A variational perspective on accelerated methods in optimization
From MaRDI portal
Abstract: Accelerated gradient methods play a central role in optimization, achieving optimal rates in many settings. While many generalizations and extensions of Nesterov's original acceleration method have been proposed, it is not yet clear what is the natural scope of the acceleration concept. In this paper, we study accelerated methods from a continuous-time perspective. We show that there is a Lagrangian functional that we call the emph{Bregman Lagrangian} which generates a large class of accelerated methods in continuous time, including (but not limited to) accelerated gradient descent, its non-Euclidean extension, and accelerated higher-order gradient methods. We show that the continuous-time limit of all of these methods correspond to traveling the same curve in spacetime at different speeds. From this perspective, Nesterov's technique and many of its generalizations can be viewed as a systematic way to go from the continuous-time curves generated by the Bregman Lagrangian to a family of discrete-time accelerated algorithms.
Recommendations
- Multiscale analysis of accelerated gradient methods
- A variational formulation of accelerated optimization on Riemannian manifolds
- Understanding the acceleration phenomenon via high-resolution differential equations
- A differential equation for modeling Nesterov's accelerated gradient method: theory and insights
- Generalized momentum-based methods: a Hamiltonian perspective
Cited in
(only showing first 100 items - show all)- Finding geodesics joining given points
- Convergence rates of inertial primal-dual dynamical methods for separable convex optimization problems
- Preconditioned accelerated gradient descent methods for locally Lipschitz smooth objectives with applications to the solution of nonlinear PDEs
- An LQR-Lagrange algorithm generated by interdisciplinary integration with optimal control and optimization
- Nesterov acceleration in optimizing over probability measures
- A differential equation for modeling Nesterov's accelerated gradient method: theory and insights
- No-regret dynamics in the Fenchel game: a unified framework for algorithmic convex optimization
- Discrete processes and their continuous limits
- Fast primal-dual algorithm via dynamical system for a linearly constrained convex optimization problem
- Online optimization of switched LTI systems using continuous-time and hybrid accelerated gradient flows
- New methods for parametric optimization via differential equations
- Online accelerated data-driven learning for optimal feedback control of discrete-time partially uncertain systems
- A fixed-time stable forward-backward dynamical system for solving generalized monotone inclusions
- Dynamical Systems Theory and Algorithms for NP-hard Problems
- An optimal high-order tensor method for convex optimization
- Triggered gradient tracking for asynchronous distributed optimization
- Inducing strong convergence of trajectories in dynamical systems associated to monotone inclusions with composite structure
- Explicit stabilised gradient descent for faster strongly convex optimisation
- Conformal mirror descent with logarithmic divergences
- Potential Function-Based Framework for Minimizing Gradients in Convex and Min-Max Optimization
- scientific article; zbMATH DE number 1268598 (Why is no real title available?)
- Linear convergence of forward-backward accelerated algorithms without knowledge of the modulus of strong convexity
- A fast primal-dual algorithm via dynamical system with variable mass for linearly constrained convex optimization
- On the connections between optimization algorithms, Lyapunov functions, and differential equations: theory and insights
- PDE acceleration: a convergence rate analysis and applications to obstacle problems
- Conformal symplectic and relativistic optimization
- Stochastic Methods for Composite and Weakly Convex Optimization Problems
- Mini-workshop: High-dimensional control problems and mean-field equations with applications in machine learning. Abstracts from the mini-workshop held December 8--13, 2024
- A general system of differential equations to model first-order adaptive algorithms
- Proximal gradient flow and Douglas-Rachford splitting dynamics: global exponential stability via integral quadratic constraints
- A new dynamical system with self-adaptive dynamical stepsize for pseudomonotone mixed variational inequalities
- Practical perspectives on symplectic accelerated optimization
- On dissipative symplectic integration with applications to gradient-based optimization
- First-order methods for convex optimization
- A primal-dual flow for affine constrained convex optimization
- Novel projection neurodynamic approaches for constrained convex optimization
- Implicit Regularization and Momentum Algorithms in Nonlinearly Parameterized Adaptive Control and Prediction
- Differential equations in data analysis
- Selection dynamics for deep neural networks
- A Recursively Recurrent Neural Network (R2N2) Architecture for Learning Iterative Algorithms
- Entropic mean-field min-max problems via best response flow
- A Systematic Approach to Lyapunov Analyses of Continuous-Time Models in Convex Optimization
- Accelerated variational PDEs for efficient solution of regularized inversion problems
- Multiscale analysis of accelerated gradient methods
- Accelerated Forward-Backward Optimization using Deep Learning
- Fast gradient method for low-rank matrix estimation
- A Continuous-Time Analysis of Distributed Stochastic Gradient
- The regularized submodular maximization via the Lyapunov method
- On the convergence of gradient-like flows with noisy gradient input
- Accelerated forward-backward and Douglas-Rachford splitting dynamics
- Continuous relaxations for the traveling salesman problem
- Accelerated optimization on Riemannian manifolds via discrete constrained variational integrators
- scientific article; zbMATH DE number 7370630 (Why is no real title available?)
