In one short sentence: "We are at the beginning of a transformation that will take us a motor to a motor information to know. "Johanna Wright, product manager at Google, summed it wishes to make the search giant with her new baby, Knowledge Graph.
This new feature, which was presented before yesterday by the search giant, will be launched gradually in coming days, the English version of the site initially.
Knowledge Graph is a first step towards the dream of Web inventor Tim Berners-Lee: the emergence of a "semantic web" which would connect not just pages of documents by links, but also concepts, ideas, facts.
Google provides its users with a database that goes in that direction. It contains about 500 million objects and 3.5 billion of facts, links between these objects.
The system knows such as Pierre-Auguste Renoir was an impressionist, was born in Limoges February 25, 1841. The database is able to compare him with other painters from the same school and to identify his work. In short, the link information that are typically isolated.
Note that the system is also able to answer certain questions in natural language ("What is the tallest building on earth" or "wife of Barack Obama" for example) but it is far from complete and the coup frustrating.
For Google, this is just the beginning: "It is an essential first step to build the next generation of Internet search, which exploits the collective intelligence of Web and understands the world like the people" wrote Amit Singhal, vice president of Google, the blog of the company.
According to Google, this new feature improves tripling its service: First, Knowledge Graph avoids the problems of homonymy, offering you to refine your search in case of doubt.
Second, this module summarizes the most important facts related to research, to obtain as quickly as possible good result ... Finally, Knowledge Graph provides more serendipity in research, deepening, thanks to a function specifying queries Associated performed by other users.
Future mobile?
Enabling this new feature is automatic, it appears in a new column when doing a search. It is already possible to try it, by following the link test from this page.
In practice, we think of a mixture of Wikipedia (from which some of these data) and Wolfram Alpha, the search engine that drives the particular vocal assistant Siri ... While not go so far in the approximation, the combination data.
In this regard, we imagine, moreover, that Knowledge Graph will be more interest in research on mobile, because this new module could well serve as the basis for a future assistant close to that of Apple, we would query language natural for answers ...