Communicate to Learn at the Edge

Abstract

Bringing the success of modern machine learning (ML) techniques to mobile devices can enable many new services and businesses, but also poses significant technical and research challenges. Two factors that are critical for the success of ML algorithms are massive amounts of data and processing power, both of which are plentiful, yet highly distributed at the network edge. Moreover, edge devices are connected through bandwidth-and power-limited wireless links that suffer from noise, time-variations, and interference. Information and coding theory have laid the foundations of reliable and efficient communications in the presence of channel imperfections, whose application in modern wireless networks have been a tremendous success. However, there is a clear disconnect between the current coding and communication schemes, and the ML algorithms deployed at the network edge. In this paper, we challenge the current approach that treats these problems separately, and argue for a joint communication and learning paradigm for both the training and inference stages of edge learning. I. MOTIVATIONModern machine learning (ML) techniques have made tremendous advances in areas such as machine vision, robotics, and natural language processing. Novel ML applications emerge every day, ranging from autonomous driving and finance to marketing and healthcare -potential applications are limitless. In parallel, the fifth generation (5G) of mobile technology promises to connect billions of heterogeneous devices to the network edge, supporting new applications and verticals under the banner of Internet of things (IoT). Edge devices will collect massive amounts of data, opening up new avenues for ML applications. The prevalent approach for the implementation of ML solutions on edge devices is to amass all the relevant data at a

Resource Allocation for Intelligent Reflecting Surfaces Assisted Federated Learning System with Imperfect CSI

“…Fortunately, due to the significant increase in data processing capabilities of mobile devices, the concept of edge learning has been introduced to solve these problems, which processes data at the edge rather than at the cloud center. Federated learning (FL) is one of the most promising edge learning frameworks, where user devices (UDs) only send local models calculated by local resource to the base station (BS) without sharing local data [1,2].…”

Federated Edge Learning with Misaligned Over-The-Air Computation

“…Increasingly complex machine learning (ML) models are trained and deployed to gather intelligence from the data collected by edge devices [1], [2]. While this is conventionally done at a cloud server [3], offloading huge amounts of edge data to centralized cloud servers is not sustainable, and will potentially cause significant network congestion [4]. Moreover, data from edge devices contain user-specific features, and centralized processing also causes privacy concerns.…”

“…In the case of edge devices, often the devices that collaborate to learn a common model are within physical proximity of each other, and are coordinated by a nearby access point, e.g., a base station acting as the PS. In this, so-called federated edge learning (FEEL) scenario [4], the UL model aggregation step is particularly challenging as the wireless medium is shared among all the participating devices. Traditional radio access network (RAN) technologies distribute channel resources among the devices by means of orthogonal multiple-access technologies [8] (e.g., TDMA, CDMA, OFDMA).…”

Hierarchical Over-the-Air Federated Edge Learning

“…In FL, several data owners called mobile users (MUs) are selected based on some criteria such as their computing capability, data quality, available power, and location [3]. Each MU in the federation trains a local model using its own data and computing power in every iteration.…”