Three Forms of Textual Question Answering

             Three Forms of Textual Question Answering


                       Sanda Harabgiu
                Department of Computer Science
                University of Texas at Dallas


Textual Question Answering Q/A re-emerged in the past few years as an
exciting application, in which a question expressed in natural
language is answered by a text snippet containing the required
information. In 1999 the Text Retrieval Conference (TREC) first
performed Textual Q/A evaluations for open-domain questions that have
answers in a very large text collection. The TREC evaluations were an
unqualified success and now open-domain information seeking textual
question answering continues to be annually evaluated by TREC. For
information seeking textual question answering, the recognition of the
expected answer type plays an important role in answer extraction.

A different form of question answering is imposed by the reading
comprehension task, which requires a natural language question to be
answered by pinpointing a simple sentence from a short story, as an
approximation of text comprehension. High precision question answering
for text comprehension requires an extended set of expected answer
types as well as thematic role recognition.

A third form of textual questioning arises when multiple interactions
between a user and the Q/A system are considered. Interactive question
answering is characterized by the inference of the question context
and by more complex inquiries that need to be processed.

In this talk all three forms of textual question answering will be
covered and details of question processing and answer extraction will
be provided for each of them.


If you would like to meet with the speaker, please contact:

                     Feiyu Xu

This seminar series is jointly organized by the Department of
Computational Linguistics and Phonetics and the European Post-Graduate
College in Language Technology and Cognitive Systems.

A current version of the program for this term can be found at:

http://www.coli.uni-sb.de/colloquium/