Are Knowledge Graphs Just Robust RDFs?

Whether you know it or not, you likely interact with a knowledge graph on a weekly if not daily basis. When you ask your favorite voice assistants like Siri, Alexa, Cortana or others to field your questions, you are relying on underlying knowledge graphs to contextualize your questions and link to a breadth of structured and unstructured data to provide robust and comprehensive answers.

Not a Typical Graph

So, what is a knowledge graph? A knowledge graph is a set of interlinked data that models the knowledge of a domain in a programmatic way, with the interlinked data augmented by subject matter experts and machine learning algorithms. A knowledge graph is generally built on top of databases to connect the contained data together in a meaningful way. A key characteristic of knowledge graphs is that they have formal semantics that can be used both for interpretation and inference. The formal semantics are like a contract between the knowledge graph developer and the user, whether that is a person or an application, regarding the meaning of the data in the knowledge graph.

More Robust Than RDFs

Resource Description Framework is the standard model for exchanging data on the worldwide web. RDF already supports the merging of disparate data and facilitates schema evolution without requiring changes from anyone consuming data from an evolving schema. Every RDF graph, however, is not a knowledge graph nor is every knowledge base a knowledge graph. In order to be a knowledge graph, the graph must have the key characteristics of interlinked data with formal structure and semantics. Knowledge graphs are typically used for information-dense services such as content recommendation that is contextually aware, semantic searches, regulatory document information discovery and advanced drug safety analytics.

Enterprise Knowledge Graphs

Huge technology companies like Google, Facebook and Amazon have spent millions of dollars creating robust and comprehensive knowledge graphs. The connection of disparate data in meaningful ways is very strategic for these companies. When you search for something on Google, you are interacting with its knowledge graph as it takes your searches, breaks the search strings down and provides answers by fetching the correct answer for what it understands you to be asking. Companies don’t have to be technology giants to benefit from knowledge graphs. All organizations can benefit from combining siloed data and connecting structured and unstructured data. Not only does this give companies a leg up on employee enablement, but it makes them more competitive by allowing them to get to deeper insights more quickly.

As the data intelligence industry continues to move away from the concept of big data and demand that all data must be smart data, knowledge graphs become more essential. Machine learning and artificial intelligence has amazing potential to unlock insights deep within your data, but that can’t happen if they are bound by inflexible schemas. Knowledge graphs provide the necessary background and concept awareness to interpret data, extract key facts and then further enhance the knowledge graph, promoting a positive cycle of valuable analysis that drives results plus data enrichment.

If you have any questions, please ask below!