Automatic clip**: Differentially private deep learning made easier and stronger

Z Bu, YX Wang, S Zha… - Advances in Neural …, 2024 - proceedings.neurips.cc
Per-example gradient clip** is a key algorithmic step that enables practical differential
private (DP) training for deep learning models. The choice of clip** threshold $ R …

Convergence of adagrad for non-convex objectives: Simple proofs and relaxed assumptions

B Wang, H Zhang, Z Ma… - The Thirty Sixth Annual …, 2023 - proceedings.mlr.press
We provide a simple convergence proof for AdaGrad optimizing non-convex objectives
under only affine noise variance and bounded smoothness assumptions. The proof is …

Generalized-smooth nonconvex optimization is as efficient as smooth nonconvex optimization

Z Chen, Y Zhou, Y Liang, Z Lu - International Conference on …, 2023 - proceedings.mlr.press
Various optimal gradient-based algorithms have been developed for smooth nonconvex
optimization. However, many nonconvex machine learning problems do not belong to the …

DPSUR: accelerating differentially private stochastic gradient descent using selective update and release

J Fu, Q Ye, H Hu, Z Chen, L Wang, K Wang… - ar** for non-convex optimization
A Reisizadeh, H Li, S Das, A Jadbabaie - ar** is a standard training technique used in deep learning applications such
as large-scale language modeling to mitigate exploding gradients. Recent experimental …

A theory to instruct differentially-private learning via clip** bias reduction

H ** and communication compression
B Li, Y Chi - IEEE Journal of Selected Topics in Signal …, 2025 - ieeexplore.ieee.org
Achieving communication efficiency in decentralized machine learning has been attracting
significant attention, with communication compression recognized as an effective technique …