PIA-WAL: Peripheral Instance Augmentation for End-to-End Anomaly Detection using Weighted Adversarial Learning
In this work, we develop a weighted generative model by leveraging a few labelled anomalies for anomaly detection, named PIA-WAL. The goal of our model is to learn representative descriptions of normal instances in order to reduce the amount of false positives while maintaining accurate anomaly detection.
Zong W., Zhou F., Pavlovski M., Qian W., "Peripheral Instance Augmentation for End-to-End Anomaly Detection using Weighted Adversarial Learning"., Proceedings of the 27th International Conference on Database Systems for Advanced Applications (DASFAA-2022) , April, 2022.