(062) previous ~ index ~ next
To: Jaime Carbonell <jgc@nl.cs.cmu.edu>
From: George Doddington <doddington@nist.gov>
Subject: Re: TDT3, deconflating the issues
Date: Mon, 29 Mar 1999 22:24:48 -0500
> 4. Task clarification -- sorry if I missed this in earlier communications:
> Are we required/advised to:
>
> - Do the TDT tasks independenly in each language?, or
>
> - Do the tasks jointly? This implies that we may get
> only Mandarin training stories for tracking English and
> vice-versa. It also implies joint clusters of English and
> Madarin stories if these are on the same topic for detection.
The tasks are unchanged by the introduction of multiple languages.
(Consider the language to be an attribute of the source.) Thus topic
tracking includes multiple languages in the training stories, and topic
detection includes multiple languages in the story clusters. You bring
up an interesting point, however, by suggesting monolingual processing
followed by post-processing integration. Hmmm, do you think that that
approach would be a win? Perhaps. Interesting...
--
George Doddington in McLean, VA: doddington@nist.gov or 703/556-3434
(062) previous ~ index ~ next
Last updated Thu May 13 09:28:21 1999