Kinect for Azure re-packages the tech from Microsoft’s shuttered Xbox motion-sensor
MICROSOFT HAS GONE HEAVY on artificial intelligence (AI) at its Build 2018 conference, surprising pretty much no one.
That’s because every tech bod and their dog are working on AI at the moment and Microsoft proved that by revealing a load of AI-centric stuff; here are the noteworthy ones, in our most humble of opinions.
First out of the blocks is Redmond’s work and a limited beta of squashing Cortana and Alexa together to allow the two virtual assistants to chat together.
Talking of talking to AI, Microsoft highlighted an overhaul of its conversational AI bots, which should make it easier for developers to get such smart collections of code up and running on the firm’s Azure cloud platform or their own servers, with more options for customisation and integration.
For example, a new project dubbed Project Conversation Learner will let developers feed conversations into the bot framework and let Redmond’s machine learning tech figure out new dialogue sequenced on top of them.
Speaking of projects, Microsoft also showed off Project Brainwave, a deep learning neural network architecture that’s designed, in theory, to make Azure the fastest cloud platform for running AI systems on.
Then there’s the open sourcing of Azure IoT Edge Runtime, which is Microsoft’s effort to bring AI smarts to Internet of Things (IoT) devices at – you guessed it – the edge of networks of connected gadgets.
The open sourcing of the runtime will give developers more scope to modify, debug and control the machine learning intelligence they work into IoT devices locally rather than need to connect back to a central server or cloud. This helps reduce the latency and improving the speed at which said gadgets smartly crunch and serve up data or requests. Part of this included Azure Cognitive Services, which give developers the ability to more readily plonk AI algorithms for say language interpretation into their gadgets.
A similar service in the form of Custom Vision is also coming to Azure IoT Edge, which will enable things like industrial equipment and drones to gain smart image recognition and computer vision capabilities running on them locally.
Azure IoT Edge looks to be positioned as the backbone of future connected and smart devices, while the cognitive services on offer will allow developers to easily create specific smart capabilities for said devices.
Another part of this will involve Microsoft’s partnership with Qualcomm to create a “vision AI developer kit” for Azure IoT Edge. This will arrive in the form of camera-based IoT tech with that makes use of Qualcomm’s AI engine and its smart vision tech combined with the Azure cloud to deliver computer vision and image recognition smarts, as well as stream analytics and other cognitive services on devices locally rather than relying on a constant cloud connection.
Kinect is also making a comeback in the form of Project Kinect for Azure, which mixes the scrapped motion-sensing camera hardware with Redmond’s cloud-powered AI tech to provide a hardware and services package; think the camera sensors and lenses that can be popped into devices like smart home tech or drones and make use of Redmond’s AI tech, rather than the OG Kinect bar.
Microsoft touted the “countless new opportunities” for mixing Kinect with machine learning but didn’t really name an example.
Rather, Redmond wants to see what developers do when they combine depth-sensors tech data with AI, as tech tinkerers have previously used Kinect in all sorts of applications beyond its original Xbox origins; the camera tech has popped up in medical, industrial, robotics and education applications, according to Microsoft chief exec Satya Nadella.
Rounding everything off, Microsoft also announced a £25m five-year commitment to help fund the use of AI to help people with disabilities. Developers, universities, NGOs, and other people working on worthy uses of AI could be up for a slice of the funding, which looks to be Redmond’s efforts to see AI put to good use rather than develop smart tech that nicks jobs or destroys humanity.
All in all, it looks like Microsoft is moving away from the “mobile first, cloud first” maxim it has been previously touting for the past few years and is now positioning itself as an AI and intelligent edge firm.
Naturally, the mobile and cloud stuff will still have a part to play, but that’s become an expected part of the technology industry; embracing AI is the new thing big tech firms like to spout about these days, which isn’t surprising given the impact smarter tech could have on the way we do everything from watching TV to getting around. µ
Source : Inquirer