20th December 2023 (Europe 1) - Developer Documentation
Skip to content

20th December 2023 (Europe 1)

Scope

We have worked hard to deliver a great Insights Hub experience. These release notes are relevant for Insights Hub Release in December 2023 and include new and adjusted features, fixed issues. If you find other issues, please give us your feedback by contacting our Support.

Predictive Learning Essentials

Applicable for:

  • Capability Packages Subscribers

What's new?

  • With this release,
    • Predictive Learning essentials app will be offered to standard capability package tenants.
    • We are introducing Quick bootup environments for job executions.
      • Users can launch their jobs on newly introduced quick boot up environments.
      • Quick boot up environments are faster in launching the jobs as compared to legacy moderate boot up environments.
      • User's can launch 'on-demand' and 'schedule based' jobs with the newly introduced quick boot up environments.
    • We are introducing, new templates for environment configurations with multiple instance types
      • In Predictive learning essentials user can choose the relevant instance types which would be needed to run users model/script in an on demand or schedule job.
      • User can choose the right instance type as per the CPU and Memory requirements of their model and data under processing.
      • A wide variety of instances is provided from the minimum of '1 CPU and 2 GB Memory' to maximum of '16 CPU and 32 GB Memory' is provided.
      • Both Jupyter and Docker job execution environments have quick bootup and moderate bootup type of instances. User can chose an appropriate combination of environment configuration template and instance types as per their need.
    • We have improved model executions such as:
      • user can debug model/script execution by downloading the execution logs once the job execution is finished.
      • This release introduces granular level schedule windows e.g. 5 mins, 10 mins, 15 mins & 30 mins.
    • We have improved the legacy standalone Jupyter notebook is replaced with Jupyter Lab software for model validation.

Advance information

  • The two legacy applications which were specifically for tenants on old P&P will be replaced by the latest “Predictive Learning Essentials” application with all the new features and new user interface and user experience in the upcoming release.
  • For the users of “Analytics Workspace” and “EMR” in legacy application, you could have same but even better features via newly integrated Jupyter Lab environment and more options on the model validation or execution environment configuration via the latest “Predictive Learning Essentials”.

Known issues

  • The model/script execution logs are not available for the jobs running on moderate boot up type of environments.

Predict

Predict (AI for Everyone)

Applicable for:

  • MindAccess IoT Value Plan Subscribers
  • MindAccess Developer Plan Subscribers
  • MindAccess Operator Plan Subscribers
  • Capability Packages Subscribers

What's new?

  • With this release,:
    • we have enabled Y axis zoom for the charts, which helps in visualizing kpi values having minor variations.
    • new model creation in Anomaly Detection now supports more advanced configurations.
    • it is possible to enable fraction increments for sensitivity in Anomaly Detection.
    • Forecast Prediction now supports data imputation.

Last update: January 29, 2024

Except where otherwise noted, content on this site is licensed under the Development License Agreement.