Ergodic SDEs on submanifolds and related numerical sampling schemes

  title={Ergodic SDEs on submanifolds and related numerical sampling schemes},
  author={Wei Zhang},
  journal={ESAIM: Mathematical Modelling and Numerical Analysis},
  • Wei Zhang
  • Published 26 February 2017
  • Mathematics
  • ESAIM: Mathematical Modelling and Numerical Analysis
In many applications, it is often necessary to sample the mean value of certain quantity with respect to a probability measure μ on the level set of a smooth function ξ : ℝd → ℝk, 1 ≤ k < d. A specially interesting case is the so-called conditional probability measure, which is useful in the study of free energy calculation and model reduction of diffusion processes. By Birkhoff’s ergodic theorem, one approach to estimate the mean value is to compute the time average along an infinitely long… 

Figures from this paper

Non-reversible sampling schemes on submanifolds
This paper presents and analyzes a non-reversible generalisation of the projection-based scheme developed by one of the authors, and proves the consistency of this numerical scheme and provides quantitative error estimates for estimators based on finite-time running averages.
Constrained overdamped Langevin dynamics for symmetric multimarginal optimal transportation
This work designs a numerical method, built upon constrained overdamped Langevin processes to solve Moment Constrained Optimal Transport (MCOT) relaxations, and proves that there is no strict local minimizer to the resulting problem.
Order conditions for sampling the invariant measure of ergodic stochastic differential equations on manifolds
A new methodology for the construction of high order integrators for sampling the invariant measure of ergodic stochastic differential equations with dynamics constrained on a manifold is derived using an extension of the exotic aromatic Butcher-series formalism.
Multiple projection MCMC algorithms on submanifolds
We propose new Markov Chain Monte Carlo algorithms to sample probability distributions on submanifolds, which generalize previous methods by allowing the use of set-valued maps in the proposal step
Multiple projection Markov Chain Monte Carlo algorithms on submanifolds
We propose new Markov Chain Monte Carlo algorithms to sample probability distributions on submanifolds, which generalize previous methods by allowing the use of set-valued maps in the proposal step
Newton retraction as approximate geodesics on submanifolds
It is shown that Newton retraction (NR) is always stabler than the popular approach known as oblique projection or orthographic retraction: per Kantorovich-type convergence theorems, the superlinear convergence regions of NR include those of the latter.
Coulomb Gases Under Constraint: Some Theoretical and Numerical Results
This work shows that the special case of quadratic confinement and linear constraint is exactly solvable due to a remarkable factorization, and that the conditioning has then the simple effect of shifting the cloud of particles without deformation.
A uniformly accurate scheme for the numerical integration of penalized Langevin dynamics
This paper proposes a new consistent method with an accuracy independent of ε for solving penalized dynamics on a manifold of any dimension and converges to the constrained Euler scheme when ε goes to zero.
Jarzynski’s Equality, Fluctuation Theorems, and Variance Reduction: Mathematical Analysis and Numerical Algorithms
In this paper, we study Jarzynski’s equality and fluctuation theorems for diffusion processes. While some of the results considered in the current work are known in the (mainly physics) literature,
Stochastic Deep-Ritz for Parametric Uncertainty Quantification
This work proposes a deep learning based numerical method for solving elliptic partial differential equations (PDE) with random coeficcients by elucidate the stochastic variational formulation for the problem by recourse to the direct method of calculus of variations.


Projection of diffusions on submanifolds: Application to mean force computation
It is proved that the constrained stochastic differential equations proposed in [7, 13] are ergodic with respect to this restricted distribution and constructed numerical schemes for the integration of the constrained diffusions.
Hybrid Monte Carlo methods for sampling probability measures on submanifolds
This work shows here how to generalize to GHMC a procedure suggested by Goodman, Holmes-Cerfon and Zappa for Metropolis random walks on submanifolds, where a reverse projection check is performed to enforce the reversibility of the algorithm for large timesteps and hence avoid biases in the invariant measure.
Convergence of Numerical Time-Averaging and Stationary Measures via Poisson Equations
Numerical approximation of the long time behavior of a stochastic differential equation (SDE) is considered. Error estimates for time-averaging estimators are obtained and then used to show that the
Langevin dynamics with constraints and computation of free energy differences
A consistent discretization of the overdamped Langevin (Brownian) dynamics on a submanifold is obtained, also sampling exactly the correct canonical measure with constraints, in some limiting regime.
Level set approach to mean curvature flow in arbitrary codimension
We develop a level set theory for the mean curvature evolution of surfaces with arbitrary co-dimension, thus generalizing the previous work [8, 15] on hypersurfaces. The main idea is to surround the
Monte Carlo on Manifolds: Sampling Densities and Integrating Functions
An MCMC sampler is given for probability distributions defined by unnormalized densities on manifolds in euclidean space defined by equality and inequality constraints and used to develop a multistage algorithm to compute integrals over such manifolds.
High Order Numerical Approximation of the Invariant Measure of Ergodic SDEs
A systematic procedure based on the framework of modified differential equations for the construction of stochastic integrators that capture the invariant measure of a wide class of ergodic SDEs (Brownian and Langevin dynamics) with an accuracy independent of the weak order of the underlying method.
Weak Backward Error Analysis for SDEs
A weak backward error analysis result is shown that the generator associated with the numerical solution coincides with the solution of a modified Kolmogorov equation up to high order terms with respect to the stepsize, implying that every invariant measure of the numerical scheme is close to a modified invariants measure obtained by asymptotic expansion.
Degenerative convergence of diffusion process toward a submanifold by strong drift
Let be a diffusion process on a Riemannian manifold Nhaving an infinitesimal generator . We investigate the asymptotic behavior of as the strength β of the drift becomes large under the situation
A Family of MCMC Methods on Implicitly Defined Manifolds
A general constrained version of Hamiltonian Monte Carlo is proposed, and conditions under which the Markov chain is convergent are given, which define a family of MCMC methods which can be applied to sample from distributions on non-linear manifolds.