Pith. sign in

REVIEW 3 cited by

Enhancing Convergence of Decentralized Gradient Tracking under the KL Property

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 2412.09556 v1 pith:BNL4NNOL submitted 2024-12-12 math.OC cs.LGcs.SYeess.SYstat.ML

Enhancing Convergence of Decentralized Gradient Tracking under the KL Property

classification math.OC cs.LGcs.SYeess.SYstat.ML
keywords thetaconvergencewhendecentralizedfunctionpropertyratetextbf
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
abstract

We study decentralized multiagent optimization over networks, modeled as undirected graphs. The optimization problem consists of minimizing a nonconvex smooth function plus a convex extended-value function, which enforces constraints or extra structure on the solution (e.g., sparsity, low-rank). We further assume that the objective function satisfies the Kurdyka-{\L}ojasiewicz (KL) property, with given exponent $\theta\in [0,1)$. The KL property is satisfied by several (nonconvex) functions of practical interest, e.g., arising from machine learning applications; in the centralized setting, it permits to achieve strong convergence guarantees. Here we establish convergence of the same type for the notorious decentralized gradient-tracking-based algorithm SONATA. Specifically, $\textbf{(i)}$ when $\theta\in (0,1/2]$, the sequence generated by SONATA converges to a stationary solution of the problem at R-linear rate;$ \textbf{(ii)} $when $\theta\in (1/2,1)$, sublinear rate is certified; and finally $\textbf{(iii)}$ when $\theta=0$, the iterates will either converge in a finite number of steps or converges at R-linear rate. This matches the convergence behavior of centralized proximal-gradient algorithms except when $\theta=0$. Numerical results validate our theoretical findings.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

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

  1. A Unified Algorithm for Nonconvex Decentralized Nonlinear Optimization

    math.OC 2025-11 conditional novelty 7.0

    A unified framework for decentralized nonconvex optimization that encompasses gradient tracking and quasi-Newton methods, with convergence analysis under nonconvex and Kurdyka-Lojasiewicz conditions and new efficient ...

  2. A Unified Fractional Regularization Framework for Sparse Recovery

    cs.IT 2026-04 unverdicted novelty 5.0

    A unified ℓ1/ℓp^q fractional regularizer equates to the subtractive ℓ1 - α ℓp model at stationary points, supplies a new RIP-based recovery condition, and is solved by a provably convergent majorization-minimization a...

  3. A Unified Fractional Regularization Framework for Sparse Recovery

    cs.IT 2026-04 unverdicted novelty 4.0

    A unified ℓ₁/ℓ_p^q fractional regularization framework for sparse recovery is shown to be equivalent to subtractive ℓ₁−αℓ_p models at stationary points, with RIP-based recovery guarantees and a convergent MM algorithm.