Connect
About
Thomas is a Postdoctoral Researcher in Computational Social Science at the Department of Methodology.
He earned his PhD in Mathematics from the University of Cambridge in 2025, and holds Master's and Bachelor's degrees in Physics and Mathematics. His research focuses on hybrid neural models that integrate deep learning with mechanistic modelling to study complex dynamical systems, with applications in social and economic processes as well as infectious disease dynamics. He is a member of the Humanet Lab, working with Milena Tsvetkova on human–machine and machine–machine interaction
Thomas’s research focuses on integrating deep learning with mechanistic models (such as PDEs and SDEs) to study complex dynamical systems. Social systems provide particularly rich applications, as they exhibit self-organisation, interdependence, and latent dynamics. By augmenting—rather than replacing—mechanistic models with neural networks, his work aims to develop more expressive yet interpretable models that also mitigate overfitting. He has applied this framework to human migration flows, COVID-19 infection dynamics, and the global trade of food and agricultural products. Beyond hybrid modelling, he is interested in network dynamics and social network analysis, and has examined populist mobilisation on Twitter. His current research investigates the interactions between humans and machines in networked social systems.
Publications
Deep learning four decades of human migration (2025) with Guy Abel; under review at Nature.
Populist narrative power in a globalised infosphere (2024) with Zhu Yi; under review at Humanities and Social Sciences Communications.
Modelling Global Trade with Optimal Transport (2024) with G. Demirel, M-T. Wolfram, A. Duncan; under review at Nature Communications.
Neural parameter calibration and uncertainty quantification for epidemic forecasting (2024) with T. Conrad, G. Pavliotis, C. Schütte. PLoS ONE 19 (10): e0306704.
Inferring networks from time series: A neural approach (2024) with G. Pavliotis and M. Girolami. PNAS Nexus 3: 4.
Neural parameter calibration for large-scale multi-agent systems (2023) with G. Pavliotis and M. Girolami. PNAS 120 (7).