REVIEW 1 cited by
Non-ergodic Convergence Analysis of Heavy-Ball Algorithms
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
Non-ergodic Convergence Analysis of Heavy-Ball Algorithms
read the original abstract
In this paper, we revisit the convergence of the Heavy-ball method, and present improved convergence complexity results in the convex setting. We provide the first non-ergodic O(1/k) rate result of the Heavy-ball algorithm with constant step size for coercive objective functions. For objective functions satisfying a relaxed strongly convex condition, the linear convergence is established under weaker assumptions on the step size and inertial parameter than made in the existing literature. We extend our results to multi-block version of the algorithm with both the cyclic and stochastic update rules. In addition, our results can also be extended to decentralized optimization, where the ergodic analysis is not applicable.
Forward citations
Cited by 1 Pith paper
-
Unified convergence analysis for gradient descent optimization methods in the training of deep neural networks
Bounded trajectories of a broad class of GD optimizers (Adam, RMSprop, NAG, Adan, etc.) converge with polynomial rates to critical points of KL objectives with locally Lipschitz gradients, covering analytic-activation...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.