Azure Machine Learning
Enterprise-grade machine learning service to build and deploy models faster
Accelerate the end-to-end machine learning lifecycle
Empower data scientists and developers with a wide range of productive experiences to build, train, and deploy machine learning models and foster team collaboration. Accelerate time to market with industry-leading MLOps—DevOps for machine learning. Innovate on a secure, trusted platform, designed for responsible machine learning.
Machine Learning for all skill levels
Productivity for all skill levels, with Jupyter notebooks, drag-and-drop designer, and automated machine learning.
End-to-end MLOps
Robust MLOps capabilities that enable creation and deployments of models at scale using automated and reproducible machine learning workflows.
Responsible machine learning innovation
Rich set of built-in responsible capabilities to understand, protect, and control data, models and processes.
Open and interoperable
Best-in-class support for open-source frameworks and languages including MLflow, Kubeflow, ONNX, PyTorch, TensorFlow, Python, and R.
Boost productivity with machine learning for all skills
Rapidly build and deploy machine learning models using tools that meet your needs regardless of skill level. Use built-in Jupyter Notebooks with Intellisense or the drag-and-drop designer. Accelerate model creation with automated machine learning, and access powerful feature engineering, algorithm selection, and hyperparameter-sweeping capabilities. Increase team efficiency with shared datasets, notebooks, models, and customizable dashboards that track all aspects of the machine learning process.
Operationalize at scale with MLOps
Take advantage of MLOps to streamline the machine learning lifecycle, from building models to deployment and management. Create reproducible workflows with machine learning pipelines, and train, validate and deploy thousands of models at scale, from the cloud to the edge. Use Azure DevOps or GitHub Actions to schedule, manage, and automate the machine learning pipelines, and use advanced data-drift analysis to improve model performance over time.
Build responsible machine learning solutions
Access state-of-the-art responsible machine learning capabilities to understand, protect, and control your data, models, and processes. Explain model behavior during training and inferencing, and build for fairness by detecting and mitigating model bias. Preserve data privacy throughout the machine learning lifecycle with differential privacy techniques, and use confidential computing to secure machine learning assets. Automatically maintain audit trails, track lineage, and use model datasheets to enable accountability.
Innovate on an open and flexible platform
Get built-in support for open-source tools and frameworks for machine learning model training and inferencing. Use familiar frameworks like PyTorch, TensorFlow, or scikit-learn, or the open and interoperable ONNX format. Choose the development tools that best meet your needs, including popular IDEs, Jupyter Notebooks, and CLIs, or languages such as Python and R. Use ONNX Runtime to optimize and accelerate inferencing across cloud and edge devices. Track all aspects of your training experiments using MLfLow.
Build your machine learning skills with Azure
Learn more about machine learning on Azure and participate in hands-on tutorials with this 30-day learning journey. At the end of this learning journey, you'll be prepared to take the Azure Data Scientist Associate Certification.
Advanced security, governance, and hybrid infrastructure
- Train models on your hybrid infrastructure using Kubernetes clusters on-premises, across multicloud environments, and at the edge with Azure Arc interoperability.
- Access security capabilities such as role-based access, custom machine learning roles, virtual networks, and private links. Manage governance with policies, audit trails, quota, and cost management.
- Streamline compliance with a comprehensive portfolio spanning 60 certifications, including FedRAMP High and DISA IL5.
Key service capabilities
Collaborative notebooks
Maximize productivity with IntelliSense, easy compute and kernel switching, and offline notebook editing.
Automated machine learning
Rapidly create accurate models for classification, regression, and time-series forecasting. Use model interpretability to understand how the model was built.
Drag-and-drop machine learning
Use machine learning tools like designer with modules for data transformation, model training, and evaluation, or to easily create and publish machine learning pipelines.
Data labeling
Prepare data quickly, manage and monitor labeling projects, and automate iterative tasks with machine learning–assisted labeling.
MLOps
Use the central registry to store and track data, models, and metadata. Automatically capture lineage and governance data. Use Git to track work and GitHub Actions to implement workflows. Manage and monitor runs, or compare multiple runs for training and experimentation.
