Entity-Based Query Interpretation

Web search queries can be rather ambiguous: Is "paris hilton" meant to find the latest news on the celebrity or to find a specific hotel in Paris? And in which of the worldwide more than 20 "Parises"? We propose to solve this ambiguity problem by deriving entity-based query interpretations: given some query, the task is to link suitable parts of the query to semantically compatible entities in a background knowledge base. Our suggested approach to identify the most reasonable interpretations of a query based on the contained entities focuses on effectiveness but also on efficiency since web search response times should not exceed some hundreds of milliseconds. In our approach, we propose to use query segmentation as a pre-processing step that finds promising segment-based "skeletons". These skeletons are then enhanced to "interpretations" by linking the contained segments to entities from a knowledge base and then ranking the interpretations in a final step. An experimental comparison on a corpus of 2,800 queries shows our approach to have a better interpretation accuracy at a better run time than the previously most effective query entity linking methods.