The Google Knowledge Graph, introduced by Google in may 2012, is an attempt to analyse and present information with a focus on the relationships between the elements that are deemed important for a particular search. The data about all possible relationships are created and stored on beforehand on Google’s computers, in a giant database, which, in turn, form the base for the Knowledge Graph.

How does the Knowledge Graph work?

Let assume that you want information about the type of food that people ate in ancient Egypt at the time of the pyramids and that you start your Google Search by typing in “the pyramids”. Nothing more. When you commence that search, the Google search engine will utilize the data in the Knowledge Graph and select a couple of suggestions for you. You are presented with not only links to the home pages across the globe that contain the words “the pyramids” but also a set of themes. These themes may be, for example, “the history of the pyramids”, “the facts about the pyramids”, “the scholarly literature about the pyramids” and so on. Since you are interested in what people ate in ancient times, your select the history theme. Ta da! You are henceforth presented with on-screen information that (in Google’s opinion, based on the information stored in the Knowledge Graph database) is more related to the history of the pyramids than anything else. Then you just continue your search. In summary, you never had to write the word “history” in the search field, you just had to commence your search with a short introductory phrase ("the pyramids") and then select a theme and continue your search.

The idea behind Google Knowledge Graph

Words are not just words beacuse they also refer to relationships, and Google knows this and wants to digest and analyse these relationships in order to produce better search results. Google's Knowledge Graph, and its database, is thus a way to analyse the interrelated connections between items, concepts, et cetera, and by analysing and interpreting these relationships, Google hope to do a better job finding and presenting search result to you, the user. In the end, it is suppose to be faster and more accurate.

Utilizing information based on previous searches

The system also utilizes the information that Google has stored regarding a lot of previous searches that have been performed, by others, regarding the search phrase that you have entered. If you are searching and using the phrase "what is a scientific article", the Google Knowledge Graph will not only try and locate pages that contain these words based on an analysis of the content on web pages around the globe, but it will also utilize information based on its analysis of other peoples' previous searches. It means that previous searches performed by other Google users across the world, using that particular search phrase, is stored, analysed and used as a base for making new decisions today and in the future about what pages to show in the search result.

The launch | Google Blog posts and YouTube Channel

Google introduced and discussed the Knowledge Graph in a blog post on 16 May 2012. Here, they highlighted the fact that a lot of search activities are inductive by nature, meaning that the user also wants to be surprised and find new, relevant information while performing the search. The user is searching for what he or she believes is the focus of the search but at the same time is open for any other related information that might be of value. Google wants to provide those related issues and pieces of information by presenting adjacent and (what they guess are) relevant information in the search results.

Google also discussed the Google new search engine in a video released on the same day on their YouTube channel. The video was entitled "Introducing the Knowledge Graph" and it highlighted and reminded us of the fact that words refer to real world things and relationships. The Google Knowledge Graph is Google's way of harvesting the information that lies within these relationships.