Autoscaling compute
Use managed compute to distribute training and to rapidly test, validate, and deploy models. Share CPU and GPU clusters across a workspace and automatically scale to meet your machine learning needs.
Deep integration with other Azure services
Accelerate productivity with built-in integration with Microsoft Power BI and Azure services such as Azure Synapse Analytics, Azure Cognitive Search, Azure Data Factory, Azure Data Lake, Azure Arc and Azure Databricks.
Hybrid and multicloud support
Run machine learning on existing Kubernetes clusters on-premises, in multi-cloud and at the edge with Azure Arc. Use the simple one-click deploy ML agent to start training models securely, wherever your data lives.
Reinforcement learning
Scale reinforcement learning to powerful compute clusters, support multi-agent scenarios, and access open-source reinforcement learning algorithms, frameworks, and environments.
Responsible machine learning
Get model transparency at training and inferencing with interpretability capabilities. Assess model fairness through disparity metrics and mitigate unfairness. Protect data with differential privacy and confidential machine learning pipelines.
Enterprise-grade security
Build and deploy models more securely with network isolation and private link capabilities, role-based access control for resources and actions, custom roles, and managed identity for compute resources.
Cost management
Better manage resource allocations for Azure Machine Learning compute instances with workspace- and resource-level quota limits.
Pay only for what you need, with no upfront cost
Mastering Azure Machine Learning
Master expert techniques for building automated and highly scalable end-to-end machine learning models and pipelines in Azure using TensorFlow, Spark, and Kubernetes.
Principles of Data Science
Many people working with data have developed skills in math, programming, or domain expertise, but proper data science calls for all three. This comprehensive e-book helps fill in the gaps.
A Forrester Wave Leader 2020
Forrester names Microsoft Azure Machine Learning a leader in The Forrester Waveâ„¢: Notebook-Based Predictive Analytics And Machine Learning, Q3 2020.
How to use Azure Machine Learning
Go to your studio web experience
Build and train
Deploy and manage
Author new models and store your compute targets, models, deployments, metrics, and run histories in the cloud.
Use automated machine learning to identify algorithms and hyperparameters and track experiments in the cloud. Author models using notebooks or the drag-and-drop designer.
Deploy your machine learning model to the cloud or the edge, monitor performance, and retrain it as needed.
Resources
Beginner tutorials
Advanced tutorials
Additional resources
Start using Azure Machine Learning today
Get instant access and a $200 credit by signing up for an Azure free account.
Sign in to the Azure portal.
Explore the documentation and tutorials, and browse quickstarts and developer resources.
Customers using Azure Machine Learning
Jolie Vitale, Director of BI and Analytics, Carhartt"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."
Ignasi Paredes-Oliva, Lead Data Scientist, Nestlé Global Security Operations Center"MLOps is at the core of our product. Because of its reproducible ML pipelines, ... registered models, and automatic model scoring, we're definitely detecting things that we missed before. Which, in terms of risk management, is really, really important."
Joey Chua, Senior Manager Machine Learning Engineering, AGL"Azure Machine Learning allows us to manage the entire lifecycle, from experimentation and development to production and enhancements."
Daniel Engberg, Head of Data Analytics and Artificial Intelligence, Scandinavian Airlines"With model interpretability in Azure Machine Learning, we have a high degree of confidence that our machine learning model is generating meaningful and fair results."
Dean Riddlesden, Senior Data Scientist, Global Analytics, Walgreens Boots Alliance"If I have 200 models to train—I can just do this all at once. It can be farmed out to a huge compute cluster, and it can be done in minutes. So I'm not waiting for days."
Alex Mohelsky, Partner and Advisory Data, Analytic, and AI Leader, EY Canada"We see Azure Machine Learning and our partnership with Microsoft as critical to driving increased adoption and acceptance of AI from the regulators."
Azure Machine Learning updates, blogs, and announcements
Frequently asked questions about Azure Machine Learning
-
The service is generally available in several countries/regions, with more on the way.
-
The service-level agreement (SLA) for Azure Machine Learning is 99.9 percent uptime.
-
Azure Machine Learning studio is the top-level resource for Machine Learning. This capability provides a centralized place for data scientists and developers to work with all the artifacts for building, training, and deploying machine learning models.