Million Query (1MQ) Track
This is where information on the Million Query Track of TREC 2008 appears.
URGENT UPDATES (on June 16th):
The goal of this track is to run a retrieval task similar to standard
ad-hoc retrieval, but to evaluate large numbers of queries
incompletely, rather than a small number more completely. Participants will run
10,000 queries and a random 1,000 or so will be evaluated. The corpus is the terabyte track's
GOV2 corpus of roughly 25,000,000 .gov web pages, amounting to just under
half a terabyte of data.
- Track submission deadline extended to Tuesday, June 17th
- trec.nist.gov is currently down; you can submit your runs at http://ir.nist.gov/trecsubmit/mq.html (you will need the participant username and password).
Participants in the track will:
For full details, please see the latest draft of the track guidelines.
- Run 10,000 (short) queries against the collection
of 25,000,000 .gov web pages (just under a half terabyte of data)
- (Optionally) participate in a web-based evaluation process,
judging as many queries as they can in July and August
(perhaps continuing into the fall). On average, someone is likely to
judge 4-40 documents per query.
The track is being organized by James Allan at UMass Amherst and Jay Aslam at Northeastern University.
If you are interested in participating in the track, you must apply to be a part of TREC 2008.
The official deadline for applying is February, 2008.
Late applications are sometimes accepted, so if you're genuinely
interested and you've missed the deadline, you should try.
To subscribe to the track's mailing list, see the
list page. Subscribing will require a confirmation phase via
email. You do not need to be part of TREC 2008 to subscribe to the mailing list. There is an archive of the mailing list as well as an archive of last year's mailing list.
You must be subscribed to the mailing list to send messages to it. At
the moment, messages from non-subscribers will be held for moderator
approval, but once spam hits the list, such messages will be
discarded without a glance.