Abstract:Error bounds are derived for sampling and estimation using a discretization of an intrinsically defined Langevin diffusion with invariant measure $\text{d}\mu_\phi \propto e^{-\phi} \mathrm{dvol}_g $ on a compact Riemannian manifold. Two estimators of linear functionals of $\mu_\phi $ based on the discretized Markov process are considered: a time-averaging estimator based on a single trajectory and an ensemble-averaging estimator based on multiple independent trajectories. Imposing no restrictions beyond a nominal level of smoothness on $\phi$, first-order error bounds, in discretization step size, on the bias and variance/mean-square error of both estimators are derived. The order of error matches the optimal rate in Euclidean and flat spaces, and leads to a first-order bound on distance between the invariant measure $\mu_\phi$ and a stationary measure of the discretized Markov process. This order is preserved even upon using retractions when exponential maps are unavailable in closed form, thus enhancing practicality of the proposed algorithms. Generality of the proof techniques, which exploit links between two partial differential equations and the semigroup of operators corresponding to the Langevin diffusion, renders them amenable for the study of a more general class of sampling algorithms related to the Langevin diffusion. Conditions for extending analysis to the case of non-compact manifolds are discussed. Numerical illustrations with distributions, log-concave and otherwise, on the manifolds of positive and negative curvature elucidate on the derived bounds and demonstrate practical utility of the sampling algorithm.

Moment estimates, exponential integrability, concentration inequalities and exit times estimates on evolving manifolds

Convergence to the uniform distribution of moderately self-interacting diffusions on compact Riemannian manifolds

Measure estimation on a manifold explored by a diffusion process

Entropy-minimizing dynamical transport on Riemannian manifolds

From bounds to Gromov–Hausdorff convergence of Riemannian manifolds

Theory and Algorithms for Diffusion Processes on Riemannian Manifolds

Sampling and estimation on manifolds using the Langevin diffusion

Smoothing effects and extinction in finite time for fractional fast diffusions on Riemannian manifolds

Estimates on moments of the solutions to stochastic differential equations with respect to martingales in the plane

The Long Time Limit of Diffusion Means

The continuity equation for Hermitian metrics: Calabi estimates, Chern scalar curvature and Oeljeklaus-Toma manifolds

Second-order differential operators, stochastic differential equations and Brownian motions on embedded manifolds

Stability and moment estimates for the stochastic singular $Φ$-Laplace equation

Some Estimates of Fundamental Solution on noncompact manifolds with time-dependent metrics

Curvature-dimension condition, rigidity theorems and entropy differential inequalities on Riemannian manifolds

A practical global existence and uniqueness result for stochastic differential equations on Riemannian manifolds of bounded geometry

Functional Inequalities for Brownian Motion on Riemannian Manifolds with Sticky-Reflecting Boundary Diffusion

Important Moments in Systems and Control

Uniform estimates of Green functions and Sobolev-type inequalities on real and complex manifolds

Smooth center-stable/unstable manifolds and foliations of stochastic evolution equations with non-dense domain

Rational Points Near Manifolds, Homogeneous Dynamics, and Oscillatory Integrals