| − | Google seems to populate the graph database based on English Wikipedia results and then expands on the data once the main object is created. For instance, if a Wikipedia biography is created, KG creates a similar object about a week later and does some of its own research to expand upon the biography, obtaining information about family relations and such and making liberal use of images that Wikipedia might reject for copyright reasons. If such a biography get deleted from Wikipedia, it seems that it is also deleted from KG within a few weeks.<ref>https://googleblog.blogspot.co.uk/2012/05/introducing-knowledge-graph-things-not.html</ref> | + | Google seems to populate the graph database based on English Wikipedia results and then expands on the data once the main object is created. For instance, if a Wikipedia biography is created, KG creates a similar object about a week later and does some of its own research to expand upon the biography, obtaining information about family relations and such and making liberal use of images that Wikipedia might reject for copyright reasons. If such a biography get deleted from Wikipedia, it seems that it is also deleted from KG within a few weeks.<ref>https://googleblog.blogspot.co.uk/2012/05/introducing-knowledge-graph-things-not.html</ref> The semantic information built up in KG is used in the Google Instant feature and in the Google Search results. |