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

TL;DR

This paper introduces SCVAE, a novel unsupervised anomaly detection model for IIoT edge devices that reduces model size and inference time without sacrificing performance.

Contribution

The paper presents a Squeezed Convolutional Variational AutoEncoder that incorporates Fire Modules to improve efficiency in anomaly detection for edge computing in IIoT.

Findings

Model achieves comparable performance to existing methods.

Model size and inference time are significantly reduced.

Effective on both benchmark and real-world data.

Abstract

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