Read about the knowledge graph and about how many enterprises are already embracing the idea of benefiting from it.
Realize that everything connects to everything else.
Leonardo da Vinci
The poet Simonides was at a banquet hosted by a Thessalian nobleman whom Simonides described as ‘as too ignorant to be beguiled by poetry’. During the festivities, Simonides was told there were two messengers waiting for him outside. He stepped out into the warm night, but the streets were empty. Just as he went to look around the corner, an earthquake knocked him to the ground.
When it was over, he saw that the palace where he was dining had collapsed and he had escaped certain death. As they cleared the rubble of the banquet hall, the guests had been crushed beyond recognition. As the sole survivor, Simonides was asked to identify those who had been killed. He could do so by matching their identities to the place they were sitting at the table.
Thus, the first knowledge graph was invented! Actually, this was called the ‘memory palace’ technique, and the idea of matching information with context still enables humans to memorize an astounding amount of information. The current record for the number of playing cards memorized was achieved by Munkhshur Narmandakhat at the 2017 World Memory Championship. In one hour, he memorized 1924 playing cards. That’s 37 full decks worth!
The ‘memory palace’ mnemonic technique is a hack for something our brains do naturally. We do not hold memories in isolation. They are in the context of other memories and knowledge. It’s also the secret for why knowledge graph databases are so powerful and work intuitively with how we, as humans, store and use information. But what is a knowledge graph?
Every bit of information is related to every other bit. Think of how social media organizes our connections. There is your profile and it is connected to other people’s profiles who are in turn connected to some of the same people that you are, but also others. That social network is an example of a graph. And guess what, every social media platform (e.g., Twitter, Facebook, Google, etc.) stores their information in a knowledge graph. Why would they do it otherwise?
Traditional relational databases only have information. Graphs have information and relationships, the connections between that information. Like the ill-fated dining companions of Simonides who had their identities and their position around the table in relation to each other. For example, at the head of the table was the flattened nobleman Scopas and to his left was his father Aleuas. ‘To the Left of’ and ‘To the Right of’ are relationships that can be represented and most importantly reasoned about.
Graphs treat the connections between information as equally important, a first-class citizen in the world of data. Click To TweetAs new information is put into a knowledge graph, it is connected to the current information, giving it its context. That context enables you to easily traverse the connections. The internet is old enough now that we don’t have to explain the value of links and traversing those links as we consume information.
So far, we’ve talked about information. But, where does the knowledge part of the knowledge graph come from? By representing relationships between information, it is possible to encode the meaning, also known as the semantics, of the information. This is done by describing the data in terms of an ontology or schema using the same language that is used to describe the data and their relationships. Let’s give an example.
A Person is to the left of a Person.
A Person is to the right of a Person.
A Poet is a Person.
Scopas is a Person.
Aleuas is a Person.
Simonides is a Poet.
Scopas is to the left of Aleuas.
That’s one of the amazing properties of a database based on a knowledge graph versus traditional relational databases. A knowledge graph not only gives you the data but also provides how the computer is to interpret and understand that data. Click To Tweet It is hard to overstate just how empowering that is for creating incredibly sophisticated applications and services.
Look back at the English sentences used for the example of a knowledge graph in the previous section. Those statements are ridiculously close to the RDF language used by Ontotext’s GraphDB. It would be trivial for even a beginner to translate those English statements to a form that a computer could understand and reason about.
Knowledge graphs represent information that is similar to how humans understand information. When querying enormous knowledge graphs (e.g., every time you use google, you use a knowledge graph) you do so in a language near to natural language using humanly-understandable terms and not just computer code. This enables a more intuitive search and discovery, narrowing the gap between human and machine language.
Knowledge graphs are graphs in the mathematical sense. ‘The Seven Bridges of Königsberg’ by Leonhard Euler, considered the first paper in the history of graph theory, was published in 1736. It means that by using a knowledge graph database like GraphDB you can take advantage of nearly 300 years of research for optimizing search as well as traversing and analyzing your data.
Having your schema, the read instructions for your data, described alongside your data, you unlock one of the most interesting properties of a knowledge graph: flexibility and dynamism. One of the biggest problems with relational databases is their rigidity.
Organizations grow and change at a rate inconceivable when the first relational databases were developed. A knowledge graph expands by adding relationships and nodes. If the information is no longer relevant or correct, the relationships and the information is deleted without worrying about the unseen consequences. A knowledge base can change as the world and your organization’s understanding of that world changes.
Additionally, knowledge graphs were made for an internet world made of constantly changing and heterogenous data. It is a perfect model for capturing provenance for data. A knowledge graph will be able to tell you what it knows, how it knows it and why.
What’s not to love?