Training Machine Learning models at the Edge - A Survey - Edge learning training methods overview

Aug 3, 2026

Upload by Hazel

Edge computing accelerates AI capability placement by moving intelligence closer to devices, yet training ML models at the edge remains comparatively underexplored. This survey investigates edge learning and how to optimize ML training directly on edge resources, synthesizing approaches, comparing techniques, and mapping challenges and future trends. Using Scopus and Web of Science advanced search, the survey identifies literature with emphasis on distributed learning, especially federated learning, and discusses available frameworks, libraries, and simulation tools.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages30

Reading Time1 h 16 min

AccessFree to download

Introduction

Edge learning and motivation

Challenges of edge-side training

Existing approaches and related methods

Survey scope and contribution

Training Machine Learning models at the Edge: A Survey Aymen Rayane Khouas, Mohamed Reda Bouadjenek, Hakim Hacid, and Sunil Aryal arXiv :2403 .02619v3 [ cs .LG] 11 Oct 2024 Abstract—Edge computing has gained significant traction in recent years, promising enhanced efficiency by integrating artificial intelligence capabilities at the edge. While the focus has primarily been on the deployment and inference of Machine Learning (ML) models at the edge, the training aspect remains less explored. This survey, explores the concept of edge learning, specifically the optimization of ML model training at the edge. The objective is to comprehensively explore diverse approaches and methodologies in edge learning, synthesize existing knowledge, identify challenges, and highlight future trends. Utilizing Scopus and Web of science advanced search, relevant literature on edge learning was identified, revealing a concentration of research efforts in distributed learning methods, particularly federated learning. This survey further provides a guideline for comparing techniques used to optimize ML for edge learning, along with an exploration of the different frameworks, libraries, and simulation tools available. In doing so, the paper contributes to a holistic understanding of the current landscape and future directions in the intersection of edge computing and machine learning, paving the way for informed comparisons between optimization methods and techniques designed for training on the edge. Index Terms—Machine Learning; Edge Computing; Edge AI; Edge Learning; On-Device Training; Edge intelligence; Artificial Intelligence; IoT. I. INTRODUCTION In recent years, the fields of Artificial Intelligence (AI) and Machine Learning (ML) have witnessed significant growth, and have demonstrated remarkable success across various industrial applications [1] . ML’s essence lies in the interplay between algorithmic models and large quantities of data, asthe latter is often required to successfully train ML models. Traditionally, datasets have been collected in cloud storage, databases, and data lakes. These datasets are then processed in central cloud servers to train various ML models. Conversely, the rapid proliferation of smart devices and sensors in recent years has led to an explosion of data generation at the edge of the network. With edge devices generating vast quantities of data closer to the source, growing concerns about privacy and security, as well as the desire to optimize the bandwidth consumption on the increasing number of edge devices and reduce the computational load on cloud servers, have driven a paradigm shift towards edge computing. In this context, computational processes are decentralized and A.R. Khouas, M. R. Bouadjenek, and A. Aryal are with the School of Information Technology, Deakin University, Waurn Ponds Campus, Geelong, VIC 3216, Australia. H. Hacid is with the Technology Innovation Institute, UAE. E-mail: [a.khouas@deakin.edu.au](a.khouas@deakin.edu.au) (corresponding author) Manuscript received XXX YY, ZZZZ; revised XXX YY, ZZZZ. migrated to edge devices. This sets the stage for a novel intersection between ML and edge computing. This shift has sparked growing interest towards edge ML. A union between machine learning and edge computing, deploying ML models at the edge, closer to end devices, enabling inference or training to occur at the edge. Edge learning is a subset of Edge ML that involves training ML models directly at the edge. Traditionally, ML models have relied on cloud infrastructure for training and deployment. However, this approach poses several challenges. These include high latency, significant communication overheads, and concerns around data privacy and security. By processing data closer to its source, edge learning tackles these challenges while enabling real-time decision-making and reducing cloud resource usage. Furthermore, this enables innovative ML applications, such as pr

FAQs

What problem does the survey address in edge computing and machine learning?

It focuses on the relatively less explored training side of machine learning on the edge, aiming to optimize ML model training where resources are constrained.

How does the survey identify and synthesize prior work?

It uses Scopus and Web of Science advanced search to collect relevant edge learning literature and then synthesizes the resulting body of knowledge.

Which research direction receives particular emphasis in the survey?

Distributed learning methods are emphasized, especially federated learning, as a key approach to enable collaborative training across edge devices.

Aug 3, 2026

Upload by Hazel