REVIEW 5 cited by
Sampling via Gradient Flows in the Space of Probability Measures
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Sampling a target probability distribution with an unknown normalization constant is a fundamental challenge in computational science and engineering. Recent work shows that algorithms derived by considering gradient flows in the space of probability measures open up new avenues for algorithm development. This paper makes three contributions to this sampling approach by scrutinizing the design components of such gradient flows. Any instantiation of a gradient flow for sampling needs an energy functional and a metric to determine the flow, as well as numerical approximations of the flow to derive algorithms. Our first contribution is to show that the Kullback-Leibler divergence, as an energy functional, has the unique property (among all f-divergences) that gradient flows resulting from it do not depend on the normalization constant of the target distribution. Our second contribution is to study the choice of metric from the perspective of invariance. The Fisher-Rao metric is known as the unique choice (up to scaling) that is diffeomorphism invariant. As a computationally tractable alternative, we introduce a relaxed, affine invariance property for the metrics and gradient flows. In particular, we construct various affine invariant Wasserstein and Stein gradient flows. Affine invariant gradient flows are shown to behave more favorably than their non-affine-invariant counterparts when sampling highly anisotropic distributions, in theory and by using particle methods. Our third contribution is to study, and develop efficient algorithms based on Gaussian approximations of the gradient flows; this leads to an alternative to particle methods. We establish connections between various Gaussian approximate gradient flows, discuss their relation to gradient methods arising from parametric variational inference, and study their convergence properties both theoretically and numerically.
Forward citations
Cited by 5 Pith papers
-
Harnessing the Power of Reinforcement Learning for Adaptive MCMC
A contrastive-divergence-based reward stabilizes reinforcement learning of position-dependent step sizes for gradient-based MCMC, beating constant-step tuning on most of 44 benchmark posteriors.
-
An equivalence of moment closure and nonlinear variational approximation of the Fokker-Planck equation for dilute polymeric flow
For the linear Hookean chain, the Dirac–Frenkel Gaussian variational approximation of the Fokker–Planck equation reproduces exactly the classical diffusive Oldroyd-B moment-closure equations for the conformation tensor.
-
Fast Convergence for High-Order ODE Solvers in Diffusion Probabilistic Models
A TV convergence bound O(d^{7/4} ε^{1/2} + d(dH)^p) is proved for p-th order (exponential) Runge-Kutta samplers of probability-flow ODEs under C² smoothness of the learned score.
-
Gradient Flow Sampler-based Distributionally Robust Optimization
Entropy-regularized Wasserstein DRO can be solved by sampling from a Gibbs worst-case distribution with gradient-flow samplers, giving new WFR/SVGD algorithms and a principled recovery of WRM.
-
Solving Inverse Problems via Diffusion-Based Priors: An Approximation-Free Ensemble Sampling Approach
A weighted-particle sampler evolves the posterior through the diffusion model's reverse dynamics, with theoretical error bounds and improved image reconstructions.
Discussion (0). Sign in to comment.