Stochastic AC-FGM achieves optimal O(1/√ε) iteration complexity and O(1/ε²) sample complexity while being fully adaptive to smoothness, horizon, and noise under bounded conditional variance.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
math.OC 3years
2026 3verdicts
UNVERDICTED 3representative citing papers
PFUGS is the first parameter-free gradient sliding method for composite convex problems with unknown Hölder and Lipschitz constants, using O((M_ν/ε)^{2/(1+3ν)}) subgradient evaluations of f and O((L/ε)^{1/2}) gradient evaluations of g.
The paper proposes an auto-conditioned framework for Frank-Wolfe algorithms that replaces global smoothness constants with local estimators computed from first-order information, achieving convergence to stationary points in nonconvex settings and sublinear rates in convex settings without prior kno
citing papers explorer
-
Stochastic Auto-conditioned Fast Gradient Methods with Optimal Rates
Stochastic AC-FGM achieves optimal O(1/√ε) iteration complexity and O(1/ε²) sample complexity while being fully adaptive to smoothness, horizon, and noise under bounded conditional variance.
-
Universal and Parameter-free Gradient Sliding for Composite Optimization
PFUGS is the first parameter-free gradient sliding method for composite convex problems with unknown Hölder and Lipschitz constants, using O((M_ν/ε)^{2/(1+3ν)}) subgradient evaluations of f and O((L/ε)^{1/2}) gradient evaluations of g.
-
Auto-Conditioned Frank-Wolfe Algorithms
The paper proposes an auto-conditioned framework for Frank-Wolfe algorithms that replaces global smoothness constants with local estimators computed from first-order information, achieving convergence to stationary points in nonconvex settings and sublinear rates in convex settings without prior kno