Guest lecture by Evimaria Terzi: Query Suggestion
on Wednesday, 16th January at 14 in C222.
Abstract: Keyword queries are the main form of interaction between users and search engines. Failure to formulate the right query means bad-quality results and user dissatisfaction. The number of possible alternatives to a given query is huge, even if the space of suggestions is restricted to a given query log. In this talk I will present a methodology for picking a subset of alternative queries from a given query log to be presented as suggestions to the user. The alternative queries are picked in such a way that they remain within the scope of the initial query and at the same time they allow for better exploration of the search space.
I will describe the combinatorial formulation of the query-suggestion problem and present algorithms for solving it. I will also give some experimental results that demonstrate the practical utility of our setting.
This is a joint work with Rakesh Agrawal and Sreenivas Gollapudi from Microsoft Search Labs.
Welcome!