CHI-ANNOUNCEMENTS Archives

ACM SIGCHI General Interest Announcements (Mailing List)

CHI-ANNOUNCEMENTS@LISTSERV.ACM.ORG

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jeff Dalton <[log in to unmask]>
Reply To:
Jeff Dalton <[log in to unmask]>
Date:
Thu, 6 Jun 2019 00:51:05 +0100
Content-Type:
text/plain
Parts/Attachments:
text/plain (71 lines)
-----------------------------------------------------------------------
TREC 2019 Conversational Assistance Track (CAsT)
-----------------------------------------------------------------------

Webpage: http://treccast.ai
Twitter: @treccast <https://twitter.com/treccast>
Mailing List: https://groups.google.com/forum/#!forum/trec-cast

 --- GOALS  ---
The goal is to advance Conversational Information Seeking research and
create large-scale reusable test collections for open-domain conversational
search systems.

This is the first year of TREC CAsT, which is running as a track in the
Text Retrieval Conference (TREC). The main task is that, given a sequence
of utterances that a user might ask to a conversational agent like the
Google Assistant, Cortana, Siri, or Alexa, to retrieve relevant content
that satisfies the user's information need. This requires:
 * Tracking the evolution of the information need in the conversation and
identifying salient information for each turn
 * Retrieving candidate response passages from a large heterogeneous
collection of paragraphs

--- Task Description ---
See the website (http://treccast.ai) for full details, including the Year 1
Task Guidelines
<https://docs.google.com/document/d/1SH6UdZ9xQUZzhxnCzlsJvXtPEUCuiow5MXlrsA22pKs/edit?usp=sharing>
and
trainings topics.

For year 1, participants of the track will be challenged with retrieving
relevant paragraphs for each turn in a conversational topic.
* Topics: Manually developed information seeking topics
* Data collection: (1) Wikipedia paragraphs, (2) Microsoft MARCO web
passages, (3) Washington Post news

Note: Where possible we link our topics to the recently released MS MARCO
Conversational Search corpus
<https://microsoft.github.io/MSMARCO-Conversational-Search/>, a large
collection of artificial information seeking sessions linked to the MS
MARCO dataset <http://www.msmarco.org/dataset.aspx>.

--- Important Dates ---
ASAP: TREC Registration:
https://ir.nist.gov/trecsubmit.open/application.html
Note: To register, send an email to [log in to unmask] to obtain a registration
key.

Test topic release: June 12th
Run submission: August 16th
System papers due: Late October
Conference at NIST: November 13-15 2019

--- Organizing Committee ---
Jeff Dalton, University of Glasgow
Chenyan Xiong, Microsoft Research AI
Jamie Callan, Carnegie-Mellon University

--- Contact ---
Mailing list: [log in to unmask]

    ---------------------------------------------------------------
    For news of CHI books, courses & software, join CHI-RESOURCES
     mailto: [log in to unmask]

    To unsubscribe from CHI-ANNOUNCEMENTS send an email to
     mailto:[log in to unmask]

    For further details of CHI lists see http://listserv.acm.org
    ---------------------------------------------------------------

ATOM RSS1 RSS2