The link below is a web view of the in-progress conceptual model for threat and risk information sharing illustrating the semantic federation use case. This is a UML representation of the conceptual model. I would note that the UML representation has some compromises but is effective for the purpose. The model is not complete but is progressing.
We then have mapping models that show how data schema represent concepts in this model.
One thing that is made very clear by this experience is that the representation relationships are frequently not 1:1, this requires that the concepts be more “semantic” and we understand the composition of the concepts. There can be patterns on either the conceptual side or the data side. This is why just “pointing at a definition” is not sufficient, you really need to understand the concepts, including their relations, context and parts.