An Introduction to Knowledge Graphs

I couldn&#x27;t stress how important Wikidata (and its predecessor DBpedia) is as a public example of a huge knowledge graph (versus the ones hidden at big tech companies) but also as a Wikipedia-like collaborative project to organise all the knowledge among existing semantic web &#x2F; linked data publishers, government open data, libraries, galleries, archives...<p>Also remember that Wikidata is open source and you can fire up your own knowledge graph as docker containers on your laptop: <a href="https:&#x2F;&#x2F;wikiba.se&#x2F;" rel="nofollow">https:&#x2F;&#x2F;wikiba.se&#x2F;</a><p>If you have been disappointed by RDF-based technologies before, I would say Wikidata&#x2F;Wikibase have significantly innovated on top of them. For example, they allow each statement to have qualifiers, references, depreciation&#x2F;preferredness attached to them in a user-friendly way while also keeping simple queries simple.

Knowledge graphs for text (the focus of the article) seem narrowly-scoped since they require &quot;objective&quot; facts and relations to be practical. Capturing the subjective and transient perspective of observations made by multiple observers (which is what we actually have access to) is more complicated.<p>For example, asking the same person the same question may yield different answers based on their mood or other environmental or situational factors. Who&#x27;s asking the question can also matter, as does the specific phrasing of the question.

I wish all of you not to fall in the trap of ontologies. I worked very hard in this domain my conclusion is that all ontologies fail to scale eventually. I would recommend people in the field to go towards &quot;perspectivism&quot;.

On a side note, I love the idea of researchers writing “articles” in this format. No paywall, no complex two-column format, no PDFs. As a researcher myself, I wish this is what my “productivity” was judged upon, I’d probably have a lot more fun and motivation to work and produce!

Love to see that relationships (edges) are directional. I hate to see graph models where relationships are bi-directional as it loosens up data rules far too much with very little benefit.<p>I&#x27;ve worked with systems where the relationships are typed and can have attributes just like the vertices allowing the system to model data in a more intuitive fashion.

The definition seems faulty to me, since the pair (E: subset(N × N), f: E → L) does not admit of multiple edges with different labels, connecting the same ordered pair of nodes. Of course this is most often allowed in practical KG&#x27;s.

Slightly off-topic, does anyone recommend any open-source ANSI C implementations of knowledge graphs&#x2F;graph databases? I had toyed around a bit with Neo4J and am interested in something in a language I&#x27;m more familiar with.<p>I&#x27;ve created some useful output from Wikidata&#x27;s dataset with some-space saving decisions like focusing on certain languages and an arbitrary data structure. The dump is quite big and space at a premium.

So this is the ‘semantic web’ from ~15 years ago?

KG are cool, but I haven&#x27;t find a practical framework of combining simple logical predicates with temporal facts (things that are true at a certain moment in time) and information provenance (the truthiness of information given the origin). There might be ways to encode this information in a hyper graph but they are far from practical.

Is this really the semantic web debacle all over again? Curation doesn&#x27;t scale and if you&#x27;re going to do it just do it in a database already; standards lead to committees that end up choking progress because we have all the standards we really need already; NLP only pretends to understand written text but all it does really is tokenise badly. Just cache in a database already and move on!

SQL might be a good fit to model Knowledge Graphs, since FOREIGN KEYs can be named, using the CONSTRAINT constraint_name FOREIGN KEY … syntax. We thus have support to label edges.<p>Nodes = Tables<p>Edges = Foreign keys<p>Edge labels = Foreign key constraint names

I&#x27;m surprised there&#x27;s not many knowledge graph databases but there are graph databases eg Neo4J and there is GraphQL. It seems they are more popular