Amazon Web Services brings machine learning to edge IoT

Public cloud provider @Amazon Web Services ( #AWS) has updated its edge computing platform, #Greengrass, to incorporate #machinelearning. The new offering brings together Internet of Things ( #IoT) capabilities, machine learning ( #ML), and edge computing to help organisations use AWS to build, train, and test ML models. These can then be deployed to small, low-powered, intermittently-connected IoT devices, including those running in factories, vehicles, fields, and homes. The latest version will be able to run #Apache #MXNet and TensorFlow Lite models locally on edge IoT devices using @NVIDIA #Jetson TX2 and @Intel #Atom architectures, according to @JeffBarr, AWS evangelist, in a blog post. The optimised libraries can take advantage of GPU and FPGA hardware accelerators at the edge in order to provide fast, local machine learning inferences. According to Barr, customers can also use Amazon SageMaker and other cloud-based ML tools to build, train, and test their models before deploying them to IoT devices. Users can also deploy 32-bit Raspberry Pi devices for local testing, he added. Use cases Barr said that Amazon has tossed the IoT, machine learning, and edge computing “into a blender” with the new offering. Among the potential use cases for Greengrass ML are precision farming, he said. Using the system, intelligent devices in the field could process images of soil, plants, pests, and crops, take local corrective action based on changing conditions, and send status reports to the cloud. The service could also be used for physical security, he added. Smart devices, such as Amazon’s deep-learning-enabled camera, DeepLens, could process images and scenes locally, looking for objects, watching for changes, and even detecting faces. “When something of interest or concern arises, the device can pass the image or the video to the cloud and use Amazon Rekognition to take a closer look,” explained Barr. The technology could also be used in industrial maintenance. Sensors would be able to run inference operations on power consumption, noise levels, and even vibration to flag anomalies, predict failures, and detect faulty equipment, he said. Internet of Business says With Dell, Microsoft, and a range of other vendors now focusing their IoT efforts on the critical edge environment and the distributed core, the race is on the make edge deployments faster, smarter, more intuitive – and cheaper – so that IoT projects can realise their full potential. In the long run, as the IoT rises, many organisations will find that their cloud investments will become more diffused. For many processing tasks, computing will need to be brought much closer to the point of delivery

https://internetofbusiness.com/aws-brings-machine-learning-inference-to-edge-iot/