Nesterov acceleration in optimizing over probability measures

DOI10.48550/arXiv.2607.23008MaRDI QIDQ7284439arXivFDO

Authors Wilfrid Gangbo, Jiaqi Tang, Qin Li

Publication date 25 July 2026

Published in arXiv

Tldr We develop Nesterov acceleration over probability measure spaces with non-asymptotic convergence guarantees, extending the classic method from Euclidean space to probability measures while overcoming computational challenges via two complementary lifting procedures.

Full work available at URL https://doi.org/10.48550/arXiv.2607.23008

arXiv classification

cs.LGmath.OC

Cites work

• A blob method for diffusion
• A differential equation for modeling Nesterov's accelerated gradient method: theory and insights
• A mean field view of the landscape of two-layer neural networks
• A variational perspective on accelerated methods in optimization
• Accelerating optimization over the space of probability measures
• Analysis and design of optimization algorithms via integral quadratic constraints
• Birth–death dynamics for sampling: global convergence, approximations and their asymptotics
• Building population-informed priors for Bayesian inference using data-consistent stochastic inversion
• Diffusion Schrödinger bridges for Bayesian computation
• From Optimization to Sampling Through Gradient Flows
• Global well-posedness of Master equations for deterministic displacement convex potential mean field games
• Mean field analysis of neural networks: a law of large numbers
• Mean field games master equations with nonseparable Hamiltonians and displacement monotonicity
• Mean-Field Langevin Dynamics: Exponential Convergence and Annealing
• Non‐linear stability for the Vlasov–Poisson system—the energy‐Casimir method
• On differentiability in the Wasserstein space and well-posedness for Hamilton-Jacobi equations
• On the Global Convergence of Gradient Descent for Over-parameterized Models using Optimal Transport
• On the rate of convergence in Wasserstein distance of the empirical measure
• Overparameterization of deep ResNet: zero loss and mean-field analysis
• Projected Wasserstein Gradient Descent for High-Dimensional Bayesian Inference
• Reduction and a concentration-compactness principle for energy-Casimir functionals
• Sampling as optimization in the space of measures: The Langevin dynamics as a composite optimization problem
• Some methods of speeding up the convergence of iteration methods
• The first Fundamental Theorem of Calculus for functions defined on Wasserstein space
• The Master Equation and the Convergence Problem in Mean Field Games
• The Variational Formulation of the Fokker--Planck Equation
• Understanding the acceleration phenomenon via high-resolution differential equations
• Variational inference via Wasserstein gradient flows

This page was built for publication: Nesterov acceleration in optimizing over probability measures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284439)