On the Private Data Synthesis Through Deep Generative Models for Data Scarcity of Industrial Internet of Things

Due to the data-driven intelligence from the recent deep learning based approaches, the huge amount of data collected from various kinds of sensors from industrial devices have the potential to revolutionize the current technologies used in the industry. To improve the efficiency and quality of machines, the machine manufacturer needs to acquire the history of the machine operation process. However, due to the business secrecy, the factories are not willing to do so. One promising solution to the abovementioned difficulty is the synthetic dataset and an informatic network structure, both through deep generative models such as differentially private generative adversarial networks. Hence, this article initiates the study of the utility difference between the abovementioned two kinds. We carry out an empirical study and find that the classifier generated by private informatic network structure is more accurate than the classifier generated by private synthetic data, with approximately 0.31–7.66%.

I. Introduction

Thanks to the rapid rise of the Internet of Things, there are increasing demands and novel user scenarios for human life. Smart appliances [1], autonomous driving [2], intelligent robots [3] are exemplar applications with a considerable number of devices connected to each other. For industry, wireless communications and artificial intelligence (AI) jointly promote the development of IIoT [4], [5]. In particular, IIoT has been witnessed to significantly improve manufacturing efficiency, reduce product cost, and upgrade the manufacturing process by integrating various sensors and controllers with intelligent analysis. While the intelligence in IIoT is the core component that leads to the abovementioned benefits, a critical part behind the scene is the abundance of the data.