Crowdtraining: Architecture and incentive mechanism for deep learning training in the internet of things

We have witnessed the great advances in the Internet of Things (IoT) that bring great opportunities for promoting industrial upgrades and even allow the introduction of the fourth industrial revolution. In IoT, data is generated at an unprecedented scale, and it needs to be analyzed efficiently to draw meaningful insights. In recent years, deep learning has emerged as a powerful tool for implementing this analysis. Unfortunately, the traditional deep learning algorithm running in the cloud data center needs a large amount of data for effective training, which takes a substantial amount of time. In this article, we propose a concept of crowdtraining, which employs edge computing units (ECUs) for big data training in IoT and pays rewards to ECUs using a gamebased incentive mechanism. Simulations have been performed to evaluate the performance of a crowdtraining game.