Google Unveils Knowledge Graph
Posted On May 24, 2012
Google introduced what it calls its “Knowledge Graph,” designed to “understand real-world entities and their relationships to one another: things, not strings.” According to the blog post from Amit Singhal, Google senior vice president of engineering, the Google Knowledge Graph is “an intelligent model—in geek-speak, a ‘graph’—that understands real-world entities and their relationships to one another: things, not strings. Google’s Knowledge Graph isn’t just rooted in public sources such as Freebase, Wikipedia, and the CIA World Factbook. It’s also augmented at a much larger scale—because we’re focused on comprehensive breadth and depth. It currently contains more than 500 million objects, as well as more than 3.5 billion facts about and relationships between these different objects. And it’s tuned based on what people search for, and what we find out on the web.”
Google reportedly won’t comment on the exact mix of sources (or the business deals involved), but search project manager Johanna Wright says that “comprehensiveness is the goal.”
By leveraging so many sources of information, Google is trying to keep browsers engaged in Google content for a longer period of time, benefitting Google, its advertisers, and its content providers, says industry analyst John Blossom, president of Shore Communications, Inc., in Westport, Conn. Google’s searches had become too efficient, quickly shunting browsers off to non-Google sites. Now the expectation is that browsers will find most of what they need without leaving the search engine.
Grant Simmons of The Search Agency, an independent, Los Angeles-based search marketing firm, says this is the first of many iterations that will be coming in which Google will be changing the way it classifies the web.
“Search is moving from a list of links to a list of entities,” Simmons says. “Google in the past has listed what it considered to be the most relevant results.” Knowledge Graph treats the queries more like questions, generating the categorized information on the right side, as well as providing quickly accessible nuggets of information.
The Knowledge Graph is embedded in the Google search engine. Now, a Google search produces not just a list of entries, as in the past, but instead a listing on the left side of the page and a categorized listing on the right side. For example, a search for “mercury” will produce full listings on the left, and on the right, an invitation to see results about Mercury, the planet, and mercury, the element. Clicking on the planet or the element will list those results on the left, as well as some specific facts about the item on the right (e.g., the mass, distance from the Sun, and other quick, short factoids about Mercury, the planet).
For items like concerts, Google Knowledge Graph will provide time, ticket information (including purchases), and similar details, Simmons adds. “It makes the searches more actionable.”
In that way, Google Knowledge Graph could change effective online marketing tactics, according to David Wolf, CEO of InBusiness, Inc., Orlando, Fla.
Google Knowledge Graph builds on the idea long used by Amazon and other online retailers of grouping similar items together, according to Wolf and several others. At Amazon, for example, a purchase leads to suggestions for other purchases. Similarly, online electronics retailers will recommend certain cables or other related accessories after a person buys a television or similar electronics item.
“Most people online navigate by searching,” says Wolf, who has blogged about Google Knowledge Graph. “They actually prefer to browse by bouncing thru links, reading pieces of info they see. [For example] Amazon’s website is designed to maximize browsing. The knowledge graph is designed to give users more connections to similar content to enhance the browsing experience. Browsing requires access to information that answers questions that haven't been asked yet.”
The value of the new tool has yet to be determined, according to Wolf. “Google has been making attempts at increasing the relevancy of its searches for a long time. Updates like ‘Caffeine’ a couple years ago as well more recent ones have all be small steps towards relevancy. This is the latest, and may be a short-term fad. Google often removes features that don’t catch on.”
Siamak Farah, founder and CEO of InfoStreet, a Tarzana, Calif.-based cloud applications provider, is much more upbeat on the value of Google Knowledge Graph. “I find it to be fantastic.”
“Google Knowledge Graph provides people with more rich content with more context for multiple devices,” Blossom says. “It leverages the concept of entity relationship mapping, a concept that’s been part of search engine for many years. This is the first time that we have seen entity relationship mapping used with such broad range of topics, leveraging a publicly available web search engine.”
Microsoft’s Bing was the first search engine to leverage entity relationship mapping, thanks to its earlier acquisition of Powerset, Blososm says. Though entity relationship mapping remains an important element, the latest redesign of Bing de-emphasizes the abilities of Powerset, Blossom adds.
Google’s use of entity relationship mapping focuses much more on academic-oriented searches, according to Blossom. In that way, it complements Google’s Chrome Books, which are popular in many classrooms.
“They’re focusing on training a new generation of Googlers,” Blossom says, likening it to the strategy Apple used a generation ago when it focused on the education market.
Simmons expects that future iterations of Google Knowledge Graph and Google search will further build on the current version to provide people with even more capabilities. Blossom adds that a natural extension of the new capabilities is to enable voice-activated and voice-response search, much as Apple has done with Siri on the iPhone and iPad.
“It doesn't work quite like Siri but the knowledge graph is an important layer of data analysis that will be required to meet the goal of surpassing Siri’s functionality,” adds Wolf. “Because Siri is just an application, it is limited to what it can do. Google will own both the application, and the data in which it is analyzing, [that is] a powerful combination.”