Umberto Picchini (Chalmers University of Technology and University of Gothenburg, Sweden): Scalable Bayesian inference in stochastic nonlinear mixed-effects models using semi-amortised likelihood and posterior approximations

Time and place:
March 31st, 15–16
CK111 (Exactum), University of Helsinki

Abstract: The analysis of data from multiple experiments, such as observations of several individuals, is commonly approached using mixed-effects models, which account for variation between individuals through hierarchical representations. This makes mixed-effects models widely applied in fields such as biology, pharmacokinetics, and sociology. In this talk, we discuss the methodology in [1] for scalable Bayesian inference in hierarchical mixed-effects models. Our framework first constructs amortized approximations of the likelihood and the posterior distribution, which are then rapidly refined for each individual dataset, to ultimately approximate the parameters posterior across many individuals. The framework is easily trainable, as it uses mixtures of experts but without neural networks, leading to parsimonious yet expressive surrogate models of the likelihood and the posterior. We demonstrate the effectiveness of our methodology using challenging stochastic models, such as mixed-effects stochastic differential equations emerging in systems biology-driven problems. However, the approach is broadly applicable and can accommodate both stochastic and deterministic models. We show that our approach can seamlessly handle inference for many parameters. Additionally, we applied our method to a real-data case study of mRNA transfection. When compared to exact pseudomarginal Bayesian inference, our approach proved to be both fast and competitive in terms of statistical accuracy.
[1] H. Häggström, S. Persson, M. Cvijovic and U. Picchini (2026). Simulation-based inference for stochastic nonlinear mixed-effects models with applications in systems biology, Statistics and Computing, 36(99).

Bio: Umberto Picchini is interested in Bayesian inference for stochastic modelling, stochastic differential equations, simulator-based methods for inference, applications in biology and medicine. He prefers music pre 2000s. He is a professor in Mathematical Statistics at Dept Mathematical Sciences in Göteborg, Sweden. Previously he was Associate Professor in Lund (Sweden) and a lecturer in Durham (UK).