Google’s Knowledge Graph Deconstructed

The term “World Wide Web” has been used to describe the Internet for many years, despite the fact that the Internet is less a web than a repository of random, unconnected information. That is, until now.

Both Google and Microsoft are taking their respective search platforms to the next level by attempting to make relevant and relational connections between all searchable concepts on the Web, as opposed to continuing forward with what has until now been search based on syntax. Google’s attempt is dubbed the Knowledge Graph, while Microsoft has labeled their effort Satori.

In other words, the two search competitors intend to make the Web more like an actual web by providing users with content that is connected to their core search based on what previous users have searched for following the initial query. The data used to provide the related content include other user searches and their subsequent linking within the primary search as well as what kind of content is contained on the Web page.

For example, a Google search for “lions” will return the usual list of results down main area of the results page and also provide a snapshot of related content to the right of the list (a space that is constantly being repurposed by Google). In this instance, there are only two subcategories: the animal and the Detroit Lions. In broader searches, however, users can expect to find a richer field of subcategories and related content.

The explanation for how this is done is naturally very complex, but generally involves the indexing of literally trillions of pieces of information so that they can be instantaneously recalled and catalogued based on their relationships to primary searches. Analysts estimate Google has mapped 500 million Web “objects” like monuments, celebrities, landmarks and the like, and have processed 3.5 billion facts regarding their relationships among one another.

Google engineers call their cache of past user searches the “zeitgeist” – a detailed map for all things relevant of this time and age. Using this information to better calibrate searches and follow-on data is akin to polling hundreds of millions of people about a given topic to further distill its relevant essence for the individual. In many ways it’s like a Magic 8 Ball that can predict what people want to know about a topic.

Even so, critics are quick to point out that Google does not have access to Facebook’s treasure trove of data, as Microsoft does, which could give Satori an edge in the race to spin a truly connected Web.

Leave a Reply

Connect With Us