Because Google's search results are so good, it's possible (and maybe even common nowadays) to use Google's rank(someQuery, someDocument) as a relevance label to train your own ranker on, but what Google alleged here is even more interesting: they claim Microsoft used IE to telemetrize users' Google searches, scraping the query and results -- `(someQuery, [document1, document2, ...])` -- phoning that home, and using that _directly as their IR_, presumably by caching it or something. Kind of a cheeky move by Bing!
Edit: Just checked and Chrome wasn't the dominant web browser until mid-2012 or so[0], so IE definitely would have seen a big enough share of Google searches to make this viable. I wonder if this factored into Google's strategy for Chrome and monopolizing search, etc.
Because Google's search results are so good, it's possible (and maybe even common nowadays) to use Google's rank(someQuery, someDocument) as a relevance label to train your own ranker on, but what Google alleged here is even more interesting: they claim Microsoft used IE to telemetrize users' Google searches, scraping the query and results -- `(someQuery, [document1, document2, ...])` -- phoning that home, and using that _directly as their IR_, presumably by caching it or something. Kind of a cheeky move by Bing!
Edit: Just checked and Chrome wasn't the dominant web browser until mid-2012 or so[0], so IE definitely would have seen a big enough share of Google searches to make this viable. I wonder if this factored into Google's strategy for Chrome and monopolizing search, etc.
[0] https://en.wikipedia.org/wiki/Usage_share_of_web_browsers#/m...