Adaptive graph-based algorithms for conditional anomaly detection and semi-supervised learning

2026-05-05

AI Summary

This work addresses the computational and storage bottlenecks of graph-based methods in large-scale or streaming data settings, as well as the challenges posed by marginal and isolated points in conditional anomaly detection. To this end, the authors propose a similarity-graph-based semi-supervised learning framework that constructs approximate graphs incrementally, compresses the graph size via local representative points to maintain low distortion, and enhances stability through regularized harmonic solutions. Additionally, a non-parametric graph strategy is introduced to effectively handle the sparse structures inherent in conditional anomalies. The proposed approach is validated on a clinical patient management task for behavioral anomaly detection and has been evaluated by 15 intensive care specialists, who confirmed its reliability and practical utility.

Abstract

We develop graph-based methods for semi-supervised learning based on label propagation on a data similarity graph. When data is abundant or arrive in a stream, the problems of computation and data storage arise for any graph-based method. We propose a fast approximate online algorithm that solves for the harmonic solution on an approximate graph. We show, both empirically and theoretically, that good behavior can be achieved by collapsing nearby points into a set of local representative points that minimize distortion. Moreover, we regularize the harmonic solution to achieve better stability properties. We also present graph-based methods for detecting conditional anomalies and apply them to the identification of unusual clinical actions in hospitals. Our hypothesis is that patient-management actions that are unusual with respect to the past patients may be due to errors and that it is worthwhile to raise an alert if such a condition is encountered. Conditional anomaly detection extends standard unconditional anomaly framework but also faces new problems known as fringe and isolated points. We devise novel nonparametric graph-based methods to tackle these problems. Our methods rely on graph connectivity analysis and soft harmonic solution. Finally, we conduct an extensive human evaluation study of our conditional anomaly methods by 15 experts in critical care.

Problem

• conditional anomaly detection
• semi-supervised learning
• graph-based methods
• fringe points
• isolated points

Innovation

• graph-based learning
• conditional anomaly detection
• online label propagation
• harmonic solution regularization
• nonparametric graph methods