Azure Machine Learning Roadmap

Azure Machine Learning allows customers to execute mathematical algorithms on data to draw conclusions based on known results in similar data and to categorize data. This predictive analytics process, called machine learning (ML), can be used for purposes such as determining product recommendations, equipment maintenance requirements, and financial credit scores. It can also be used as the basis for generative AI, which drives applications such as chat agents that communicate with users in natural language.

Customers can use Azure Machine Learning to develop and host models (ML configurations trained from existing data to produce results given new data) or by using models developed elsewhere, such as Azure AI Services.

Product: Machine Learning

Upcoming changes to the service in the near and long term include:

Near Term (2024–2025)

Portions of Azure Machine Learning that are in preview include:

Administration. Signals from the model monitoring feature, which helps administrators monitor production ML models by collecting inference data from deployed endpoints can help administrators determine whether a model should be retrained for better performance and accuracy. The feature observes signals from the data, such as those that can measure data, prediction, and feature attribution drift and data quality. Alerts can be generated from the monitored results, and historical collected data can be reviewed. Another administrative feature allows workspaces to be moved across Azure subscriptions