How Google Knowledge works: Click for the full graphic.
The Google Search of the future is here. Now. Today. The long-talked-about semantic web — Google prefers “Knowledge Graph” — is rolling out across all Google Search tools, and our most fundamental online task may never be the same again.
Starting today, a vast portion of Google Search results will work with you to intuit what you really meant by that search entry. Type in an ambiguous query like “Kings” (which could mean royalty, a sports team or a now-cancelled TV show), and a new window will appear on the right side of your result literally asking you which entity you meant. Click on one of those options and your results will be filtered for that search entity.
To understand the gravity of this change, you need to know about the fundamental changes going on behind the scenes at Google Search. As we outlined in our report earlier this year, Google is switching from simple keyword recognition to the identification of entities, nodes and relationships. In this world, “New York” is not simply the combination of two keywords that can be recognized. It’s understood by Google as a state in the U.S. surrounded by other states, the Atlantic Ocean and with a whole bunch of other, relevant attributes.
As Ben Gomes, Google Fellow, put it, Google is essentially switching “from strings to things.”
To build this world of things, Google is tapping a variety of knowledge databases, including Freebase, which it bought in 2010, Wikipedia, Google Local, Google Maps and Google Shopping. Currently, Google’s Knowledge Graph has over 500 million people, places and things and those things have at least 3.5 billion attributes.
That’s a lot of things. According to Google, search users will see these new knowledge graph results at least as often as they see Google Maps in results. In fact, this update will have a greater initial impact than the updates that brought Google Images, videos, news and books, combined. It’s big and it’s probably going to be everywhere. …
In addition to the window which will help users find the right “thing,” Google will also surface summaries for things, which, again, will try to be somewhat comprehensive by tapping into the various databases of knowledge. A search for Frank Lloyd Wright, for instance, will return a brief summary, photos of Wright, images of his famous projects and perhaps, most interestingly, related “things.” People who search for Wright are also looking for other notable architects. It’s a feature that may remind users of Amazon’s penchant for delivering “people who liked this book also bought or searched for this one” results.
Gomes said that the search results are tailored to deliver information that best relates to the initial search result. So the details delivered about a female astronaut will likely outline her space travel record, because that’s what people who search for her are, according to Google, most interested in. ….
Of course, not every “thing” is the right thing. Wikipedia is, for example, a community-sourced encyclopedia that is known for both its breadth and depth of information and the occasional whoppers of misinformation it stores. Google’s Knowledge Graph includes an error reporting system. When users find misinformation, Google will share it with the source and the knowledge graph will get just a little bit smarter …