STEP-GAN: A One-Class Anomaly Detection Model with Applications to Power System Security

Abstract

Smart grid systems (SGSs), and in particular power systems, play a vital role in today's urban life. The security of these grids is now threatened by adversaries that use false data injection (FDI) to produce a breach of availability, integrity, or confidential principles of the system. We propose a novel structure for the multigenerator generative adversarial network (GAN) to address the challenges of detecting adversarial attacks. We modify the GAN objective function and the training procedure for the malicious anomaly detection task. The model only requires normal operation data to be trained, making it cheaper to deploy and robust against unseen attacks. Moreover, the model operates on the raw input data, eliminating the need for feature extraction. We show that the model reduces the well-known mode collapse problem of GAN-based systems, it has low computational complexity and considerably outperforms the baseline system (OCAN) with about 55% in terms of accuracy on a freely available cyber attack dataset.

DATIS: Data Augmentation for Trust Intensity Prediction in Incomplete Signed Networks

“…STEP-GAN [18,19] is a semi-supervised approach that has been specifically developed to detect anomalies from normal data as a binary classification task in cyber-security. This method is an extension of the regular GAN architecture and has several key differences.…”

“…DATIS is novel in addressing the imbalances in signed networks for improving NTI prediction task and regression model performance. We drew inspiration from STEP-GAN [18,19], which was initially developed for cyber-security purposes, specifically for anomaly detection. However, we propose a novel technology to generate synthetic data for data augmentation in imbalanced signed networks.…”