Non-quadratic Mirror Descent exhibits exponential initialization sensitivity in convex settings, shown via 3D constructions and KL-regularized simplex examples, with Bregman anchoring proposed for stabilization.
Empirical risk minimization with shuffled sgd: A primal-dual perspective and improved bounds
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
citation-role summary
method 1
citation-polarity summary
years
2026 5roles
method 1polarities
use method 1representative citing papers
Dual Averaging Power-Prox method provides the first convergence analysis for incremental gradients with heavy-tailed noise and shows asymptotically better rates than i.i.d. SGD.
Combining random reshuffling and Richardson-Romberg extrapolation yields cubic bias refinement and better MSE for constant-step SGD on structured non-monotone variational inequalities.
SGD on multiclass cross-entropy loss alternates between curvature-driven oscillations and stable regimes but self-stabilizes to enable best-iterate convergence with large learning rates for linear and two-layer models.
Continual classification in homogeneous models is sequential projections onto margin sets, with local linear convergence under regularity properties for random and cyclic tasks, extended to regression.
citing papers explorer
-
Mirror Descent Beyond Euclidean Stability: An Exponential Separation in Initialization Sensitivity
Non-quadratic Mirror Descent exhibits exponential initialization sensitivity in convex settings, shown via 3D constructions and KL-regularized simplex examples, with Bregman anchoring proposed for stabilization.
-
The Dual Averaging Power-Prox Method with Application to Heavy-Tail Incremental Gradient
Dual Averaging Power-Prox method provides the first convergence analysis for incremental gradients with heavy-tailed noise and shows asymptotically better rates than i.i.d. SGD.
-
Shuffling the Data, Stretching the Step-size: Sharper Bias in constant step-size SGD
Combining random reshuffling and Richardson-Romberg extrapolation yields cubic bias refinement and better MSE for constant-step SGD on structured non-monotone variational inequalities.
-
SGD at the Edge of Stability: Stochastic Stabilization with Large Learning Rates
SGD on multiclass cross-entropy loss alternates between curvature-driven oscillations and stable regimes but self-stabilizes to enable best-iterate convergence with large learning rates for linear and two-layer models.
-
Convergence of Continual Learning in Homogeneous Deep Networks
Continual classification in homogeneous models is sequential projections onto margin sets, with local linear convergence under regularity properties for random and cyclic tasks, extended to regression.