Bio
Peter is a cross-disciplinary researcher with a track record of publications in machine learning, statistical physics and applied mathematics. He is also an honorary research fellow at Imperial College London, where he is involved in collaborations and co-supervision of MSc/PhD projects.
A full list of publications, CV and teaching statement are available on Peter’s personal website.
Peter's work at The Alan Turing Institute lies at the interface of applied probability, computational statistics and machine learning, with a particular focus on complex systems exhibiting dynamics across multiple spatial and temporal scales, including many-body systems, multi-agent systems, fluids, and, more recently, climate physics.
Peter works on methodology and tools which underpin digital twins for complex physical and engineering systems. Here Peter developed a number of novel data-driven techniques and algorithms, including data-driven coarse-graining of many-particle systems using physics-informed Bayesian learning, automated discovery of low-dimensional data manifolds and digital twinning of complex black-box systems using priors based on generative models. While these challenges were motivated and informed by problems in data-centric engineering, their potential impact reaches far beyond the general sphere of engineering.