Global Optimization via Schrödinger–Föllmer Diffusion

DOI10.1137/22M1477738zbMATH Open1522.90120arXiv2111.00402OpenAlexW3209115427MaRDI QIDQ6057791FDOQ6057791

Authors: Yin Dai, Yu Ling Jiao, Lican Kang, Xiliang Lu, Jerry Zhijian Yang 

Publication date: 26 October 2023

Published in: SIAM Journal on Control and Optimization (Search for Journal in Brave)

Abstract: We study the problem of finding global minimizers of $V(x):mathbb{R}^{d}ightarrowmathbbR$ approximately via sampling from a probability distribution $m{u}_{sigma}$ with density $p_{sigma}(x)=dfracexp(-V(x)/sigma)in{t}_{mathbb{R}^d}exp(-V(y)/sigma)dy$ with respect to the Lebesgue measure for $sigmain(0,1]$ small enough. We analyze a sampler based on the Euler-Maruyama discretization of the Schr{"o}dinger-F{"o}llmer diffusion processes with stochastic approximation under appropriate assumptions on the step size $s$ and the potential $V$. We prove that the output of the proposed sampler is an approximate global minimizer of $V(x)$ with high probability at cost of sampling $mathcalO(d^3)$ standard normal random variables. Numerical studies illustrate the effectiveness of the proposed method and its superiority to the Langevin method.

Full work available at URL: https://arxiv.org/abs/2111.00402

Recommendations

• Diffusions for Global Optimization
• Diffusion for Global Optimization in $\mathbb{R}^n $
• Convergence analysis of a global optimization algorithm using stochastic differential equations
• Global optimization and stochastic differential equations
• Publication:4509896

zbMATH Keywords

global optimizationstochastic approximationSchrödinger-Föllmer diffusion

Mathematics Subject Classification ID

Computational methods for problems pertaining to probability theory (60-08) Nonconvex programming, global optimization (90C26)

Cites Work

• A Relationship Between Arbitrary Positive Matrices and Doubly Stochastic Matrices
• Computational optimal transport. With applications to data sciences
• Smooth minimization of non-smooth functions
• Title not available (Why is that?)
• Rate of convergence for ergodic continuous Markov processes: Lyapunov versus Poincaré
• Title not available (Why is that?)
• Title not available (Why is that?)
• Representation formula for the entropy and functional inequalities
• Title not available (Why is that?)
• Analysis and geometry of Markov diffusion operators
• Log-Sobolev inequalities: different roles of Ric and Hess
• Title not available (Why is that?)
• A survey of the Schrödinger problem and some of its connections with optimal transport
• Title not available (Why is that?)
• The Markov processes of Schr�dinger
• Title not available (Why is that?)
• A stochastic control approach to reciprocal diffusion processes
• Stochastic processes and applications. Diffusion processes, the Fokker-Planck and Langevin equations
• Asymptotic probabilities and differential equations
• Title not available (Why is that?)
• Stein's lemma for elliptical random vectors
• Laplace's method revisited: Weak convergence of probability measures
• Asymptotics of the spectral gap with applications to the theory of simulated annealing
• High-dimensional Bayesian inference via the unadjusted Langevin algorithm
• Poincaré and logarithmic Sobolev inequalities by decomposition of the energy landscape
• Diffusion for Global Optimization in $\mathbb{R}^n $
• Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
• Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
• Convergence rates for annealing diffusion processes
• Unbiased estimation using a class of diffusion processes
• Coupling and convergence for Hamiltonian Monte Carlo
• Simulation and Inference for Stochastic Processes with YUIMA
• Sampling can be faster than optimization
• On stochastic gradient Langevin dynamics with dependent data streams in the logconcave case
• Interacting Langevin diffusions: gradient structure and ensemble Kalman sampler
• Stability of the logarithmic Sobolev inequality via the Föllmer process
• User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient
• Convergence of Langevin MCMC in KL-divergence
• Accelerated information gradient flow
• A consensus-based global optimization method for high dimensional machine learning problems
• Stochastic control liaisons. Richard Sinkhorn meets Gaspard Monge on a Schrödinger bridge
• On Stochastic Gradient Langevin Dynamics with Dependent Data Streams: The Fully Nonconvex Case
• Bayesian learning via neural Schrödinger-Föllmer flows
• Improved bounds for discretization of Langevin diffusions: near-optimal rates without convexity
• A Generalization of Laplaces's Method
• Nonasymptotic estimates for stochastic gradient Langevin dynamics under local conditions in nonconvex optimization

Cited In (2)

• Weak approximation of Schrödinger-Föllmer diffusion
• Global optimization using diffusion perturbations with large noise intensity

This page was built for publication: Global Optimization via Schrödinger–Föllmer Diffusion

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6057791)