NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
31 results for “keyword:"sgd"” in ICLR 2027. 7 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Same Privacy Guarantee, Different Robustness: How Gradient Clipping Shapes Adversarial Training▲3.0k2.33
[PDF]A Solvable Model of Adaptive Learning Rate Rescaling: Acceleration, Stability & Scaling▼5.8k2.33
[PDF]Optimization Risk Bounds for Kolmogorov–Arnold Networks Trained by DP-SGD with Correlated Noise▲5103.00
[PDF]From Pixels to Privacy-Sensitive Features: An Information-Theoretic Framework for Reconstruction Bounds▲2623.67
[PDF]Zeroth-Order SGD with Momentum Resetting Achieves Dimension-Dependent Optimality in Nonsmooth Nonconvex Optimization▼4.5k2.67
[PDF]Avoiding Bias in Clipped SGD for Overparameterized Models under Generalized Smoothness▲3.0k2.33
[PDF]Memory-Efficient Subspace Optimization with Residual Correction: From Momentum SGD to Muon▲5103.00
Withdrawn or desk rejected (7)
[PDF]Attested but Unaccounted: DP-SGD Privacy Accounting as a Precondition of Every StepWithdrawn1.67
[PDF]Stability and Generalization of Noisy signSGD: From Sign Discontinuity to Gaussian SmoothingWithdrawn
[PDF]WHEN BATCH SCHEDULES CANNOT BE CERTIFIED: CAPACITY OF GLOBALLY VALID MONOMIAL CERTIFICATES ON INTEGER LOG-HULLSWithdrawn1.33
[PDF]EquiConfig: Efficient Mitigation of Disparate Impact in Differentially Private LearningWithdrawn2.00