AI Paper Atlas

Research map / Graph learning / GNN architectures

Full-Graph vs. Mini-Batch Training: Comprehensive Analysis from a Batch Size and Fan-Out Size Perspective

Mengfan Liu, Da Zheng, Junwei Su, Chuan Wu

ICLR 2026 · Graph learning › GNN architectures · approach: GNN generalization theory · 2 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
tune batch size and fan-out for mini-batch instead of full-graphGraphSAGE · Redditbest test accuracy: 96.13 -> 96.32Improved
tune batch size and fan-out for mini-batch instead of full-graphGraphSAGE · ogbn-productsbest test accuracy: 77.92 -> 78.80Improved
tune batch size and fan-out for mini-batch instead of full-graphGraphSAGE · ogbn-arxivbest test accuracy: 70.96 -> 71.16Improved
tune batch size and fan-out for mini-batch instead of full-graphGraphSAGE · ogbn-papers100Mbest test accuracy: 59.54 -> 58.52Hurt
use iteration-to-accuracy instead of time-to-accuracy across hardwareconvergence metric · hardware comparisonmax variation across hardware: 2787.05% -> 41.28%Improved
increase batch size with fixed fan-out (MSE)one-layer GNN · convergence theoryiterations to convergence: T proportional to b^(5/2)Hurt
increase batch size with fixed fan-out (CE)one-layer GNN · convergence theoryiterations to convergence: T proportional to b^-1Improved
increase fan-out size under fixed batch sizeone-layer GNN · convergence theoryiterations to convergence: T proportional to beta^(-1/2) (MSE), beta^(-5/2) (CE)Improved
increase fan-out size to very large values (over about 15) under CE (ablation)GraphSAGE · ogbn-productstest accuracyHurt
increase batch size beyond half of training nodes under CE (ablation)GraphSAGE · ogbn-productstest accuracyHurt
use MSE instead of CE for large batch or fan-outGraphSAGE · ogbn-productstest accuracyImproved

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