Connectionists: Pascal Large Scale Learning Challenge

[Elad Yom-Tov]

Apologies for any cross-postings.

Pascal Large Scale Learning Challenge
==================================

The Large Scale Learning challenge is aimed at comparing the scalability
and efficiency of existing classification algorithms when applied to
large data. The goal of this challenge is to encourage research in
Machine Learning and related applications by providing several large
datasets, uniform scoring procedures, and a forum for organizations
interested in comparing their results. The experience in fields such
as Information Retrieval has shown that such challenges are extremely
useful in advancing the state of the art and in fostering new approaches
to the problem. 

Time Line 
=========
27 February 2008   Challenge Announcement
12 June 2008       Test sets available
26 June 2008       End of the competition
27 June - 8 July   Re-calibration of all entries
9 July 2008        ICML'08 Large Scale Learning workshop
10 July - ...      Re-run of the top ten methods

Call for Participation
======================
We encourage participants to download the datasets, submit their results
to the challenge web site for instant feedback on their performance
over a validation set, and to send their predictions for classification of
a test set. 

Prizes
======
We will award travel grants or cash prizes to the winners.

Evaluation
==========
The results of the challenge will be discussed at an ICML'08 workshop,
where the best and most innovative approaches will be presented.
Participants will be invited to submit to submit a paper to a JMLR
special topic on large scale learning (*TBC*).

Detailed information can be found at:
http://largescale.first.fraunhofer.de/