Introduction to Altair® Knowledge Studio® 2022.0

Hugh_R_Heinsohn

Knowledge Studio 2022.0 introduces several important improvements that make the software easier than ever to use and deploy.

• Docker is no longer required to support any Knowledge Studio functionality, which makes deployments easier, eliminates certain compliance issues, and saves money.
• The software now offers a new decision tree tab and we have made enhancements to the SHAP tab and the decision tree tab provides a visual and easy-to-understand way to explain black box models.
• A single standardized scoring node is now applicable to all native models, ARIMA Forecasting, Keras Deep Learning, Novelty & Outlier Detector, GLM, and XGB models; this improves usability, simplifies the process of building workflows, and removes clutter from the Action function palette.
• Users can also now specify custom Python and R execution environments in the Python and R Code nodes, which allows them to use any version of Python or R with Knowledge Studio.