Geometry and Dynamics for Markov Chain Monte Carlo

Article, Preprint English OPEN
Barp, A; Briol, F-X; Kennedy, AD; Girolami, M;
(2017)
  • Publisher: Annual Reviews
  • Identifiers: doi: 10.1146/annurev-statistics-031017-100141
  • Subject: ALGORITHMS | Markov chain Monte Carlo | hep-lat | Mathematics | SIMULATION | math.NA | Computer Science - Learning | LANGEVIN | cs.LG | symplectic integrators | stat.ML | Physical Sciences | High Energy Physics - Lattice | PHASE-SPACE | SYSTEMS | INVERSE PROBLEMS | CONSTRUCTION | Mathematics - Numerical Analysis | FERMIONS | Statistics - Machine Learning | Statistics - Computation | Hamiltonian mechanics | DIFFUSIONS | stat.CO | Science & Technology | information geometry | Mathematics, Interdisciplinary Applications | Statistics & Probability | PARAMETERS | shadow Hamiltonians

Markov Chain Monte Carlo methods have revolutionised mathematical computation and enabled statistical inference within many previously intractable models. In this context, Hamiltonian dynamics have been proposed as an efficient way of building chains which can explore p... View more