Learnable Kernel Density Estimation for Graphs and Its Application to Graph-Level Anomaly Detection
arXiv:2505.21285v4 Announce Type: replace Abstract: This work proposes a framework LGKDE that learns kernel density estimation for graphs. The key challenge in graph density estimation lies in effectively capturing both structural patterns and semantic variations while maintaining theoretical guarantees. Combining graph kernels and kernel density estimation (KDE) is a standard approach to graph density estimation, but has unsatisfactory performance due to the handcrafted and fixed features of kernels. Our method LGKDE leverages graph neural networks to represent each graph as a discrete distri
The continuous advancements in graph neural networks and the increasing complexity of graph-structured data necessitate more sophisticated anomaly detection methods for robust AI systems.
Improved graph anomaly detection could significantly enhance the reliability and security of complex AI applications across various domains, from cybersecurity to fraud detection and critical infrastructure monitoring.
The ability to learn kernel density estimation for graphs directly empowers AI systems to identify unusual patterns in highly interconnected data structures with greater precision and adaptability.
- · AI/ML researchers
- · Cybersecurity sector
- · Financial services
- · Critical infrastructure operators
- · Legacy anomaly detection systems
- · Attackers relying on subtle graph manipulations
More accurate and adaptive anomaly detection in complex, interconnected systems.
Reduced false positives and improved threat intelligence in areas like network security and transactional fraud.
Enhanced resilience of AI-driven systems against sophisticated adversarial attacks and emergent failures.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG