The sheer amount of information available nowadays and the near exponential growth of the web have resulted in a continuously increasing role of search engines, to the point they have become the central entry points to the web and a crucial factor in the experience of internet users. Web search engines are increasingly used for satisfying not only the informational needs, but also the navigational and the transactional needs of users. In this context, query logs of large-scale web search engines can provide a large amount of information about the web users, and can be viewed as an implicit source of collective endorsement about the average user's knowledge, needs, and preferences in any particular time frame.
In this talk, I will present how statistics extracted from web search engine query logs can be used in several language processing, information retrieval, and information extraction tasks, among which spelling correction (including spelling correction for the medical domain), re-ranking of search results, extraction of semantically related terms, and user-centric fact extraction for question answering.
