Graph foundation models for relational data

Interesting representation. Not rows in the database as samples and columns as features, but a whole graph.

Makes training much more flexible, and fine tuning as well. Now, when a new data in terms of samples or new tables are connected to the olds ones you just extend the existing graph, without changing its existing morphology much.

Although it is unclear if it is scientific: "Look how cool we can do" or business result: "Look how much value do we get from this representation"