AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Self-Supervised Graph Embedding Clustering

Fangfang Li, Quanxue Gao, Xiaoke Ma, Ming Yang, Cheng Dan Deng

TPAMI 2025 · Graph learning › Graph self-supervised learning · approach: Heterophily and imbalance on graphs · 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
add self-supervised label-guided manifold clustering (Our-LPP) over best prior one-step LDA methodUn-RTLDA · ORLACC: 0.663 -> 0.895Improved
add self-supervised label-guided manifold clustering (Our-LPP) over best prior baselineUn-TRLDA · UMISTACC: 0.501 -> 0.817Improved
add self-supervised label-guided manifold clustering (Our-MFA) over best prior baselineUn-RTLDA · ARACC: 0.561 -> 0.601Improved
use Our-MFA instead of Our-LPPOur-LPP · USPSACC: 0.826 -> 0.782Hurt
use Our-MFA instead of Our-LPPOur-LPP · ORLACC: 0.895 -> 0.848Hurt
use Our-MFA instead of Our-LPPOur-LPP · MSRCACC: 0.886 -> 0.929Improved
add dimensionality reduction before clusteringK-means · benchmark datasetsACC, NMI, PurityImproved
use one-step dimensionality reduction clustering instead of two-steptwo-step PCA/LPP/LDA + K-means · benchmark datasetsACC, NMI, PurityImproved
use local manifold LPP instead of global linear PCA before K-meansPCA-KM · benchmark datasetsACCImproved
increase neighbor number k from 2 toward n/c (ablation)Our-LPP · ORL, MSRC, Yalefaceclustering performance: peaks slightly below n/c then declinesMixed
increase neighbor number k (ablation)Our-LPP · UMISTclustering performance: peak at k=2Hurt
increase reduced dimension (ablation)Our-LPP · ORL, MSRC, UMIST, Yalefaceclustering performanceMixed

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