AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

HSMAD: Heterophily-Driven Spectral and Manifold Learning for Graph Anomaly Detection

Chen Zhu, Yaying Zhang

ICML 2026 · Graph learning › Graph self-supervised learning · approach: Graph anomaly detection

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 Heterophily-Weighted Spectral Filtering (HWSF) (ablation)HSMAD · Average of 6 datasetsF1-Macro: 0.8809 -> 0.8953Improved
add Heterophily-Routed Manifold Update (HRMU) (ablation)HSMAD · Average of 6 datasetsF1-Macro: 0.8801 -> 0.8953Improved
add HWSF (ablation)HSMAD · Average of 6 datasetsAUROC: 0.9482 -> 0.9549Improved
add HRMU (ablation)HSMAD · Average of 6 datasetsAUROC: 0.9456 -> 0.9549Improved
add HWSF (ablation)HSMAD · YelpF1-Macro: 0.8306 -> 0.8682Improved
add HRMU (ablation)HSMAD · T-SocialF1-Macro: 0.9249 -> 0.9634Improved
add HRMU (ablation)HSMAD · WeiboAUROC: 0.9900 -> 0.9944Improved
heterophily-weighted edges versus unweighted or random-weighted edgesHSMAD · Amazon and Yelp subgraphs (2,000 nodes)spectral energy concentrationImproved
increase spectral filter order CHSMAD · YelpF1-Macro and AUROCImproved
vary edge mask quantile tau from 10% to 30%HSMAD · YelpF1-Macro and AUROCMixed
HSMAD versus PMP and CGADM on very sparse graphHSMAD · T-SocialAUROC: 0.9888 vs 0.9946 (PMP)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