AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

GOOD: A Graph Out-of-Distribution Benchmark

Shurui Gui, Xiner Li, Limei Wang, Shuiwang Ji

NeurIPS 2022 · Graph learning › Graph self-supervised learning · approach: Graph out-of-distribution generalization · 30 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
evaluate under OOD split instead of ID splitERM · GOOD-HIV size covariateROC-AUC (%): 83.72 -> 59.94Hurt
evaluate under OOD split instead of ID splitERM · GOOD-CMNIST color covariateaccuracy (%): 77.96 -> 28.60Hurt
use DIR instead of ERM on concept shiftDIR · GOOD-HIV size conceptROC-AUC (%): 63.26 -> 72.61Improved
use DIR instead of ERM on covariate shiftDIR · GOOD-HIV size covariateROC-AUC (%): 59.94 -> 57.11Hurt
use DIR instead of ERM on regressionDIR · GOOD-ZINC scaffold covariateMAE: 0.1995 -> 0.6493Hurt
use VREx instead of ERMVREx · GOOD-Motif base covariateaccuracy (%): 68.66 -> 71.47Improved
use GroupDRO instead of ERMGroupDRO · GOOD-ZINC scaffold covariateMAE: 0.1995 -> 0.1934Improved
use Mixup instead of ERM on node tasksMixup · GOOD-Cora degree conceptaccuracy (%): 60.54 -> 63.65Improved
use Mixup instead of ERM on graph tasksMixup · GOOD-HIV scaffold covariateROC-AUC (%): 69.58 -> 68.88No clear effect
use IRM instead of ERMIRM · GOOD-HIV scaffold covariateROC-AUC (%): 69.58 -> 67.97No clear effect
use EERM instead of ERM on WebKBEERM · GOOD-WebKB university covariateaccuracy (%): 14.29 -> 24.61Improved
use OOD validation set instead of ID validation set for model selectionERM · GOOD splitsOOD test scoreImproved

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