Microsoft is giving Windows developers another reason to look forward to the Windows 10 feature update, code-named Redstone 4, this spring.
At the Windows Developers Day event on March 7, local time, Microsoft announced the launch of the artificial intelligence platform Windows ML, allowing developers to use pre-trained machine learning systems in their applications to create applications with more features.
Microsoft's Windows Machine Learning Model is designed to run on a variety of different devices including PCs, IoT devices, servers, data centers and HoloLens headsets, and AI processors such as Intel's Movidius VPU will also be supported.
Encoders that are difficult to solve using traditional algorithms will be able to take a pre-trained machine learning model and display it as "first-class object" in Visual Studio, which developers can then use to add AI functionality to the application. For example, a manufacturer can use image classifiers based on good photos of existing photos and damaged parts and use the resulting models to build differentiated applications. Microsoft team project manager VedBrat said.
Why evaluate machine learning models on your local PC instead of using AI-enabled cloud products, including Microsoft's own Azure platform?
While the cloud does enable artificial intelligence workloads to harness the computing power of very large data centers, there are some important benefits of using a client-based approach, explained Kevin Gallo, Microsoft vice president.
He said: "The AI assessment conducted on the local system" can reduce latency and provide you with real-time results. “ Users will be able to use the processing power of the Windows desktop or other devices to run large amounts of video, images and other local data analysis.
It is understood that Windows ML can efficiently utilize hardware for any given artificial intelligence workload and intelligently distribute work across multiple hardware types, including Intel's Visual Processing Unit (VPU).
"On Windows, we have used AI in a variety of ways," said Kam Ved Brat, Microsoft team project manager, who explained Windows ML on the Microsoft Developer Day Webcast. "We use it for handwriting recognition, isolating security threats, and services like Cortana... Then why don't we consider bringing AI to Windows?" & rdquo;
Leifeng learned that in the area of IT planning and management, companies can use server logs and any other types of information collected by their systems to inform AI-enabled applications to assess past patterns and help users make more informed technical decisions .
In addition, VedBrat revealed that "Windows 10 will support ONNX", an open source format for AI framework interoperability that was originally developed by Microsoft and Facebook. In a preview of Visual Studio 15.7, users will be able to add ONNX files to the UWP (Common Windows Platform) project and the IDE will automatically create the necessary model interfaces.