OML - Oracle Machine Learning

The advent of AI and ML is near and fast approaching. Offerings such as ChatGPT are changing the way we go about doing research and just ways of working in general. To this end, it is worth noting that great strides have already been made relative to data science and analysis, and that Oracle has a strong offering that allows descriptive and predictive analytics to be performed where your data resides, which can be different from Generative AI, like ChatGPT, but just as powerful based on context.

If you have an autonomous database in Oracle Cloud Infrastructure, then you already have default access to the Oracle Machine Learning features within it. After performing a few steps from the OCI Admin console within your autonomous database area, you can access the Oracle Machine Learning Home Page, where you can create a Workspace, which encapsulates your ML projects, and you can also create notebooks, which is an interface for the creation of complex queries using SQL. There's also full permission sets where you can grant viewer, developer or manager to users for your workspace and child attributes.

The below graphic from Oracle showcases the features of the OML offering inside the ADW/ATP:

The below graphic from Oracle shows how OML is "different" from generic ML:

So why is this important? This offering gives you the ability to do data discovery and analytics on top of your data, without having to move it out of the database which can create performance issues, security issues, etc. Using the below features of the OML engine, you can find hidden patterns in your data that can both benefit your operations and processes, bottom line, and more.

You can also use Python and R via the OML interfaces provided if you are more comfortable with that toolset, versus PLSQL. Additionally, there's API support for OML itself, so that a tool like Postman can be used versus the web UI, the Git Hub project with Oracle provided collections can be found at: https://github.com/oracle-samples/oracle-db-examples/tree/main/machine-learning/oml-services/postman-collection-examples

The below graphics from Oracle showcase the capabilities of the OML service, in terms of what algorithms and functionality are provided:

Overall, for an embedded service of the autonomous database at no additional cost, this is a great feature that allows several algorithms to be leveraged out of the box, plus the capability to build your own, and definitely worth checking out further! Oracle has a simple course over at Oracle University and there's even a certification for it and can be found at: Become an Associate on Oracle Machine Learning with Autonomous Database - Oracle MyLearn