Squeezed Convolutional Variational AutoEncoder for Unsupervised Anomaly Detection in Edge Device Industrial Internet of Things

摘要

In this paper, we propose Squeezed Convolutional Variational AutoEncoder(SCVAE) for anomaly detection in time series data for Edge Computing inIndustrial Internet of Things (IIoT). The proposed model is applied to labeledtime series data from UCI datasets for exact performance evaluation, andapplied to real world data for indirect model performance comparison. Inaddition, by comparing the models before and after applying Fire Modules fromSqueezeNet, we show that model size and inference times are reduced whilesimilar levels of performance is maintained.