AI Paper Atlas

Research map / Graph learning / GNN architectures

Explainability in Graph Neural Networks: A Taxonomic Survey

Hao Yuan, Haiyang Yu, Shurui Gui, Shuiwang Ji

TPAMI 2022 · Graph learning › GNN architectures · approach: GNN explainability · 767 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 subgraph-level explanation (SubgraphX) instead of edge masks (GNNExplainer)explanation method · Graph-SST2Fidelity+ (prob): 0.0953 -> 0.3152Improved
use SubgraphX instead of random importance scoresexplanation method · BA-2MotifsFidelity+ (prob): 0.4903 -> 0.8642Improved
use flow-based GNN-LRP instead of GNNExplainerexplanation method · Graph-TwitterFidelity+ (prob): 0.1452 -> 0.1931Improved
use GNNExplainer on node classification instead of graph classificationGNNExplainer · BA-Shapes vs Graph-SST2Fidelity+ (prob): 0.0953 -> 0.2925Mixed
use DeepLift (no graph-specific design) instead of random scoresDeepLift · BA-ShapesFidelity+ (prob): 0.1884 -> 0.0411Hurt
use gradient x input rule (GNN-GI) instead of random scoresGNN-GI · BA-2MotifsFidelity+ (prob): 0.4903 -> 0.0466Hurt
use SubgraphX instead of random importance scoresexplanation method · Graph-TwitterFidelity- (prob): 0.2825 -> -0.1085Improved
use GNN-LRP instead of GNN-GIexplanation method · BA-shapesexplanation accuracy (AUC-ROC): 0.8369 -> 0.9243Improved
use GNNExplainer instead of PGExplainerexplanation method · BA-Communityexplanation accuracy (AUC-ROC): 0.6843 -> 0.9194Improved
use DeepLift instead of GNN-LRPexplanation method · BA-shapesexplanation stability: 0.0432 vs 0.1872Mixed
raise the Sparsity thresholdsoft-mask explanation methods · all datasetsFidelity+Hurt

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