Centre of Data Innovation Research (CoDIR)

Purkinje cellGalaxy formation

The Centre of Data Innovation Research (CoDIR) is a virtual research centre that combines data science and machine learning expertise from astrophysics and biocomputation. We have come to recognise that many difficult problems in both fields can be tackled by the same techniques of data analysis and machine learning. Often techniques developed for the benefit of what seems to be a highly specific problem in one field can be productively deployed to solve seemingly unrelated problems in another. Our distinctive plan is to not only pioneer innovative data science techniques – in particular exploiting expertise in machine learning and computational neuroscience – to advance the field of survey astronomy, but also develop a strategy to translate these techniques to other disparate areas, with a particular focus on medicine, neuroscience, defence and agritech. By embedding this translational approach in our research programme, which has the potential to deliver genuine impact, we seek to become an exemplar research centre that will serve as a blueprint for others aiming to deliver economic and societal impact from blue skies research.

Within the Centre for Astrophysics Research (CAR) we perform world-leading research spanning the discovery of exoplanets, the nature of star formation within our Milky Way and the formation and evolution of galaxies. Computational neuroscience and machine learning, a focus of the Biocomputation Research Group in the Centre for Computer Science and Informatics Research (CCSIR), are thriving areas of interdisciplinary research at the interface between computer science, mathematics and biology. The approach taken by the Biocomputation group is unique as it combines, in addition to computational modelling and mathematical analyses, cutting-edge work in neuromorphic computing, neural data analysis and machine learning. In the past few years we have (1) recognised an urgent need to apply the very latest data science and machine learning techniques to analyse ‘big’ astronomical and biological data, and (2) identified clear opportunities to apply our expertise to problems in fields outside our discipline, not only in other scientific fields (such as Life and Medical Sciences), but also in practical areas closely aligned to the Government's Industrial Strategy (e.g. agritech, data economy). To this end we have initiated an internal cross-disciplinary collaboration between CAR and CCSIR that is delivering not only scientific results (Hocking et al. 2018), but also demonstrated the potential for impact in the private sector (e.g. through a successful on-going collaboration with an industrial partner). Currently, this is a small cross-disciplinary research group led by Geach and Steuber.