Knowledge Graphs

Knowledge graphs are used to describe people, things, ideas and anything else you can imagine. They capture knowledge in a way that both humans and computers can understand.


What are they?

For humans it is intuitive to communicate using graphs. There are labels/terms we use for ideas and things that we agree on, then we share new ideas by linking them to these known terms.

Knowledge graphs are a digital representation of this mechanism. They define terms (classes) to describe things, and properties for classes that store data or links to other classes.

Sometimes graphs are nicely structured with orderly rows and columns. Other times they are jagged and incomplete, because that's all the data you have. This is not a problem, and is naturally how our brains work when we're trying to understand something new.

The use of knowledge graphs by computers is quite exciting because they are reliable, flexible, and able to scale to massive volumes as Google, Facebook and others have proven. The use of knowledge graphs is exploding around the world, and they are being used to support software and AI by all the big tech companies.


What are the potential benefits?

Smarter data, open standards, AI fluency and automated reasoning

  • Without question, the best way to capture metadata (data about data)
  • Based on widely used, web-centric, open standards like RDF, OWL and JSON-LD
  • Provides a common vocabulary for humans and computers, that can be enhanced by a rich language of logical assertions
  • Designed to support efficient and flexible automated reasoning

How can I use them?

There are 2 ways you can use any graph/vocabulary (custom or public)

  1. Reference them
  2. Automate with them

Referencing them is free. It just means you are aware of them, and that data fields you use can be mapped to them where there is overlap. For example, you can manually fill a custom pick list in forms with codes you copy from a public graph.

Automating with them is even better. It makes you digitally compatible with other members of your community/network, which can lead to new opportunities. And it offers quality and productivity benefits similar to smart content. Some automation is free and some requires customization as discussed below.

The free Community subscription provides access to some automated features. These include exporting SEO scripts using schema.org (to inform search engines) and exporting Open Graph metadata (to inform social networks). These are discussed further here.

The reason the rest of this section is located within Enterprise features, is these features currently require a certain amount of customization.

In particular, Tag includes several popular graph libraries including Apache Jena and OWL-API. These are not yet exposed in the general use tool, but can be accessed via customization.

There is also a visual knowledge graph editor under development that will become a general use Tag app.

Public vocabulary integration

The support page includes a list of Public Resources which list several well-known public knowledge graphs. These graphs contain information that drive business process (e.g., health billing codes, financial codes, other professional industry standards) and improve global communication.

Tag can integrate parts of these vocabularies into smart content templates and forms. For example, instead of generating *.docx documents, you can generate XML fragments described by public graphs (e.g., FHIR resources).

Custom knowledge graphs

Tag can also store custom knowledge graphs created in other tools (e.g., Protege) in a persistent Jena triplestore (local or network based). This is useful when working with large graphs.

We can also be helpful when creating custom graphs, which are sometimes created by converting other documents or models.


How do I learn more?

Contact us for more information about any of the above features.