What's New

1 What's New

Provides a summary of the latest enhancements and features for Oracle Machine Learning User Interface on Oracle Autonomous Database.

Table 1-1 New Features

Features	Description
Support for email notifications	Oracle Machine Learning User Interface has been enhanced with the functionality to send email notifications for jobs. Jobs allow you to schedule the running of notebooks. You can now send email notifications to the specified users' email addresses about the selected events for the job. See Create Jobs to Schedule Notebooks for more information on how to send email notifications.
Support for NVIDIA GPU compute capabilities in OML Notebooks	Oracle Machine Learning Notebooks offers support for NVIDIA GPUs (Graphics Processing Unit) compute capabilities. With the new GPU capabilities in Oracle Machine Learning Notebooks, you can run advanced machine learning algorithms such as deep learning models, transformers (embedding models) for generating vectors, and small LLMs. The GPU feature is enabled for Oracle Autonomous Data Warehouse Serverless or Oracle Autonomous Transaction Processing Serverless instances with 16 or more ECPUs specified for the OML application. Both the Licensed and the Bring Your Own Licence (BYOL) versions are available. For cost details, refer to the Oracle PaaS and IaaS Universal Credits Service Descriptions document available on the Oracle Cloud Services contracts page. Note: GPU resources are available only on paid Oracle Autonomous Database Serverless instances. GPU resources are not available on Always Free Oracle Autonomous Database Serverless or Oracle Autonomous Database Serverless instances with fewer than 16 ECPUs allocated.
Oracle Machine Learning Notebooks Classic deprecated	Oracle Machine Learning Notebooks Classic has been deprecated since June 11, 2024. On October 29, 2024, you will no longer be able to create Classic notebooks, save them as templates, or select them for job scheduling. Existing Classic notebooks can be opened in read-only mode. You can continue to convert Classic notebooks to the new format using the Copy to OML Notebooks button on the Notebook Classic listing page. On December 31, 2024, Classic notebooks will no longer be available. The ADMIN user can access Classic notebooks in read-only mode and convert them to the new format. Jobs that still use Classic notebooks will show a status of `Disabled`. Associated job logs will not be accessible. On June 4, 2025, the ADMIN user will no longer have access to Classic notebooks, and any remaining notebooks will be deleted. If you have Classic notebooks or Classic template notebooks (personal or shared) that you wish to keep, you must convert these to the new format. If you have jobs that rely on Classic notebooks, these jobs must either be updated with a new notebook or recreated with a new notebook.
Oracle Machine Learning Notebooks update	Oracle Machine Learning User Interface offers an enhanced notebook environment. Initially released as Notebooks EA (Early Adopter) in Oracle Autonomous Database Serverless, it is now accessed using Notebooks under the left navigation menu and home page. The enhanced notebook interface supports SQL, SQL Script, R, Python, Conda, and Markdown interpreters. You can write code, text, create rich visualizations, and perform data analytics including machine learning in the enhanced notebooks. Note: The original Zeppelin-based notebook interface is still available for a limited time under the left navigation menu item Notebooks Classic.
Support for model monitoring in Oracle Machine Learning User Interface	Oracle Machine Learning User Interface offers support for model monitoring. It allows you to create model monitors. The model monitors enable you to monitor the quality of model predictions over time, and provides you with insights on the underlying causes.
Support for data monitoring in Oracle Machine Learning User Interface	Oracle Machine Learning User Interface offers support for data monitoring. It allows you to monitor your data and evaluate how your data evolves over time. It helps you with insights on trends and multivariate dependencies in the data. It also provides you an early warning about data drift.
