Query by Annotation

Steven Bird, CSSE, University of Melbourne
Haejoong Lee, LDC, University of Pennsylvania

Abstract:
Databases of hierarchically annotated text occupy a central place in
linguistic research and language technology development.  I will
describe a new approach to tree query called "Query by Annotation".
Users express a query by annotating a tree, and the annotation is
compiled into an expression in a path language and thence to SQL for
execution on a relational database.  Result trees are annotated with
the original query in order to demonstrate why the query matches the
tree.  Since queries and results are annotated trees, users can easily
refine and resubmit their queries.  The approach to Query by
Annotation is motivated and exemplified using a popular kind of
linguistic database known as a treebank.  An open source
implementation, distributed with the Natural Language Toolkit, will be
demonstrated.  This research forms part of a recently concluded
NSF project "Querying Linguistic Databases", conducted in
collaboration with Susan Davidson and Mark Liberman.