Microsoft, on Thursday, announced a series of improvements to its Azure cloud platform that are designed to give developers increased capabilities for developing artificial intelligence (AI), machine learning, mixed reality, blockchain and Internet of Things (IoT) solutions, via both hybrid cloud and edge architectures.
Azure AI and Machine Learning:
The company had previewed a number of new services and tools expanding Microsoft’s cloud-based AI capabilities for developers and data scientists. The company says that over 1.3 million developers use Azure Cognitive Services, and it’s now announcing a new category called Decision. It is meant to provide recommendations to help people to make decisions. The feature includes Personalizer, which picks up knowledge from experience and then offers recommendations to help people make decisions.
“We are able to take reinforcement learning and ship it in a way that’s accessible to developers and doesn’t require a data scientist,” said Lance Olson, director of program management for applied AI at Microsoft. “That will be very impactful for customers.”
Microsoft also announced another new service called Content Moderator and talked of other decision services such as Ink Recognizer for embedding digital ink recognition, and Form Recognizer to better automate data entry. The recently announced Anomaly Detector is also in the Decision category of Azure Cognitive Services.
Microsoft also announced the general availability of ONNX Runtime support for Nvidia Tensor RT and Intel nGraph, along with hardware-accelerated inferencing models on FPGAs. There are also new MLOps features for Azure DevOps, automated machine learning advancements, and a new visual machine learning interface. Microsoft boasted that it is now an active contributor to the MLflow project, an open source platform for the machine learning lifecycle.
Microsoft can be said as a fan of the Open Neural Network Exchange (ONNX) runtime to create deployment models that work across multiple platforms. To that end it has made available support for NVIDIA TensorRT and Intel nGraph chips for high-speed inferencing.
Another announcement today is Azure Blockchain Service.