5 ms·
EdgeML: Algorithms for edge devices
- zitterbewegung 9y agoThis is a neat project and having this framework could really address two things. 1. If we bake all the intelligence on the device then it won't have to go outside to a network which could increase security. 2. If the intelligence is on the device then the users could control their data better.
- styfle 9y agoI assumed Microsoft EdgeML was related to Microsoft Edge browser...it is not. It sounds like and “edge device” refers to an IoT device.
- m3kw9 9y agoWhat is an Edge device??
- tyingq 9y ago"The trained models can be loaded onto edge devices such as IoT devices/sensors, and used to make fast and accurate predictions completely offline." Then, from: https://www.microsoft.com/en-us/research/project/resource-efficient-ml-for-the-edge-and-endpoint-iot-devices/ https://www.microsoft.com/en-us/research/project/resource-ef... "Our objective is to develop a library of efficient machine learning algorithms that can run on severely resource-constrained edge and endpoint IoT devices ranging from the Arduino to the Raspberry Pi."
- satya71 9y agoEdge, as in the edge of the internet, as opposed to data centers.
- eggy 9y agoEdge devices are what MS is calling IOT devices on cloud networks, I think: https://blogs.microsoft.com/iot/2017/05/10/microsoft-azure-iot-edge-extending-cloud-intelligence-to-edge-devices/ https://blogs.microsoft.com/iot/2017/05/10/microsoft-azure-i...
- rpeden 9y agoI used to work at a company whose software had to interact with security cameras. In the security/surveillance industry, at least, the term 'edge devices' has long been used to describe IP cameras and other devices like motion sensors operating in the field, transmitting their data back to a central server. Interestingly, 'edge analytics' has been a thing in that industry for a relatively long time, too. This is used to refer to cameras that do things like motion detection, person counting, and other analytics in the camera instead of having to post-process the video streams at a central location to extract that data. So in this case at least, is looks like MS is just using terminology that was already widely used in IoT devices even before "Internet of Things" was even really a phrase that was used to describe these devices.