English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We reveal a one-class homophily phenomenon, which is one prevalent propertywe find empirically in real-world graph anomaly detection (GAD) datasets, i.e.,normal nodes tend to have strong connection/affinity with each other, while thehomophily in abnormal nodes is significantly weaker than normal nodes. However,this anomaly-discriminative property is ignored by existing GAD methods thatare typically built using a conventional anomaly detection objective, such asdata reconstruction. In this work, we explore this property to introduce anovel unsupervised anomaly scoring measure for GAD, local node affinity, thatassigns a larger anomaly score to nodes that are less affiliated with theirneighbors, with the affinity defined as similarity on nodeattributes/representations. We further propose Truncated Affinity Maximization(TAM) that learns tailored node representations for our anomaly measure bymaximizing the local affinity of nodes to their neighbors. Optimizing on theoriginal graph structure can be biased by nonhomophily edges (i.e., edgesconnecting normal and abnormal nodes). Thus, TAM is instead optimized ontruncated graphs where non-homophily edges are removed iteratively to mitigatethis bias. The learned representations result in significantly stronger localaffinity for normal nodes than abnormal nodes. Extensive empirical results on10 real-world GAD datasets show that TAM substantially outperforms sevencompeting models, achieving over 10% increase in AUROC/AUPRC compared to thebest contenders on challenging datasets. Our code is available athttps://github.com/mala-lab/TAM-master/.

Truncated Affinity Maximization: One-class Homophily Modeling for Graph Anomaly Detection