Overview
Michael Betancourt is a statistician and physicist known for his work on Hamiltonian Monte Carlo, the differential geometry of MCMC, and Bayesian workflow, and as a core developer of the Stan probabilistic-programming language.
Key Contributions / Features
- HMC for hierarchical models (Betancourt–Girolami 2015): geometric account of the funnel pathology and the role of non-centered parameterization and divergence diagnostics. In Current Trends in Bayesian Methodology with Applications, Chapman & Hall/CRC.
- Riemannian-manifold and general-metric HMC: contributions to position-dependent metrics for HMC (with Girolami and others).
- Expository work: widely used conceptual introductions to HMC and Bayesian workflow.
Related