Technology companies today are using the power of Knowledge Graphs to revolutionize their complex businesses - uncovering hidden insights and accelerating product development cycles. They have also found a niche in areas where a high degree of competency is required, such as building and configuring spacecraft, nuclear reactors, and operational networks.
Although Knowledge Graphs seem to have suddenly emerged just a few years ago, they are built on a firm framework of tried and tested data models and query languages. These have been patiently developed since the dawn of the information age, waiting for their moment to shine.
Knowledge Graphs represent an interface between people and the hidden value of deep data analytics. They give human decision-makers a comprehensive overview of connected data, even when these are in numerous and scattered siloes of formats, languages, and sources – unlocking the full potential of machine-readable big data by making it graspable by human minds.
What are Knowledge Graphs and What’s the Benefit?
A Knowledge Graph can be defined as a graphical representation of the entities and relationships between entities in a semantic data model (or knowledge model).
While the concept behind Knowledge Graphs is not entirely new, they have flourished and proliferated in recent years due to wider awareness of their value in data-driven businesses.
The graph itself is an accurate visual representation of all the ‘knowledge’ which is contained within a semantic data model. It has value because it visibly draws the connections between data points and shows the context behind the data. They are not the same thing as ‘data visualization’ - they’re much more than that. While the model can exist without the visual representation, the Knowledge Graph is entirely supervenient on the model it is based on – it is inextricable from the data model itself, and this is what makes it special.
In simpler terms, if we can think of the semantic data model as a treasure chest, packed full of potential insights and connections, then the Knowledge Graph is the key that unlocks this treasure chest and lets us ‘mere humans’ access the riches within.
How Knowledge Graphs got so Famous
In 2012, Google unveiled their Knowledge Graph, which brought the concept to the attention of a wider audience. This was part of their new ‘things not strings’ approach. Google’s new Knowledge Graph-based paradigm was used to create an improved ‘answer engine’ that would provide meaningful results based on an understanding of the context behind people’s search queries.
However, Google’s Knowledge Graph stands on the shoulders of the previous successes of other ‘behind-the-scenes’ projects, such as WordNet.
WordNet was born in 1985, at the Cognitive Science Laboratory at Princeton University. It’s a lexical database that has grown from its original English-only version to include lexical relations between concepts expressed in more than 200 languages. It’s a perfect example of a Knowledge Database. It puts data into a meaningful context that can be read by machines, and iIt’s used to power several AI and (natural) language processing applications. If you added a visual representation of these relationships, then you would have a Knowledge Graph that is understandable for humans too.
Semantic data models like these have evolved as the result of an attempt to make data readable and understandable to machines. It’s also a vital step towards building a machine learning algorithm or AI.
Moreover, these semantic data models have become the foundation that Knowledge Graphs are built on.
Making complex relationships between data visible for people is the primary value of Knowledge Graphs. It has the secondary benefit of being able to extract hidden layers of knowledge that are not immediately apparent. Knowledge Graphs can unpick the complexity and shine a light on the patterns between seemingly unconnected entities. This improved insight is like having a superpower - one that can guide better decisions, by using incredibly large and complex data sets and converting them into something that human minds can use.
Closing Gaps in Knowledge – Knowing the Unknown
There are always gaps in our knowledge; we just don’t always know what they are. To quote the once US Secretary of Defense Donald Rumsfeld:
“…there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns—the ones we don't know we don't know.” - Donald Rumsfeld, 2002.
While we have no way of knowing what the ‘unknown unknowns’ might be, Knowledge Graphs can help to make what we do know more useful.
They can draw connections that are not immediately apparent, and lead to discovering ‘new’ knowledge that can be extracted from existing data using deduction. This gives decision makers the amazing ability to derive previously unknown knowledge, a priori, using large data sets of incredible complexity. Sherlock Holmes would be proud.
Being able to draw these hidden connections is what makes it possible for machine learning and artificial intelligence to discover new drugs, predict crime, detect fraud and money laundering or enhance productivity. Turning these connections into a visible Knowledge Graph maintains the connection between these insights and the human decision-makers in charge.
This capability is notably being harnessed by numerous tech companies to make their businesses more effective, more addictive, and more profitable.
Who uses Knowledge Graphs?
Thanks to the revolutionary insights that Knowledge Graphs can provide, they have found themselves incredibly useful in a number of situations where data can impact decision-making.
Examples of Knowledge Graphs in action:
NASA – The National Aeronautics and Space Administration (NASA) has started employing Knowledge Graphs for various purposes. Their complex operations cover all elements of design and commission of spacecraft, operations management, and personnel. To manage these incredibly data-heavy areas, NASA uses Knowledge Graphs to map out their Air Traffic and Aviation data, manage systems engineering, and to find the best talent for their skill-specific roles.
Netflix – Not only does it use its data model to provide algorithmically-derived recommendations, Netflix uses a combination of machine learning and Knowledge Graphs to guide the creative process itself when deciding on new content concepts.
“… the real power of the graph is the information contained in the relational structure. At Netflix, we apply this concept to the knowledge contained in the content universe.” - Netflix Technology Blog
Microsoft – The software and technology company is incredibly active in Knowledge Graph research, and has built several Knowledge Graphs. These are being used to probe how they might optimize scientific research, cyber security, and Natural Language Processing.
Google – As already mentioned, Google’s Knowledge Graph is being used to power better search results, based on answers that reflect context. The 500+ billion facts contained in the Knowledge Graph is also used by voice assistants like Google Assistant to answer queries.
Siri, Alexa, etc. – Virtual assistants can provide meaningful answers to requests because they are powered by Knowledge Graphs and the semantic models they’re based on. Natural Language Processing (NLP) software also uses semantic data models like Princeton’s WordNet to train and maintain their understanding of spoken language.
Semantic Web (Web 3.0) – This is an ongoing project with the goal of making all information flowing through the internet machine-readable. The dream is to make it so that all daily transactions are handled automatically by intelligent machines. This might include the automatic execution of contingent actions based on blockchain and smart contracts, for example.
Social Platforms – Social platforms like Facebook are based on a social graph structure, formally integrated into their own Knowledge Graph, which they call the Entities Graph. This has been used to power Facebook’s Graph Search (now discontinued) and is almost certainly still used for behind-the-scenes processes. LinkedIn uses their own Knowledge Graph to discover data insights that lead business decisions and find new ways to increase member engagement.
Commercial optimization – Uber uses their Knowledge Graphs to optimize their ride-sharing platform, as well as powering Uber Eats’ food recommendations. Many technology-based commercial enterprises including eBay, and Amazon have spent a lot of time developing their own proprietary Knowledge Graphs to guide their operations more effectively.
Wherever there is a vast amount of data with complex interconnections and valuable insights to be extracted, there’s a use-case for a Knowledge Graph. The tricky part is creating one.
As evidenced by most of the cases described above, semantic data models and Knowledge Graphs are typically created on a case-by-case basis. This can be a time-consuming task, but it is seen as the only option when highly specialized systems, data types, and architectures are involved. As many technology companies already have the dedicated staff capable of creating and delivering a proprietary solution, a custom-made Knowledge Graph is frequently the result. But this takes a lot of time and effort.
How get started using Knowledge Graphs even faster
There are several service providers who can create a fully customized data model and knowledge graph that perfectly matches the form and fit for your company - but is this the solution you need? A custom-built approach is often too slow for businesses where time-to-market is a primary concern.
In settings where knowledge graphs have the most value – such as the fast-moving SaaS and Tech sectors, time is of the essence. This means companies need to be able to start extracting the benefits of Knowledge Graphs as soon as possible – without waiting for a custom-made data model.
An off-the-shelf solution would provide the instant access required, but these are comparatively few and generally lack the flexibility required to adapt around each unique situation. But there is one solution that matches these criteria.
Weaver Technologies has developed a flexible data platform which can integrate data of any type, source or format and turn it into a powerful and fully customizable Knowledge Graph. The wvr.io knowledge graph platform is free to use from the start (while you learn how to make the most of it), then once you start adding more users and workspaces you pay based on a transparent pricing system that reflects the actual use.
This platform is an important step forward in the evolution of Knowledge Graphs, and will kick-start data-led decision making for a lot more companies – particularly in the tech sector, where the data is already there just waiting to be used.