Andrew Duncan is RAEng Lecturer (Assistant Professor) in the Department of Mathematics at Imperial College London and group leader for the Data-Centric Engineering Programme at The Alan Turing Institute.
Andrew's research interests span applied stochastic modelling, modelling and inference in a variety of applications including aerospace engineering, energy systems and predictive health monitoring for complex industrial processes.
His theoretical research pertains to methods for approximate inference, anomaly and change-point detection for complex statistical models based on Stein’s method and related methodology. His recent work includes developing methods for assessing convergence of MCMC samplers and for learning intractable generative models.
Andrew's interests lie generally within the intersection between computation, analysis and probability, with a particular focus on applications in biology and chemistry, particularly systems involving stochasticity and/or multiple scales. These include the analysis and construction of MCMC-based methods for sampling from probability distributions, coarse graining of stochastic models involving multiple scales, classical and stochastic homogenisation of PDEs and SDEs, and the Bayesian formulation of inverse problems.