AI Paper Atlas

Research map / Optimization / Optimizers

A Momentumized, Adaptive, Dual Averaged Gradient Method

Aaron Defazio, Samy Jelassi

JMLR 2022 · Optimization › Optimizers · approach: Adam convergence and adaptive learning rates

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 stage-wise learning rate schedule instead of sqrt-decay a/sqrt(i+b) for SGD (ablation)SGD without momentum · CIFAR-10test accuracy (%): 84.25 -> 94.15Improved
increase offset b in sqrt-decay schedule from 1 to 10000 (ablation)SGD without momentum · CIFAR-10test accuracy (%): 84.25 -> 93.55Improved
add adaptivity (AdaGrad denominator) to dual averaging under stage-wise schedule (ablation)Dual averaging (DA) · CIFAR-10test accuracy (%): 84.48 -> 93.66Improved
use stage-wise instead of flat learning rate for AdaGrad-DA (ablation)AdaGrad-DA · CIFAR-10test accuracy (%): 91.85 -> 93.66Improved
use stage-wise instead of flat learning rate for plain dual averaging (ablation)Dual averaging (DA) · CIFAR-10test accuracy (%): 87.03 -> 84.48Hurt
use dual averaging AdaGrad instead of mirror descent AdaGrad (stage-wise) (ablation)AdaGrad · CIFAR-10test accuracy (%): 92.05 -> 93.66Improved
MADGRAD instead of AdamPreResNet152 · CIFAR-10test accuracy (%): 94.69 -> 95.38Improved
MADGRAD instead of SGD+MResNet-50 · ILSVRC 2012 ImageNettest accuracy (%): 76.09 -> 76.22Improved
MADGRAD instead of AdamResNet-50 · ILSVRC 2012 ImageNettest accuracy (%): 73.03 -> 76.22Improved
MADGRAD instead of AdamVarNet 2.0 · fastMRI Kneetest SSIM: 0.9111 -> 0.9119Improved
MADGRAD instead of AdamLSTM · IWSLT14 De-Entest loss: 4.35 -> 4.33Improved
MADGRAD instead of AdamRoBERTa · BookWiki masked LMtest loss: 2.09 -> 2.07Improved

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