Proximal Markov chain Monte Carlo algorithms

Marcelo Pereyra

This paper presents a new Metropolis-adjusted Langevin algorithm (MALA) that uses convex analysis to simulate efficiently from high-dimensional densities that are log-concave, a cl…