Search papers, labs, and topics across Lattice.
Yunnan University
3
0
2
Decoupling topological and textual prototypes in TAG anomaly detection eliminates the noise amplification that obscures subtle anomalous signals.
A single graph foundation model can now achieve state-of-the-art anomaly detection across diverse graph domains, thanks to a new theory of "Anomaly Disassortativity" that tackles domain shift.
Forget rigid thresholds: GCTAM leverages contextual and global affinity to boost graph anomaly detection by 15-20% on real-world datasets, outperforming previous TAM-based methods.
