Connectomics challenge

[Isabelle]

Dear Causality Enthusiasts,

We just launched a new challenge that you should be interested in:

First neural connectomics challenge:

from imaging to connectivity

http://connectomics.chalearn.org/

Your algorithms of cause-effect pairs or causal network reconstruction should be highly relevant.

==> What you get: Time series of the activity of 1000 neurons.

==> What you predict: The directed connections between neurons.

30 teams entered in less than a week! The sample code makes it easy to quickly get started, as testified by this forum post:

"Just wanted to pay my compliments to the authors of the sample code and visualizations.  Incredibly well documented and thorough work.  Really a joy to run through."

There will be $3000 in prizes awarded, and the winners will be invited to present at WCCI 2014 and ECML 2014, see the schedule. The best papers will be invited to contribute a full length paper to a special topic of Frontiers in Neuroscience.

February 5, 2014: Start of the challenge (development period).

May 5, 2014: End of challenge.

May 20, 2014: All teams must turn in fact sheets. The three top ranking teams must turn in their code to qualify for prizes. The post-challenge verifications start.

June 1, 2014: End of the post challenge verifications. Release of the official ranking.

June 20, 2014: ECML workshop paper submission deadline.

July 6-11, 2014: Result presentation at WCCI 2014, Beijing, China.

August 30, 2014: Revised papers due for ECML workshop proceedings.

September 15 or 19, 2014: ECML workshop, Nancy, France. Discussion of the results. 

Good luck!

Isabelle

[Mortal Kolle]

Good morning 

I have tried making a submission a couple of times and each time get the message stating 3 columns were found in a particular record (1769966 in fact) where 2 columns are expected - since the code for output was exactly the same throughout, I found that very surprising and don't know what the problem might be. Any ideas?