Complete your Databricks User Groups profile!

Fill out a few details about yourself so the community can get to know you.
Genie Community

Exploring the Potential of Graph-Based Semantic Models

Summary: mywm4v shares insights on the growing importance of semantic models in AI-powered analytics, particularly discussing the potential for integrating graph-based semantic models to enhance data relationships and insights. They are interested in community feedback on experiences with these models, their impact on AI frameworks like Genie, and any related challenges or advantages. grazia responds by suggesting Ontos, an open-source framework from Databricks Lab, which features a Semantic Model layer linking technical assets to business concepts through a knowledge graph.
AI Summary

As we continue to deepen our understanding of AI-powered analytics and natural language processing, the role of semantic models becomes increasingly crucial for transforming data interactions. Currently, many are pondering the potential benefits of incorporating or supporting a graph-based semantic model within our projects. This model could potentially enhance the way data relationships are structured and insights are derived. E.g. direct integration of metaphactory or any other OSI supporting tool iwhtin Databricks.

We'd love to hear your thoughts on this topic: What are your experiences or insights regarding graph-based semantic models? Do you think they present a significant opportunity for refining and improving data relationships in AI frameworks like Genie? Are there specific challenges or advantages you've encountered in graph-based systems that you believe could be applicable here?

1 comment

Hi Matthias,

Have you looked into Ontos which is an open source framework developed by Databricks Lab?

https://marketplace.databricks.com/details/8c582cfa-4c6b-4fdc-93a2-f63b35d93906/Databricks_Ontos

Ontos contains a built-in Semantic Model layer that explicitly links physical technical assets to business concepts via a unified knowledge graph.