Windows ML

from Wikipedia, the free encyclopedia

Windows ML is a programming interface (API) from Microsoft for machine learning to Windows - operating systems . According to Microsoft, the aim is to bring AI applications with artificial neural networks to more client computers and to make it easier for developers to run AI software locally on the client PC. This should also save bandwidth and reduce latency, for example when evaluating images.

For this purpose, AI accelerators such as the Visual Processing Unit (VPU) Movidius Myriad X from Intel are used. Other graphics chips and the CPU are also supported. Windows ML uses the Open Neural Network Exchange Format ONNX to transmit the training data for the local neural networks (CNN) . Training of AI can in the Microsoft Azure - Cloud are made.

Windows ML should appear with the next Windows 10 update.

Individual evidence

  1. ↑ heise online: Windows ML: Machine Learning / AI for the PC. Accessed March 8, 2018 (German).
  2. ↑ Windows 10: Redstone 4 brings hardware accelerated AI support . ( winfuture.de [accessed on March 8, 2018]).
  3. ↑ Windows 10's next major update will include Windows ML, a new AI platform . In: The Verge . ( theverge.com [accessed March 8, 2018]).