The following describes the new release highlights for users of Infor OS Artificial Intelligence (Coleman AI Platform):
- Optimization for Real Time & Batch Prescriptive Use Cases: New Optimization module is intended to provide most effective solution for business problems. Powered by multiple solvers to compare results and select the most effective one for production deployment. This provides benefits to users by intuitively building mathematical models in a quest by setting variables, constraints, objective and relying on data collection feature. Real-time use cases can be deployed as endpoints, and for data intensive problems batch production is available.
Click to enlarge
- Auto-clustering algorithm for determining optimal number of clusters: Introduced as intellectual property of Infor's Augmented Intelligence suite of offerings, auto-clustering provides solutions by using a combination of K-means, K-modes, and K-prototypes automatically based on features. Allows automatic handling of all numerical, categorical, and/or mixed inputs. The model can be configured with minimal inputs to get auto-clustering outcomes while allowing customized, tailored clustering settings.
Click to enlarge
- Ability to download model insights from Custom Algorithm: The feature provides the possibility to extract model artifacts, such as coefficients, statistics, feature importance, and similar, in the form of a downloadable file for models trained with custom algorithms. The file is entirely dependent on the custom algorithm code, so users can include code within their custom algorithm to generate file and define its contents. Extracting insights from the models is valuable for debugging, informing feature engineering and importance, and building trust in AI outputs.
Click to enlarge
- Python Libraries Updates: The pre-installed Python libraries that are available for scripting & custom algorithm will be updated. Scripting is available as a part of Prepare Data activities to execute a customized Python script to perform an activity, which is not available in the catalog, whereas Custom Algorithm enables users to bring their own algorithm if the section of built-in machine learning algorithms does not meet requirements.?
Click to enlarge
To best prepare customers for the October 2023 Feature Release, Infor has hosted a series of webinars. Those webinars have included both the CloudSuite Release Highlights sessions, as well as an enablement session aimed at preparing ALL customers for consuming Infor's MT Feature Releases. We hope that you and your teams have been able to attend the sessions live, but if you couldn't, or if you just want to go back and re-watch these sessions, the recordings available in the KB article.
View KB 2309228 - Preparing our MT Cloud Customers for the October 2023 Feature Release (infor.com).
A full list of new features, bugs fixed, and other documentation can be found on the Infor Support Portal. Learn more about Infor OS here or contact us anytime.
Don't see a feature you want? Log into the Infor Support Portal and click on Resources > Enhancement Requests. You can enter your own and vote for others.
For more information on product updates, please see the release training that is available to customers on Infor Campus.
