STU-SV120-XPFv4.1.0-PV910 Release Notes

Author: Matthew Lindon, Created: 2025-12-30

Release Date

December 30, 2025

Enabled Browser UX Support.
Increased visibility of elements within the AI Services Log.
Changed Routine Run filter buttons to behave like an OR condition instead of an AND condition.
Reorganized the Runs page to improve visibility of elements.
Improved page rendering speeds.

Fixed an issue where Method Run descriptions may not format properly.
Fixed an issue where user-selected memory capacity wasn’t being used when creating a Routine Instance.
Fixed an issue where attempting to create an Instance or Run with invalid parameters could take up the Instance or Run name for subsequent attempts.

XBR-SV210-XPFv4.1.0-PV910

Xperiflow v4.1.0

If multiple users are in the SensibleAI Studio at the same time, Routine Instances and Routine Runs may be assigned to the wrong user.
If a memory override is set when creating a specific routine instance, it does not carry forward as the default for subsequent method runs. Instead, the method's own default memory capacity will be set as the default value but can be overridden separately.

In v3.0.0 and v4.0.0, a one-hot encoding related error may occur when a user selects 0 features with categorical data alongside other features. The issue may not show if the user selects at least 1 categorical feature with at least 3 distinct values.

When the Create Web App method is invoked via Direct (JSON) Invocation for an ML Regression (v2.0.0 only) or ML Classification (v4.0.0 only) routine instance, the Runs section of the Web Dashboard Artifact fails to load.
Timeouts may occur when no models are selected. All models are compared when no models are selected. This is known to occur for 15k+ rows for the ML Regression Routine and for 2M+ rows for the ML Classification Routine.
For the ML Classification and ML Regression web dashboards, artifacts may not load for datasets over 2M rows.