Support for enhanced notebooks in Autonomous Database - Serverless	Oracle Machine Learning User Interface offers a new enhanced notebook environment Notebooks EA (Early Adopter) in Autonomous Database - Serverless. The enhanced notebook supports SQL, SQL Script, R, Python, Conda, and Markdown interpreters. You can write code, text, create rich visualizations, and perform data analytics including machine learning in the enhanced notebooks. Note: The enhanced notebook is available in the Oracle Machine Learning Notebook Early Adopter release. During the Early Adopter release period, both Zeppelin and the enhanced notebooks will be available, after which all notebooks will be converted to the new notebook environment. During the Early Adopter phase, you can use both the original Zeppelin and new Early Adopter notebook interfaces. Notebooks in the original interface can be copied to the Early Adopter release. The enhanced notebook interface in Oracle Autonomous Database Serverless provides the following enhanced features and user experiences: Rich and enhanced user experience: The enhanced notebook offers modern look and feel, and richer visualization with many charting options. This will benefit users to better visualize and understand their data. In addition, it offers some useful features like side-by-side versions comparison, option to add comments to paragraphs, full screen size mode for paragraphs, option to define paragraph dependency, and so on. High availability: The enhanced notebook, a multi-tenant application is deployed to the same middle-tier as Oracle Machine Learning server, and this requires no additional resources. Therefore, it is always running and readily available to render the new enhanced notebooks. High scalability: The enhanced notebook assures high scalability in production. To scale up due to increased user demands, additional notebook instances can be easily added. There are tools to monitor system loads, and if a system is consistently overloaded, additional instance can be easily added to mitigate risks related to scalability.
Support for Python and R third-party libraries	Third-party libraries for Python and R are available on Oracle Machine Learning Notebooks. Oracle Machine Learning UI provides the Conda interpreter to install third-party Python and R libraries inside a notebook session. Conda is an open-source package and environment management system that enables the use of environments containing third-party Python and R libraries. Users with `OML_SYS_ADMIN` role can install Python and R third-party libraries and upload them to object storage for persistence. The user with `OML_SYS_ADMIN` role is the administrator, also known as the admin. Users with `OML_DEVELOPER` role can use the Conda interpreter to download and activate the third-party libraries using the Conda environment that are provisioned by the administrator. The user with `OML_DEVELOPER` role is the regular Oracle Machine Learning user.
Support for R	Oracle Machine Learning for R is supported within Oracle Machine Learning Notebooks. By using Oracle Machine Learning for R, you can perform data exploration and machine learning modeling. OML4R is available through Oracle Machine Learning Notebooks on Oracle Autonomous Database Serverless, including Autonomous Data Warehouse , Autonomous Transaction Processing and Oracle Autonomous JSON Database services.
Support for cross-region Autonomous Data Guard	Oracle Machine Learning Notebooks provide cross-region Autonomous Data Guard support in newly provisioned and migrated databases.
Oracle Machine Learning repository migrated from Serverless database to each respective Oracle Autonomous Database instance	The Oracle Machine Learning (OML) repository has been migrated from Serverless database to each respective Oracle Autonomous Database instance. The migration of the Oracle Machine Learning repository ensures: That all OML objects such as tables, jobs, stored procedures, and metadata are moved to the appropriate Oracle Autonomous Database instance. Provides support for Refreshable Clones, which enables cloning of the Oracle Machine Learning metadata as well. Note: The migration of the Oracle Machine Learning (OML) repository is expected to be completed over a period of 30 days. The OML repository version is mentioned in About in the `<user>` drop-down list on the top right corner of your Oracle Machine Learning User Interface page. If the version is `1.0.0.0.0`, it indicates that the OML metadata is still in the Serverless database. If the version is `22.x`, it indicates that the OML repository has been migrated to your Oracle Autonomous Database instance.
Oracle Machine Learning Notebook supported on all Oracle Autonomous Database clones	Oracle Machine Learning Notebook is supported on all types of Oracle Autonomous Database - Serverless clones, including: Full Clone: a new database is created with the data in the source database and metadata. Refreshable Clone: a read-only full clone is created that can be easily refreshed with the data from the source database Metadata Clone: a new database is created that includes all of the source database schema metadata, but not the source database data. Note: For a metadata clone, the Example Template notebooks are not supported.