CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: On some variations of Riemannian manifold and Lagrangian Monte Carlo Authors:  Vivekananda Roy - Iowa State University (United States) [presenting]
Abstract: Diffusions based and Hamiltonian dynamics-based methods such as the Metropolis adjusted Langevin algorithms (MALA), and Hamiltonian Monte Carlo (HMC) algorithms have emerged as powerful Metropolis-Hastings algorithms. IWe consider some variations of the Riemannian manifold HMC (RMHMC) and Lagrangian Monte Carlo(LMC) methods. In particular, we investigate the mixtures of the LMC and RMHMC transition kernels with the manifold MALA kernels. The resulting algorithms are shown to converge at a geometric rate under certain conditions. The algorithms are illustrated using several examples.