A knowledge graph inside Google Search
On 16 May 2012 Amit Singhal announced that Google Search would start answering about things rather than strings. The company claimed more than 500 million objects and more than 3.5 billion facts and relationships between them, drawn from Freebase, Wikipedia, the CIA World Factbook and other sources.
Why it matters
For twenty years knowledge bases had been an academic matter with a measurable size and no audience. On this day one of them became the thing a hundred million people look at the world through daily - and became invisible in the same move: the user saw not a graph but a box of answers beside the familiar links.
The announcement was written by Amit Singhal, SVP Engineering. It opens on the query [taj mahal], which may mean the monument, a Grammy-winning musician, a casino in Atlantic City, or the nearest Indian restaurant. The figures are Google's own and cannot be checked from outside, so the second source presents them as exactly that - a claim: Danny Sullivan in Search Engine Land writes "Google says it has compiled over 3.5 billion facts". The launch was partial: "We've begun to gradually roll out this view of the Knowledge Graph to U.S. English users." What this record does not claim. The thirty-seven per cent is not a share of queries in general but one example: "the information we show for Tom Cruise answers 37 percent of next queries that people ask about him." And the name "knowledge panels" is not Google's: the word panel occurs zero times in the announcement, and Sullivan writes the same day that the units had no name - "they have no official name (and I did ask), so I'll call them 'knowledge panels.'"