AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Private Adaptive Gradient Methods for Convex Optimization

Hilal Asi, John Duchi, Alireza Fallah, Omid Javidbakht, Kunal Talwar

ICML 2021 · Optimization › Nonconvex and smooth optimization · approach: Stochastic gradient last-iterate analysis · 14 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use PAGAN instead of DP-SGDLSTM · WikiText-2 test, ε=3perplexity: 238.44 -> 224.82Improved
Use PAGAN instead of DP-SGDLSTM · WikiText-2 test, ε=1perplexity: 285.11 -> 253.41Improved
Use PAGAN instead of DP-SGDLSTM · WikiText-2 test, ε=.5perplexity: 350.23 -> 291.41Improved
Use PASAN instead of DP-SGDLSTM · WikiText-2 test, ε=3perplexity: 238.44 -> 238.87No clear effect
Use PASAN instead of DP-SGDLSTM · WikiText-2 test, ε=1perplexity: 285.11 -> 274.63Improved
Use PASAN instead of DP-SGDLSTM · WikiText-2 test, ε=.5perplexity: 350.23 -> 332.52Improved
Adapt the private noise and gradient update to gradient geometryPAGAN · Synthetic absolute regressionsample lossImproved
Use isotropic noise with standard private AdaGrad instead of geometry-adaptive noisePrivate AdaGrad · Synthetic absolute regression, ε=4convergence speedHurt
Use PAGAN instead of non-adaptive private methodsLSTM · WikiText-2perplexityImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy