Learn about the knowledge graph and how it tells you what it knows, how it knows it and why.
You might have been hearing the phrase “knowledge graph” everywhere for quite a while. It’s now on the minds of business leaders and executives from a wide range of industries seeking to put their data to good use. But what exactly is a knowledge graph and why are so many enterprises embracing the idea of having one?
Let’s begin with the ‘what’.
Back in 1658, the Czech educator John Amos Comenius published a textbook for children, called Orbis Pictus, or Orbis Sensualium Pictus (Visible World in Pictures). The textbook was revolutionary for its time: it contained images, signs, words, artifacts, things and creatures – all presented through one or more rich networks of meaning related to them.
On an abstract level, and skipping a lot of technical details, to a machine trying to make sense of something (a data object), a knowledge graph is like the Orbis Pictus to a child who discovers the workings of the world. The reason these two ways of representing things are so similar is that their aim is to represent a thing through its connections and together with interrelated pieces of information about its ecosystem.
Both Orbis Pictus and a knowledge graph serve as tools for richly describing things, people and concepts, relying on the relationships they enter with the rest of the world. But Orbis Pictus uses metaphors and narratives to explain what something is to a child. While a knowledge graph is all about formalized descriptions of real-world objects and hyperconnectivity on a data level that serve machines to make sense of information input.
In more detail, the knowledge graph is a collection of interlinked descriptions of entities – real-world objects, events, situations or abstract concepts. Most importantly, it is an information structure built with the purpose of serving software agents and systems as a reference.
Which brings us to the ‘why’ part of our quest for understanding the knowledge graph: why are so many enterprises embracing the idea of having it.
Today, there is a knowledge graph technology underpinning the operations of virtual assistants such as Siri, Google Assistant as well as Alexa and Cortana. Companies like BBC, Elsevier and Roche use knowledge graphs to power their research and development activities (Roche), their operations with content (BBC) and last but not least their business processes.
The reason so many enterprises are embracing the idea of organizing their knowledge in a knowledge graph is the same: in the 21st century, when our information is only as good as our ability to navigate through it and derive meaning from it, enterprise knowledge management needs the help of machines. What attracts companies to building and maintaining a knowledge graph is its ability to integrate disparate data and use these data as background information, that is, as a context that serves enhanced analysis and decision-making. Click To Tweet
A knowledge graph structure not only allows an enterprise to organize, manage and discover internal data, but also to link these data to external data sources and benefit from the network effect. It is this collecting and connecting of seemingly unrelated facts that turns the technical talk of joining the nodes (see Data Daiquiri: The Power of Mixing Data) into the business walk of bringing more business through data integration. Exploring information, starting from any point and further tracing all the relationships this information enters into, takes the knowledge discovery journey to an entirely different level.
Linking data and exploring these data as a knowledge graph structure is beneficial to the enterprise for two reasons:
As explained in Exploiting Linked Data and Knowledge Graphs in Large Organisations:
[…] Knowledge Graph is changing our ways of accessing information and knowledge:
answers to implicit questions behind your searches surface directly from millions of relevant documents when you are googling […]; you are empowered better than ever with the effective access to your colleagues’ expertise in problem solving […]
In an era of highly connected people and data, doing business grows increasingly dependent on doing data integration efficiently. The more data a business is able to contextualize and personalize, the more opportunities stemming from the networked effect there will be.
How NuMedii built an expert knowledge graph to support their researchNuMedii set out to build an intelligent analytical solution supporting research activities related to the identification of new therapies for treating idiopathic pulmonary fibrosis (IPF).
Nu Medii built an expert knowledge graph covering genomics, proteomics, metabolomics, disease conditions, drug products, scientific literature and various biomedical ontologies. At the same time, they annotated large corpora of scientific literature with gene, disease, compounds and drug concepts and mined for the generic relationships between them.
Whether we’re talking about enterprise knowledge management, content integration, market intelligence or smart data, at the end of the day, what a knowledge graph brings to the table is context. Context through which the noise is turned into a signal, that is, the disparate, seemingly meaningless data are integrated and transformed into an edge. The edge only smart can give you.