lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dave Kor <>
Subject Re: ApacheCon next week
Date Tue, 27 Dec 2005 11:45:29 GMT
On 12/27/05, Koji Sekiguchi <> wrote:
> Hi Grant,
> > You stole my thunder!  :-)  Was going to post the URL after doing the
> > actual talk, but that's all right.  I will post a few changes I have
> > made on the plane tonight or tomorrow to the website below.
> >
> > Let me know if you have any questions...
> I'm curious to know "Candidate Identification for QA".
> At your PPT slide p.19,
> "Can be as short as a word or as large as multiple documents, based on
> system goals"
> Can you explain it in more detail?

In Question Answering, for questions that expect a single answer (Eg,
Which country won the most gold medals in the 1996 Atlanta Olympics?),
we typically just need to find a single document that contains the
answer we are seeking.

However there are other types of questions require extracting answers
from several different documents because a single document might not
contain all the relevant information. A simple example would be a
question with more than one valid answer (Eg, Who are the presidents
of the united states?). The QA system will have to find answers from
different documents since there might not be a single document that
contain  all the answers. Btw, in QA terminology we label this type of
question as a List question.

Another example are Definition questions, where we would like to
provide all interesting facets on a particular topic (Eg, Tell me all
there is to know about the Grand Canyon). Again, a set of documents
might each describe a single aspect about the Grand Canyon. To build a
complete picture, we may need to sample most documents that mention
the Grand Canyon.

I hope this helps.

Dave Kor.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message