A Hybrid Deep Learning Based Anomaly Detection Framework dedicated for Big Research Infrastructures

Anomaly detection has been an active artificial intelligence area in research and industrial communities with many unique problem complexities and challenges that require advanced approaches. In this research we concentrate on the analysis of anomalies occurring in big research infrastructures including signal measurement analysis as well as X-ray radiation images classification. The proposed hybrid end-to-end anomaly score learning framework consists of two deep learning based approached for anomaly detection dedicated for two different sources of high-dimensional data. The aim of the system is to identify abnormal status of sensors in certain time steps. In this study, we deploy two transfer learning architectures by examining pre-trained VGG-16, VGG-19, InceptionV3, Xception, as well as EfficientNet CNN models with an adjusted densely-connected classifiers. The system achieved overall accuracy of 94%, what is a state-of-the-art result in such research infrastructures.