2.2.0 release notes
This release of AI-Link expands existing functionality to improve interaction with GBQ and Databricks database backends and enhance the user experience for first time users.
Please refer to our API documentation for the latest syntax to use with AI-Link.
Auto Semantic Model Creation (Beta) Updates:
- Customers will now be able to use Auto Semantic Model Creation to generate new semantic layers from data hosted in Snowflake, GBQ, and Databricks.
- Customers will now have the option to use this capability with tables in local files (e.g. .csv) or directly from the source database. This flexibility offers users the opportunity to create semantic models from tables they have immediate access to. Our 2.2.0 update to include creating semantic models directly from the source db helps to eliminate potential table duplication within the databases, avoiding governance issues as it removes the potential of different table versions existing in a source database. These changes make it much easier to get started creating semantic models that are based on a centralized, managed source of truth. Please refer to Auto Semantic Model Creation documentation for more information on how to best use this capability.
Semantic Inference (Beta) Updates:
- UDF Support Expansion: Semantic Inference will now support GBQ and Snowflake (Snowpark) UDFs so users can manage and maintain workloads in their organization's desired data warehouse
- Non-UDF Logistic Regression: Semantic Inference (non-UDF) will now allow users to embed a logistic regression model (often used for classification problems) directly as a measure within the AtScale data model.
Non-Functional Updates:
- Improved performance for interaction with complex data models
- New writing and testing support for GBQ and Databricks data warehouse backends
- Improved documentation and lab environment for new users: Additional, prescriptive details to get started and improved lab environment flow for both Semantic Predictions and Auto Semantic Model Creation