The paper establishes that the optimal excess risk for ε-unlearning is the usual statistical error plus an unlearning penalty that interpolates between retraining-from-scratch and an exponentially smaller term as ε/d grows, with matching bounds for mean estimation.
Meaningful lower-bound of√ a2 +b −a when a≫b >0
8 Pith papers cite this work, alongside 55 external citations. Polarity classification is still indexing.
8
Pith papers citing it
55
external citations · external index
citation-role summary
background 1
citation-polarity summary
years
2026 8roles
background 1polarities
background 1representative citing papers
High-probability generalization bounds for D-SGD are derived at the optimal rate O(1/sqrt(mn) log(1/δ)) via pointwise uniform stability across convex and non-convex settings.
Price's gradient estimator enables black-box VI to achieve the same state-of-the-art iteration complexity as Wasserstein VI, with experiments confirming it as the main performance driver.
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.
Low-resolution data improves high-resolution model performance when high-resolution samples are limited, via KL-divergence bounds and experiments on vision transformers and CNNs.
Decentralized SGD and SGDA under Markovian sampling admit non-asymptotic generalization bounds that incorporate network topology, Markov mixing rates, and primal-dual dynamics.
AdaNAGED combines zeroth-order gradient-free training, automatic parameter adaptation, and LMO-based non-Euclidean geometry with claimed convergence guarantees, demonstrated on OPT-1.3B fine-tuning.
A survey equating offline Monte-Carlo/SAA and online stochastic-approximation sample complexities for convex stochastic optimization arising in statistics and ML.
citing papers explorer
-
Near-Optimal Pure Machine Unlearning for Smooth Strongly Convex Losses
The paper establishes that the optimal excess risk for ε-unlearning is the usual statistical error plus an unlearning penalty that interpolates between retraining-from-scratch and an exponentially smaller term as ε/d grows, with matching bounds for mean estimation.
-
Unveiling High-Probability Generalization in Decentralized SGD
High-probability generalization bounds for D-SGD are derived at the optimal rate O(1/sqrt(mn) log(1/δ)) via pointwise uniform stability across convex and non-convex settings.
-
Stochastic Gradient Variational Inference with Price's Gradient Estimator from Bures-Wasserstein to Parameter Space
Price's gradient estimator enables black-box VI to achieve the same state-of-the-art iteration complexity as Wasserstein VI, with experiments confirming it as the main performance driver.
-
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.
-
On What We Can Learn from Low-Resolution Data
Low-resolution data improves high-resolution model performance when high-resolution samples are limited, via KL-divergence bounds and experiments on vision transformers and CNNs.
-
Stability and Generalization for Decentralized Markov SGD
Decentralized SGD and SGDA under Markovian sampling admit non-asymptotic generalization bounds that incorporate network topology, Markov mixing rates, and primal-dual dynamics.
-
Zero-order Parameter-free Optimization for LMO-based Methods: Novel Approach for Efficient Fine-tuning
AdaNAGED combines zeroth-order gradient-free training, automatic parameter adaptation, and LMO-based non-Euclidean geometry with claimed convergence guarantees, demonstrated on OPT-1.3B fine-tuning.
-
Stochastic Optimization and Data Science
A survey equating offline Monte-Carlo/SAA and online stochastic-approximation sample complexities for convex stochastic optimization arising in statistics and ML.