Azure Machine Learning—what’s new from Build 2020

Machine learning (ML) is gaining momentum across a number of industries and scenarios as enterprises look to drive innovation, increase efficiency, and reduce costs. Microsoft Azure Machine Learning empowers developers and data scientists with enterprise-grade capabilities to accelerate the ML lifecycle. At Microsoft Build 2020, we announced several advances to Azure Machine Learning across the following areas: ML for all skills, Enterprise grade MLOps, and responsible ML.

ML for all skills

New enhancements provide ML access for all skills.

Enhanced notebook in preview

Data scientists and developers can now access an enhanced notebook editor directly inside Azure Machine Learning studio. New capabilities to create, edit, and collaborate make remote work and sharing easier for data science teams and the notebook is fully compatible with Jupyter.

  • Boost development productivity with features like IntelliSense, inline error highlighting, and code suggestions from VSCode, which deliver the best-in-class coding experience in Jupyter notebooks.
  • Access real-time co-editing (coming soon) for seamless remote collaboration or pair debugging.
  • Inline controls to start, stop, and create a new compute using GPU or CPU Compute Instance inside notebooks.
  • Add new kernels to the notebook editor and quickly switch between different kernels like Python and R.

Enhanced notebook editor inside Azure Machine Learning studio with three users and IntelliSense.

Real-time notebook co-editing with three users and IntelliSense.

Reinforcement learning support in preview

New reinforcement learning support in Azure Machine Learning enables data scientists to train agents who interact with the real world, such as control systems and game characters. To train agents on Azure Machine Learning, data scientists can use the SDK, studio UI, or command line interface (CLI). Azure Machine Learning simplifies running reinforcement learning at scale on remote compute clusters, including tracking experiment results in Tensorboard and Azure Machine Learning studio UI. See sample notebooks to train an agent to navigate a lava maze in Minecraft using Azure Machine Learning.

Animation of an agent navigating a maze in Minecraft.

An agent successfully navigates the maze in Minecraft.

Data labeling in preview

Projects that have a computer-vision component, such as image classification or object detection, generally require labels for thousands of images. Data labeling in Azure Machine learning gives you a central place to create, manage, and monitor labeling projects. Use it to coordinate data, labels, and efficiently manage labeling tasks. The new ML assisted labeling feature helps trigger automatic machine learning models to accelerate the labeling task and is available for image classification (multi-class or multi-label) and object detection tasks.

Enterprise-grade MLOps

New features for MLOps designed to deliver innovation faster.

Azure Private Link for network isolation in preview

To enable secure model training and deployment, Azure Machine Learning provides a strong set of data and networking protection capabilities. These include support for Azure Virtual Networks, dedicated compute hosts and customer managed keys for encryption in transit and at rest. In addition, we are enabling Private Link for network isolation to access Azure Machine Learning over a private endpoint in your virtual network, so the Azure Machine Learning workspace will not be accessible to the internet. This is critical for many scenarios in regulated industries like financial services, insurance, and healthcare.

Azure Cognitive Search integration in preview

Many enterprises have a large corpus of documents and can build cognitive search solutions to search for specific terms and find relevant results to improve productivity. To build an effective solution, often customized models are needed to enrich the search experience. Using Azure Machine Learning, developers can deliver custom search solutions by training and deploying models and now, seamlessly integrating the end points into the Azure Cognitive Search skillset.

Responsible ML

In collaboration with the Aether Committee and its working groups, we are bringing the latest research in responsible AI to Azure. The new responsible ML capabilities in Azure Machine Learning and our open-source toolkits empower data scientists and developers to understand ML models, protect people and their data, and control the end-to-end ML process. To learn more, read the responsible ML announcements from Build.

Innovating with customers

We continue to drive this innovation hand-in-hand with you, our customers. For example, Carhartt turned to Azure Machine Learning for quantitative insights to help their company get the right products to the places its customers work and live.

“The model we deployed on Azure Machine Learning helped us choose the three new retail locations we opened in 2019. Those stores exceeded their revenue plans by over 200 percent in December, the height of our season, and within months of opening were among the best-performing stores in their districts.” Jolie Vitale, Director of BI and Analytics, Carhartt.

Start building today!

We hope you will join us and start your journey with Azure Machine Learning.

Source: Azure Blog Feed

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.