X
Tech

Microsoft pitches Windows 10 as an AI platform for developers

Microsoft's latest AI-infused message is that Windows 10 Redstone 4 will give developers a unique AI platform. But what does this really mean?
Written by Mary Jo Foley, Senior Contributing Editor

Every tech vendor these days is quick to slap the AI label on products and services. Up until today, I thought Microsoft had done an admirable job in refraining from doing this with Windows.

win10asanaiplatform.jpg
Credit: ZDNet

But the shark has been jumped as of March 7, the company's latest Windows Developer Day. Cue the eye rolls.

Microsoft is telling developers that the next release of Windows 10, which we are still calling by its codename, "Redstone 4," will enable developers to "use AI to deliver more powerful and engaging experiences."

Microsoft execs say there's now an AI platform in Windows 10 that enables developers to use "pre-trained machine learning in their apps on Windows 10 devices."

What does this really mean? Your guess is as good as mine.

I *think* Microsoft is trying to figure out how to include Windows 10 PCs in its "intelligent cloud, intelligent edge" mission statement. Intelligent edge was a big buzzword at last year's Build developer conference. (And is expected to be another big theme at this year's Build in May.) Put simply, it means by performing more processing locally on devices on the edge -- IoT devices, phones, AR/VR headsets like the HoloLens and PCs -- developers and users can achieve better performance.

Officials are telling devs they can use local processing capabilities of their PCs to do analysis and processing -- something they're currently already able to do. They're also advising developers they can use Microsoft's cloud AI platform to make use of training models in Azure and to process workloads in the cloud -- both of which are also things they can do today.

Microsoft already announced that Windows supports ONNX, an evolving standard format for machine language models that's being championed by Microsoft, Facebook and Amazon. Microsoft officials said starting with Visual Studio Preview 15.7, developers can add an ONNX file to a Universal Windows Platform project, enabling them to automatically generate a model.

Perhaps all this AI rhetoric is Microsoft's way of trying to make Windows 10 more interesting to Wall Street and others who are only interested in what Redmond is doing cloud-wise. Otherwise, I'm not sure there's a lot of there, there.

With its Office services, Bing, cognitive services and certain Microsoft-made apps like its Photos app, there's more of a case to be made that Microsoft is making use of AI technologies and techniques. But by claiming Windows 10 Redstone 4 uniquely includes some kind of AI platform, the Windows team has gone around the AI bend.

I'm watching the Windows Developer Day webcast right now. If there are any more tangibles on the Windows AI front, I'll update this post.

Update: Microsoft announced a preview of a new Windows 10 application programming interface (API) called Windows ML during the webcast. Officials said the API will enable developers to build machine-learning models, trained in Azure, right into their apps using Visual Studio and run them on their PCs. Windows ML will be available on every Windows edition some time in 2018, they said.

Microsoft officials said Windows ML will work with AI-specific processors, like Intel's Movidius VPU, at some point in the future.

Microsoft's documentation for the Windows ML platform is here.

That's it.

It's not really a part of Windows unless there's an architectural diagram! Thanks to Rene Schulte for this screen shot:

windowsmlarchitecture.jpg

Update No. 2: Here's another take on what today's announcement means from a different part of the company.

From Microsoft's Machine Learning blog, the post entitled "ONNX Models to be Runnable Natively on 100s of Millions of Windows Devices" explains the way developers can get ONNX models to include in their apps to run natively on devices with hardware acceleration.

Editorial standards