Inertial Newton algorithms avoiding strict saddle points

DOI10.1007/s10957-023-02330-0arXiv2111.04596MaRDI QIDQ6145046

Camille Castera

Publication date: 8 January 2024

Published in: Journal of Optimization Theory and Applications (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2111.04596

zbMATH Keywords

convergence analysisdynamical systemsnon-convex optimizationalgorithms for machine learning

Mathematics Subject Classification ID

Nonconvex programming, global optimization (90C26) Methods of quasi-Newton type (90C53)

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• An inertial forward-backward algorithm for the minimization of the sum of two nonconvex functions
• Fast convex optimization via inertial dynamics with Hessian driven damping
• The gradient and heavy ball with friction dynamical systems: The quasiconvex case
• A second-order gradient-like dissipative dynamical system with Hessian-driven damping. Application to optimization and mechanics.
• Local convergence of the heavy-ball method and iPiano for non-convex optimization
• Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward-backward splitting, and regularized Gauss-Seidel methods
• Continuous Newton-like inertial dynamics for monotone inclusions
• Understanding the acceleration phenomenon via high-resolution differential equations
• First-order optimization algorithms via inertial systems with Hessian driven damping
• A gradient-type algorithm with backward inertial steps associated to a nonconvex minimization problem
• Tikhonov regularization of a second order dynamical system with Hessian driven damping
• An extension of the second order dynamical system that models Nesterov's convex gradient method
• Backtracking gradient descent method and some applications in large scale optimisation. II: Algorithms and experiments
• Convergence rates for an inertial algorithm of gradient type associated to a smooth non-convex minimization
• A generalization of Hartman's linearization theorem
• Fast convergence of inertial dynamics and algorithms with asymptotic vanishing viscosity
• Behavior of accelerated gradient methods near critical points of nonconvex functions
• The stable, center-stable, center, center-unstable, unstable manifolds
• A New Value Iteration method for the Average Cost Dynamic Programming Problem
• The Differential Inclusion Modeling FISTA Algorithm and Optimality of Convergence Rate in the Case b $\leq3$
• Gradient Descent Only Converges to Minimizers: Non-Isolated Critical Points and Invariant Regions
• A Dynamical Approach to an Inertial Forward-Backward Algorithm for Convex Minimization
• Rate of convergence of the Nesterov accelerated gradient method in the subcritical case α ≤ 3
• Newton-like Inertial Dynamics and Proximal Algorithms Governed by Maximally Monotone Operators
• An Inertial Newton Algorithm for Deep Learning
• Optimal Convergence Rates for Nesterov Acceleration
• A Lemma in the Theory of Structural Stability of Differential Equations
• Some methods of speeding up the convergence of iteration methods
• Limit Points of Sequences in Metric Spaces
• Fast optimization via inertial dynamics with closed-loop damping

This page was built for publication: Inertial Newton algorithms avoiding strict saddle points