Paper

Abstract

Introduction

Hearst et al. (2002) noted that the design of search interfaces depends heavily on the types of answers required by users, from specific facts to collection overviews to deep analysis of corpora. We posit that there is another dimension along which search varies: the complexity of the question.

Hearst et al. (2002) and others have noted the wide range of searches routinely performed by users. suggest that search tasks vary along a continuum of

Keyword search is most useful when the topic of interest can be described by a well-known word or phrase. Faceted navigation is useful for summarizing a document collection and helping the user refine a search to a conjunction of facet values. However, neither of these techniques work ideally when the user is interested in the specific relationships between entities. For example, an intellectual property lawyer searching a patent database may be looking for instances of, say, aluminum used to protect a laptop computer.

health science, legal case studies, patent searching, social networks

System

Interface

Triple Extraction

Usability study

Conclusions and Future Work

Acknowledgements

• Shiry Ginosar
• Rob Miller
• Boris Katz