Overview
Mark Girolami is a statistician (Sir Kirby Laing Professor of Civil Engineering and Royal Academy of Engineering Research Chair, University of Cambridge; former Chief Scientist of The Alan Turing Institute) known for geometric and gradient-based MCMC and probabilistic numerics.
Key Contributions / Features
- Riemannian-manifold MCMC (Girolami–Calderhead 2011): Hamiltonian and Langevin MCMC on a Riemannian manifold with a position-dependent metric, adapting proposals to local curvature. Journal of the Royal Statistical Society, Series B 73(2): 123–214.
- HMC for hierarchical models (Betancourt–Girolami 2015): the funnel pathology and non-centered parameterization.
- Probabilistic numerics: foundational work treating numerical computation as Bayesian inference.
Related