Alternatives to Cognitive Services

Cognitive Services does not encapsulate all of Microsoft’s machine learning–related offerings, just those provided by a particular group within Microsoft. Other Microsoft machine learning options that developers use generally fall into two major categories:

Full-featured offerings, such as Azure Machine Learning, Databricks, HDInsight, SQL Server, and ML.NET support machine learning capabilities for professional data engineers and scientists. They use a variety of machine learning languages like Python, Spark, and R to enable development of sophisticated machine learning models and scripts to address any requirement, whether simple or complex. As such, they generally require more effort to deploy and use than Cognitive Services but can enable more targeted solutions.

Self-service offerings, such as Power BI, AI Builder, and Excel, provide machine learning abilities for professional and nonprofessional users. They include basic predefined models (similar to some in Cognitive Services) that users can modify and customize to meet their needs and train with their own data. The models typically run within the product in which they were created, but some can export or publish their models to other Microsoft offerings to be part of a larger application.