Abstract
Devices comprising the Internet of Things, such as sensors and small cameras,usually have small memories and limited computational power. The proliferationof such resource-constrained devices in recent years has led to the generationof large quantities of data. These data-producing devices are appealing targetsfor machine learning applications but struggle to run machine learningalgorithms due to their limited computing capability. They typically offloaddata to external computing systems (such as cloud servers) for furtherprocessing. The results of the machine learning computations are communicatedback to the resource-scarce devices, but this worsens latency, leads toincreased communication costs, and adds to privacy concerns. Therefore, effortshave been made to place additional computing devices at the edge of thenetwork, i.e close to the IoT devices where the data is generated. Deployingmachine learning systems on such edge devices alleviates the above issues byallowing computations to be performed close to the data sources. This surveydescribes major research efforts where machine learning has been deployed atthe edge of computer networks.