Switched diffusion processes for non-convex optimization and saddle points search
From MaRDI portal
Publication:6089196
DOI10.1007/S11222-023-10315-2zbMATH Open1523.62016arXiv2303.13160OpenAlexW4387666788MaRDI QIDQ6089196FDOQ6089196
Authors: Lucas Journel, Pierre Monmarché 
Publication date: 17 November 2023
Published in: Statistics and Computing (Search for Journal in Brave)
Abstract: We introduce and investigate stochastic processes designed to find local minimizers and saddle points of non-convex functions, exploring the landscape more efficiently than the standard noisy gradient descent. The processes switch between two behaviours, a noisy gradient descent and a noisy saddle point search. It is proven to be well-defined and to converge to a stationary distribution in the long time. Numerical experiments are provided on low-dimensional toy models and for Lennard-Jones clusters.
Full work available at URL: https://arxiv.org/abs/2303.13160
Recommendations
- A Newton-based method for nonconvex optimization with fast evasion of saddle points
- Stochastic optimization with momentum: convergence, fluctuations, and traps avoidance
- A Diffusion Approximation Theory of Momentum Stochastic Gradient Descent in Nonconvex Optimization
- On the Global Convergence of Randomized Coordinate Gradient Descent for Nonconvex Optimization
- Behavior of accelerated gradient methods near critical points of nonconvex functions
metastabilitystochastic optimisationMarkov-modulated processsaddle point searchswitched diffusion process
Cites Work
- Averaging principle of SDE with small diffusion: Moderate deviations
- Multidimensional diffusion processes.
- Random Perturbations of Dynamical Systems
- Large deviations for Markov-modulated diffusion processes with rapid switching
- Minimum action method for the study of rare events
- Tail of a linear diffusion with Markov switching
- Long time behavior of diffusions with Markov switching
- Yet Another Look at Harris’ Ergodic Theorem for Markov Chains
- Free energy computations. A mathematical perspective
- Partial differential equations and stochastic methods in molecular dynamics
- Fokker-Planck-Kolmogorov equations
- Option pricing model based on a Markov-modulated diffusion with jumps
- On unique ergodicity for degenerate diffusions
- Asymptotics of the spectral gap with applications to the theory of simulated annealing
- The gentlest ascent dynamics
- An iterative minimization formulation for saddle point search
- Convergence of the kinetic annealing for general potentials
- Controllability implies mixing. I: Convergence in the total variation metric
- Dynamical systems under random perturbations with fast switching and slow diffusion: hyperbolic equilibria and stable limit cycles
- Simplified gentlest ascent dynamics for saddle points in non-gradient systems
- Convergence and cycling in Walker-type saddle search algorithms
- Exact long time behavior of some regime switching stochastic processes
- Some remarks on the model of the extended gentlest ascent dynamics
- Convergence analysis of discrete high-index saddle dynamics
- Constrained overdamped Langevin dynamics for symmetric multimarginal optimal transportation
This page was built for publication: Switched diffusion processes for non-convex optimization and saddle points search
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6089196)
