Pith. sign in

REVIEW 1 cited by

Quadratic minimization: from conjugate gradient to an adaptive Heavy-ball method with Polyak step-sizes

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2210.06367 v1 pith:MD7YBYKD submitted 2022-10-12 math.OC

classification math.OC
keywords methodclassicalgradientpolyakproblemquadraticstep-sizesadaptive
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

In this work, we propose an adaptive variation on the classical Heavy-ball method for convex quadratic minimization. The adaptivity crucially relies on so-called "Polyak step-sizes", which consists in using the knowledge of the optimal value of the optimization problem at hand instead of problem parameters such as a few eigenvalues of the Hessian of the problem. This method happens to also be equivalent to a variation of the classical conjugate gradient method, and thereby inherits many of its attractive features, including its finite-time convergence, instance optimality, and its worst-case convergence rates. The classical gradient method with Polyak step-sizes is known to behave very well in situations in which it can be used, and the question of whether incorporating momentum in this method is possible and can improve the method itself appeared to be open. We provide a definitive answer to this question for minimizing convex quadratic functions, a arguably necessary first step for developing such methods in more general setups.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. On the construction of a gradient method of quadratic optimization, optimal from the point of view of minimizing the distance to the exact solution

    math.OC 2025-06 conditional novelty 4.0 of 10

    An m-moment minimum error method is constructed for quadratic optimization in Hilbert space, with proved convergence, optimality among Krylov methods, and numerical tests on Helmholtz, heat, and thermoacoustics invers...

Pith tools