Optimal neural network approximation of Wasserstein gradient direction via convex optimization

DOI10.1137/23M1573173MaRDI QIDQ6633046zbMATH OpenFDO

Authors Yifei Wang, Peng Chen, Mert Pilanci, Wuchen Li

Publication date 5 November 2024

Published in SIAM Journal on Mathematics of Data Science (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1137/23M1573173

zbMATH Keywords

Bayesian inferenceconvex optimizationneural networksemipositive definite program

Mathematics Subject Classification ID

Approximation by other special function classes (41A30) Bayesian inference (62F15) Numerical optimization and variational techniques (65K10)

Cites work

• A Decomposition Augmented Lagrangian Method for Low-Rank Semidefinite Programming
• A fully discrete variational scheme for solving nonlinear Fokker-Planck equations in multiple space dimensions
• A Newton-CG augmented Lagrangian method for semidefinite programming
• Certified dimension reduction in nonlinear Bayesian inverse problems
• CVXPY: a Python-embedded modeling language for convex optimization
• Dimension-independent likelihood-informed MCMC
• Enumeration of Seven-Argument Threshold Functions
• Estimation of non-normalized statistical models by score matching
• scientific article; zbMATH DE number 2152346 (Why is no real title available?)
• scientific article; zbMATH DE number 1909499 (Why is no real title available?)
• scientific article; zbMATH DE number 2107836 (Why is no real title available?)
• Inverse problems: a Bayesian perspective
• Lagrangian schemes for Wasserstein gradient flows
• Multilayer feedforward networks are universal approximators
• Newton Sketch: A Near Linear-Time Optimization Algorithm with Linear-Quadratic Convergence
• Primal dual methods for Wasserstein gradient flows
• Projected Wasserstein Gradient Descent for High-Dimensional Bayesian Inference
• Scalable semidefinite programming
• Scaling Up Bayesian Uncertainty Quantification for Inverse Problems Using Deep Neural Networks
• SDPNAL+: a majorized semismooth Newton-CG augmented Lagrangian method for semidefinite programming with nonnegative constraints
• The deep minimizing movement scheme
• THE GEOMETRY OF DISSIPATIVE EVOLUTION EQUATIONS: THE POROUS MEDIUM EQUATION
• The Variational Formulation of the Fokker--Planck Equation
• Trust-region problems with linear inequality constraints: exact SDP relaxation, global optimality and robust optimization
• Wasserstein proximal of GANs

Cited in

(2)

• A score-based particle method for homogeneous Landau equation
• A library of mirrors: deep neural nets in low dimensions are convex Lasso models with reflection features

This page was built for publication: Optimal neural network approximation of Wasserstein gradient direction via convex optimization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6633046)