Generating Synthetic Time-Series Data on Edge Devices Using Generative Adversarial Networks

Synthetic data generation has recently garnered substantial attention in the research community. This method crafts data that mirrors authentic datasets and proves invaluable for machine learning endeavors, including classifier training and prediction tasks. Moreover, synthetic data addresses challenges related to data scarcity, preserving privacy, and analyzing specialized domains like healthcare and finance. This study introduces an approach to produce synthetic time series data on edge devices, leveraging the capabilities of Generative Adversarial Networks (GANs). This opens avenues to harness generative machine learning to tackle edge computing challenges. Our approach utilizes a GAN to create time series data that closely resembles datasets found in the real world. Furthermore, this GAN is implemented on an edge device, enabling realtime synthetic data generation. Our findings affirm the GAN's efficacy in producing data closely aligned with actual time series datasets, highlighting the potential of GANs in facilitating realtime synthetic data creation on edge devices for various machine learning applications.