I’m happy to share the ideas behind the federated knowledge graph in the Actian Data Intelligence Platform in this short explainer - and some details below:
The idea behind the federation of the knowledge graph is that all domains in an organization get their own unique graph, to structure and express the data in their domain, at the metadata level. Only when the domains are ready, will they share the data on the enterprise data marketplace, which is the central piece of the graph, that ties everything together.
This typically makes data architectures scale, as the federated knowledge graph absorbs domains without friction (they can define themselves) and makes end users join without operating in a user interface for enterprise wide data governance.
This architecture was awarded Bronze at La nuit de la data et de l’IA (Data and AI night), in Paris in 2024. What the data community had not understood 18 months ago, but that most are realizing now, is that the ontologies contained in the federated knowledge graph contains an unparalleled semantic richness to provide context for generative AI. But more people are getting it now, we sense it.
More on that soon.
