[Corpora-List] Task-based Dialogue Comparable Corpora

Raffaella Bernardi bernardi at inf.unibz.it
Thu Apr 1 13:34:58 CEST 2010


At the Free University of Bozen-Bolzano, we have had on line a multilingual chatter bot (English, Italian and German) -- BoB (Bolzano Bot) -- to answer user questions about the Library of FUB:


It has been online since October 2008 for English, since February 2009 for German and since September 2009 for Italian.

We are now working on cleaning up the logs from personal names for privacy reasons. Let me know if they can be of help. We have also annotated the logs with info such as: Topic Continuation, Topic Shift, etc.. More detailed info will be posted as soon as the logs are available --- hopefully by end of May.

The project has been carried out by KRDB members together with the University Library staff.


========================================================== Erasmus Mundus European Masters Programs in Computational Logic http://www.computational-logic.eu Language and Communication Technologies http://www.inf.unibz.it/mcs/lct/ ==========================================================

------------------------------------------------------- Raffaella Bernardi | Free University of Bozen-Bolzano | Faculty of Computer Science | P.zza Domenicani, 3 | I-39100 Bolzano | Room: 2.19 | Phone:+39 0471 0 16122 | Fax: +39 0471 0 16009 http://www.inf.unibz.it/~bernardi --------------------------------------------------------

On Mar 28, 2010, at 7:08 PM, Konstantina Garoufi wrote:

> Dear Carlos,
> the GIVE-2 corpus of Giving Instructions in Virtual Environments
> (currently in English and German) may be useful to you:
> http://www.give-challenge.org/research/page.php?id=give-2-corpus
> Best,
> Konstantina
> Date: Tue, 23 Mar 2010 03:13:05 -0400
> From: Carlos Gómez Gallo <cgallo at fas.harvard.edu>
> Subject: [Corpora-List] Task-based Dialogue Comparable Corpora
> To: corpora <corpora at uib.no>
> Dear Corpora-list members,
> I am working on compiling a list of task-based dialogue corpora such
> as ATIS, Maptask, and TRAINS. In the type of data set I am interested,
> there are usually two interlocutors (or one human interlocutor and a
> dialogue system) who collaborate in executing a common task. I am
> particularly interested in corpora that has been collected in two
> languages or more. In other words, that the experiment/task/session
> was run in more than one language. This corpus would allow the
> comparison of linguistic forms of those languages in contrast (hence,
> comparable corpus; and not parallel corpus which I take it to mean the
> translation of a corpus into at least another language).
> Your responses are greatly appreciated.
> Thanks,
> Carlos Gómez Gallo
> _______________________________________________
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora

More information about the Corpora mailing list