Build 2018 had a few strong themes this year, but arguably the most pervasive was the push on Artificial Intelligence. Microsoft’s AI push at build had two major pushes.
The first push was the ethics and responsibility of developers as we make use of AI, the second was making AI available to the masses. This makes now the perfect time if you have not already brought AI to your products.
“AI is Going To Shape All of What We Do” – Satya Nadella
Satya Nadella opened up the Build keynote with an important message about the consequences of Artificial Intelligence both in 2017 and 2018. Microsoft’s first announcement of Build was the formation of a Microsoft Ethics Committee whose job is to ensure that the company’s foray into cutting-edge techniques like deep learning doesn’t perpetuate societal biases in their products. This committee is focused on user privacy which Nadella refers to as a basic human right. All major tech companies are amassing large quantities of data that they are using to create more compelling products.
Microsoft announced many different new technologies or new ways to use AI and Machine Learning in their technologies. Some of the highlights include:
Project Brainwave is a way for developers to perform AI tasks quicker by using specialized FPGA chips built by Intel. This hardware is designed to reduce hardware latency by having specialized FPGA arrays for AI tasks.
Project Kinect for Azure
Kinect for Azure is bringing the now-defunct Xbox Kinect sensor and reusing the technology from both the Kinect and HoloLens to bring skeletal tracking and spatial understanding to a cloud-based IoT device.
ONNX Standard for Models
Open Neural Network Exchange (ONNX) was created by a consortium of companies to create an open ecosystem with portable Neural Network models that can work on multiple different platforms.
Containerization of Neural Networks
Use Azure Machine Learning Model Management to containerize Neural Networks and publish them to your Azure container registry and deploy them to your target in Docker containers.
ML .NET for C#
ML .NET is an open source and cross-platform machine learning framework to bring Machine Learning to .Net, allowing developers to work in either C# or F#.
A fast and easy way to collaborate between Data Scientists, Data Engineers, and Business Analysts with an Azure-based tool built on the Apache Spark platform.
AI on the Intelligent Edge
Leveraging computer power in IoT devices to bring Machine Learning to edge devices, allowing AI predictions to happen without the cloud.