- Accelerated Optimization in the PDE Framework: Formulations for the Manifold of Diffeomorphisms
- A continuous-time perspective on global acceleration for monotone equation problems
- Bregman dynamics, contact transformations and convex optimization
- Convergence rates of mixed primal-dual dynamical systems with Hessian driven damping
- Accelerated Optimization in the PDE Framework Formulations for the Active Contour Case
- Accelerated first-order primal-dual proximal methods for linearly constrained composite convex programming
- Quantum Langevin dynamics for optimization
- Search direction correction with normalized gradient makes first-order methods faster
- Accelerations for a variety of global optimization methods
- JKO schemes with general transport costs
- From Halpern's fixed-point iterations to Nesterov's accelerated interpretations for root-finding problems
- Stochastic mirror descent dynamics and their convergence in monotone variational inequalities
- An \(O(s^r)\)-resolution ODE framework for understanding discrete-time algorithms and applications to the linear convergence of minimax problems
- A gentle introduction to gradient-based optimization and variational inequalities for machine learning
- Nesterov's acceleration at the limit: first-order schemes
- On the mathematics of the natural physics of optimization
- An ordinary differential equation for modeling Halpern fixed-point Algorithm
- Acceleration via perturbations on low-resolution ordinary differential equations
- The approximate duality gap technique: a unified theory of first-order methods
- Neurodynamic optimization approaches with finite/fixed-time convergence for absolute value equations
- Exploring critical points of energy landscapes: from low-dimensional examples to phase field crystal PDEs
- scientific article; zbMATH DE number 7370590 (Why is no real title available?)
- Is there an analog of Nesterov acceleration for gradient-based MCMC?
- Mass-spring-damper networks for distributed optimization in non-Euclidean spaces
- A second order primal-dual dynamical system for a convex-concave bilinear saddle point problem
- Factor-\(\sqrt{2}\) acceleration of accelerated gradient methods
- Lagrangian and Hamiltonian dynamics for probabilities on the statistical bundle
- Fractional differential equation approach for convex optimization with convergence rate analysis
- Accelerated optimization algorithms and ordinary differential equations: the convex non Euclidean case
- Accelerated differential inclusion for convex optimization
- Analysis of optimization algorithms via integral quadratic constraints: nonstrongly convex problems
- scientific article; zbMATH DE number 7306890 (Why is no real title available?)
- Near-optimal closed-loop method via Lyapunov damping
- A continuous-time inertial Newton method for structured monotone inclusions
- User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient
- Hessian barrier algorithms for linearly constrained optimization problems
- Accelerated information gradient flow
- A control-theoretic perspective on optimal high-order optimization
- Last-iterate convergence of saddle-point optimizers via high-resolution differential equations
- Iterative ensemble Kalman methods: a unified perspective with some new variants
- Time-adaptive Lagrangian variational integrators for accelerated optimization
- Accelerated forward-backward algorithms with subgradient corrections
- Accelerated first-order optimization under nonlinear constraints
- Accelerated extra-gradient descent: a novel accelerated first-order method
- Stochastic dual dynamical systems for linear equality constrained convex optimization problems
- Fast primal-dual algorithm with Tikhonov regularization for a linear equality constrained convex optimization problem
- The application of Goeken-Johnson's Runge-Kutta methods in unconstrained convex optimization
This page was built for publication: A variational perspective on accelerated methods in optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4646231)
