Understanding MCMC Dynamics as Flows on the Wasserstein Space

02/01/2019
by   Chang Liu, et al.
0

It is known that the Langevin dynamics used in MCMC is the gradient flow of the KL divergence on the Wasserstein space, which helps convergence analysis and inspires recent particle-based variational inference methods (ParVIs). But no more MCMC dynamics is understood in this way. In this work, by developing novel concepts, we propose a theoretical framework that recognizes a general MCMC dynamics as the fiber-gradient Hamiltonian flow on the Wasserstein space of a fiber-Riemannian Poisson manifold. The "conservation + convergence" structure of the flow gives a clear picture on the behavior of general MCMC dynamics. We analyse existing MCMC instances under the framework. The framework also enables ParVI simulation of MCMC dynamics, which enriches the ParVI family with more efficient dynamics, and also adapts ParVI advantages to MCMCs. We develop two ParVI methods for a particular MCMC dynamics and demonstrate the benefits in experiments.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
10/07/2021

De-randomizing MCMC dynamics with the diffusion Stein operator

Approximate Bayesian inference estimates descriptors of an intractable t...
research
07/04/2018

Accelerated First-order Methods on the Wasserstein Space for Bayesian Inference

We consider doing Bayesian inference by minimizing the KL divergence on ...
research
05/16/2022

Ergodic variational flows

This work presents a new class of variational family – ergodic variation...
research
09/29/2020

Couplings for Andersen Dynamics

Andersen dynamics is a standard method for molecular simulations, and a ...
research
06/29/2022

Discrete Langevin Sampler via Wasserstein Gradient Flow

Recently, a family of locally balanced (LB) samplers has demonstrated ex...
research
02/23/2023

An Explicit Expansion of the Kullback-Leibler Divergence along its Fisher-Rao Gradient Flow

Let V_* : ℝ^d →ℝ be some (possibly non-convex) potential function, and c...
research
02/19/2023

Stochastic Generative Flow Networks

Generative Flow Networks (or GFlowNets for short) are a family of probab...

Please sign up or login with your details

Forgot password? Click here to reset