AI Paper Atlas

Research map / Optimization / Optimizers

Reducing the variance in online optimization by transporting past gradients

Sébastien Arnold, Pierre-Antoine Manzagol, Reza Babanezhad Harikandeh, Ioannis Mitliagkas, Nicolas Le Roux

NeurIPS 2019 · Optimization › Optimizers · approach: Stochastic gradient MCMC · 15 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
add implicit gradient transport with anytime tail averaging to Heavy BallHeavy Ball on MobileNetV2 · CIFAR10 testaccuracy: 91.55 -> 90.37Hurt
add implicit gradient transport with anytime tail averaging to Heavy BallHeavy Ball on LeNet5 · MNIST testaccuracy: 99.08 -> 99.19Improved
add implicit gradient transport with anytime tail averaging to Heavy BallHeavy Ball on logistic regression · Linear-MNIST testaccuracy: 92.52 -> 92.50No clear effect
add implicit gradient transport with anytime tail averaging to Heavy BallHeavy Ball on IMDb · IMDb testaccuracy: 86.90 -> 87.26No clear effect
add IGT gradients to AdamAdam on MAML · Mini-ImageNet testaccuracy: 63.70% -> 65.16%Improved
add IGT gradients with anytime tail averaging to Heavy BallHeavy Ball on MAML · Mini-ImageNet testaccuracy: 63.08% -> 64.57%Improved
remove anytime tail averaging from Heavyball-ITA (ablation)Heavyball-ITA on MobileNetV2 · CIFAR10 testaccuracy: 90.37 -> 12.05Hurt
remove anytime tail averaging from Heavyball-ITA (ablation)Heavyball-ITA on CIFAR10 · CIFAR10 trainingtraining errorHurt
use increasing momentum without gradient transport (ablation)Momentum with anytime tail averaging · CIFAR10training lossHurt
increase the minibatch size (ablation)IGT · CIFAR10 trainingsteps to target training lossMixed
use anytime tail averaging with IGT (ablation)Heavyball-IGT on ConvNet settings · MNIST and CIFAR10 optimizationtraining errorImproved

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