Graph Neural Networks + Semantic Web

Hello everyone! I hope you are all doing well.

I'm currently in my Ph.D. which has as one of the main areas the
application of Graph Neural Networks (GNN) to the problem of Tax Evasion.
Thus, I've created a Knowledge Graph with data about all the companies of
my state, their owners, and relationships between them. Since all the data
used to create such a KG are public, it is actually many small graphs
instead of a big heavily connected graph. Even so, I was able to get good
results.

Since my master's thesis was related to Semantic Web, I was wondering if I
can use the Semantic Web concepts to improve my research. Thus, I've
started a search for papers that uses GNN combined with Semantic Web
concepts/technologies. I've found some, for instance:
https://grlearning.github.io/papers/7.pdf
https://www.sciencedirect.com/science/article/pii/S2352711019302626
https://alammehwish.github.io/dl4kg-eswc/papers/paper%202.pdf

So, the goal of this e-mail is to start a discussion related to open
problems and challenges which the combination of GNN and Semantic Web may
solve. Any ideas?

Thanks in advance!
--
Otávio