See the newly published lessons in the "SORBA.ai Expert Webinars" Course
Webinars
In the October release of SORBA, we've added several exciting features. Machine Learning now supports custom algorithms and virtual environments. New regression and digital twin algorithms enhance modeling. Two-Factor Authentication (2FA) adds security to SORBA Identity. In IoT Unified, we've added WhatsApp as a notification channel and introduced flexible dashboard reports based on tag triggers. You can now select execution modes (Edge, Cloud, or both) for various nodes. We've expanded IoT Connectors and introduced new data processors. Python Script Editor improvements, a Vision Node for video analytics, cross-referencing, tag diagnostics, and enhanced alarm acknowledgment provide enhanced functionality. SORBA Vision is now in beta, offering features like object and motion detection, video analytics, and more, with further updates expected.
This update allows users to create and import custom machine learning algorithms beyond the suite of algorithms already present within the Machine Learning Trainer.
Ability to add additional virtual environments to the Machine Learning Trainer.
Additional algorithms have been added to the regression and algorithm machine learning model types, including Auto DNN, Bagging Regressor, and Random Forest Regressor.
Users now have the ability to set up two factor authentication with external applications in order to add additional security to their SORBA instances.
Whatsapp has been added as an additional option for notification channels.
Users can now choose to generate dashboard reports based on trigger tags instead of a set time schedule.
For Alarms, Model Instances, Scripts, and IOT Connectors, users can now select an execution mode, determining if these nodes will be run at an edge level, cloud level, or both.
New IOT Connectors have been added, including KCF, OPC UA, and Ignition.
New data processors have been added to the preprocessing, statistical, and tag filter categories. These include MaxAbs, MinMax, Robust, Standard (pre-processing), moving skewness, moving quantile, moving kurtosis (statistical), thermodynamics and new options for the vibration processor (tag filter).
A new option has been added to auto-learning. Instead of deleting or disabling old models, users can have a single model that is updated through auto-learning.
Multiple improvements have been made to the Script Engine
The SORBA vision node has been added to the Workspace (further notes on Vision below).
The new cross reference feature allows users to see where different nodes are being referenced in relation to one another.
Tag Diagnostics have been added to the Data View to check the quality of tags and to check what sources are writing to a tag.
Users can now add notes when acknowledging alarms.
SORBA Vision is now in a beta testing stage. This feature gives users access to tools that can be used to monitor and analyze video data. These tools include object detection, motion detection, and analytics that can be used based on the data collected with SORBA vision. These features include the ability to detect object duration, perform object counts, check for camera quality, and detect and differentiate between in-motion and motionless states. More information will be made available regarding this feature as it moves out of the beta testing phase.
