Near-optimal hyperfast second-order method for convex optimization

DOI10.1007/978-3-030-58657-7_15zbMATH Open1460.90133arXiv2002.09050OpenAlexW3084935096MaRDI QIDQ4965110FDOQ4965110

Authors: Dmitry Kamzolov 

Publication date: 25 February 2021

Published in: Mathematical Optimization Theory and Operations Research (Search for Journal in Brave)

Abstract: In this paper, we present a new Hyperfast Second-Order Method with convergence rate $O(N^{-5})$ up to a logarithmic factor for the convex function with Lipshitz the third derivative. This method based on two ideas. The first comes from the superfast second-order scheme of Yu. Nesterov (CORE Discussion Paper 2020/07, 2020). It allows implementing the third-order scheme by solving subproblem using only the second-order oracle. This method converges with rate $O(N^{-4})$. The second idea comes from the work of Kamzolov et al. (arXiv:2002.01004). It is the inexact near-optimal third-order method. In this work, we improve its convergence and merge it with the scheme of solving subproblem using only the second-order oracle. As a result, we get convergence rate $O(N^{-5})$ up to a logarithmic factor. This convergence rate is near-optimal and the best known up to this moment. Further, we investigate the situation when there is a sum of two functions and improve the sliding framework from Kamzolov et al. (arXiv:2002.01004) for the second-order methods.

Full work available at URL: https://arxiv.org/abs/2002.09050

Recommendations

• Superfast second-order methods for unconstrained convex optimization
• Reachability of optimal convergence rate estimates for high-order numerical convex optimization methods
• Implementable tensor methods in unconstrained convex optimization
• Fast convex optimization via a third-order in time evolution equation
• An optimal high-order tensor method for convex optimization

zbMATH Keywords

complexitytensor methodinexact methodsecond-order method

Mathematics Subject Classification ID

Convex programming (90C25)

Cites Work

• An accelerated hybrid proximal extragradient method for convex optimization and its implications to second-order methods
• Cubic regularization of Newton method and its global performance
• Worst-case evaluation complexity for unconstrained nonlinear optimization using high-order regularized models
• Lectures on convex optimization
• Tensor methods for minimizing convex functions with Hölder continuous higher-order derivatives
• Lower bounds for finding stationary points I
• Accelerated regularized Newton methods for minimizing composite convex functions
• Contracting proximal methods for smooth convex optimization

Cited In (10)

• Super-Universal Regularized Newton Method
• Machine Learning: ECML 2004
• Local convergence of tensor methods
• Accelerated first-order methods for large-scale convex optimization: nearly optimal complexity under strong convexity
• Fast Curvature Matrix-Vector Products for Second-Order Gradient Descent
• Superfast second-order methods for unconstrained convex optimization
• A control-theoretic perspective on optimal high-order optimization
• Fast convex optimization via a third-order in time evolution equation
• Hyperfast second-order local solvers for efficient statistically preconditioned distributed optimization
• Accelerations for global optimization covering methods using second derivatives

This page was built for publication: Near-optimal hyperfast second-order method for convex optimization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4965110)