Fog Enabled Distributed Training Architecture for Federated Learning

Abstract

The amount of data being produced at every epoch of second is increasing every moment. Various sensors, cameras and smart gadgets produce continuous data throughout its installation. Processing and analyzing raw data at a cloud server faces several challenges such as bandwidth, congestion, latency, privacy and security. Fog computing brings computational resources closer to IoT that addresses some of these issues. These IoT devices have low computational capability, which is insufficient to train machine learning. Mining hidden patterns and inferential rules from continuously growing data is crucial for various applications. Due to growing privacy concerns, privacy preserving machine learning is another aspect that needs to be inculcated. In this paper, we have proposed a fog enabled distributed training architecture for machine learning tasks using resources constrained devices. The proposed architecture trains machine learning model on rapidly changing data using online learning. The network is inlined with privacy preserving federated learning training. Further, the learning capability of architecture is tested on a real world IIoT use case. We trained a neural network model for human position detection in IIoT setup on rapidly changing data.

FIDEL: Fog integrated federated learning framework to train neural networks

“…Prior to this, we proposed a simulation-based federated learning architecture for machine learning tasks. 5 The work was based on synchronous FL for rapidly changing datasets. As an extension to Reference 5, this work proposes a federated learning paradigm for synchronous and asynchronous learning on volatile data using resource-constrained devices.…”

“…Motivated by the above‐mentioned challenges, we propose a fog‐integrated distributed learning framework – FIDEL( F og I ntegrate D f E derated L earning) for training neural networks on continuous datasets. Prior to this, we proposed a simulation‐based federated learning architecture for machine learning tasks 5 . The work was based on synchronous FL for rapidly changing datasets.…”

“…The work was based on synchronous FL for rapidly changing datasets. As an extension to Reference 5, this work proposes a federated learning paradigm for synchronous and asynchronous learning on volatile data using resource‐constrained devices. The novelty of the work lies in creating a generic neural network framework for training continuous data from multiple sources on failure‐prone and low‐computational resources.…”