour and determine the key drivers and dynamics of crime. In particular, to achieve a future reduction in crime arising from data-driven and informed policy decisions,” says Dr Merchant.
Their collaboration is developing new Bayesian optimisation algorithms to uncover hidden patterns in criminal activity.
“Modelling the dependence between different types of crime leads to greater understanding of the dynamics criminal behaviour. The Turing is looking forward to continuing this research with colleagues at the Centre for Translational Data Science,” said Associate Professor Damoulas.
