Scaling Limit of the Stein Variational Gradient Descent: The Mean Field Regime

DOI10.1137/18M1187611zbMath1417.35189arXiv1805.04035OpenAlexW2964015695MaRDI QIDQ4627155

James Nolen, Yulong Lu, Jian-feng Lu

Publication date: 7 March 2019

Published in: SIAM Journal on Mathematical Analysis (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1805.04035

zbMATH Keywords

samplinginteracting particle systemmean field limitStein variational gradient descent

Mathematics Subject Classification ID

Interacting particle systems in time-dependent statistical mechanics (82C22) PDEs in connection with statistics (35Q62) PDEs in connection with computer science (35Q68)

Related Items (23)

Derivative-Free Bayesian Inversion Using Multiscale Dynamics ⋮ Random Batch Methods for Classical and Quantum Interacting Particle Systems and Statistical Samplings ⋮ Particle dual averaging: optimization of mean field neural network with global convergence rate analysis* ⋮ Sparse approximation of triangular transports. I: The finite-dimensional case ⋮ Birth–death dynamics for sampling: global convergence, approximations and their asymptotics ⋮ Stein Variational Gradient Descent on Infinite-Dimensional Space and Applications to Statistical Inverse Problems ⋮ Stein variational gradient descent with learned direction ⋮ A blob method for inhomogeneous diffusion with applications to multi-agent control and sampling ⋮ Interacting Langevin Diffusions: Gradient Structure and Ensemble Kalman Sampler ⋮ Physics-Informed Probabilistic Learning of Linear Embeddings of Nonlinear Dynamics with Guaranteed Stability ⋮ Stein's method meets computational statistics: a review of some recent developments ⋮ Stein variational gradient descent: many-particle and long-time asymptotics ⋮ Gradient-free Stein variational gradient descent with kernel approximation ⋮ Mean Field Limits for Interacting Diffusions with Colored Noise: Phase Transitions and Spectral Numerical Methods ⋮ Stein variational gradient descent with local approximations ⋮ Ensemble Kalman inversion: mean-field limit and convergence analysis ⋮ Particle-based energetic variational inference ⋮ Data assimilation: The Schrödinger perspective ⋮ Fokker--Planck Particle Systems for Bayesian Inference: Computational Approaches ⋮ Stein Variational Reduced Basis Bayesian Inversion ⋮ A stochastic version of Stein variational gradient descent for efficient sampling ⋮ Hessian transport gradient flows ⋮ Ensemble Kalman Sampler: Mean-field Limit and Convergence Analysis

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Stationary states of quadratic diffusion equations with long-range attraction
• On the rate of convergence in Wasserstein distance of the empirical measure
• Initiation of slime mold aggregation viewed as an instability
• Large time behavior of nonlocal aggregation models with nonlinear diffusion
• Vlasov equations
• Exponential convergence of Langevin distributions and their discrete approximations
• From 1970 until present: The Keller-Segel model in chemotaxis and its consequences. I
• Kinetic equilibration rates for granular media and related equations: entropy dissipation and mass transportation estimates
• Sharp asymptotic and finite-sample rates of convergence of empirical measures in Wasserstein distance
• Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
• Macroscopic and large scale phenomena: coarse graining, mean field limits and ergodicity. Based on the presentations at the summer school, Enschede, the Netherlands, 2012
• THE GEOMETRY OF DISSIPATIVE EVOLUTION EQUATIONS: THE POROUS MEDIUM EQUATION
• A blob method for the aggregation equation
• A Practical Guide to Deterministic Particle Methods
• A Deterministic Approximation of Diffusion Equations Using Particles
• On the Rate of Convergence of Empirical Measures in ∞-transportation Distance
• Convergence of the point vortex method for the 2-D euler equations
• The Weighted Particle Method for Convection-Diffusion Equations. Part 1: The Case of an Isotropic Viscosity
• The Variational Formulation of the Fokker--Planck Equation
• Nonasymptotic mixing of the MALA algorithm
• One-dimensional empirical measures, order statistics, and Kantorovich transport distances
• Local and Global Existence for an Aggregation Equation
• Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
• Particle methods for dispersive equations

This page was built for publication: Scaling Limit of the Stein Variational Gradient Descent: The Mean Field Regime