Revised IPPC 2011 Results Available

[Scott Sanner]

Hi all,

Taking into account a number of issues raised since June 14th, I have now posted the final version of IPPC 2011 results:

  http://users.cecs.anu.edu.au/~ssanner/IPPC_2011/

  http://users.cecs.anu.edu.au/~ssanner/IPPC_2011/IPPC_2011_Presentation.pdf

MDP: Rankings for the *competition* are unchanged; a corrected version of the SPUDD entrant (server communication bug fixed, algorithm unchanged) has been included as a *post-competition* result for purposes of having an accurate value iteration baseline.

POMDP: Some results for POMDPX_NUS had been excluded because they used two different client names... this has been fixed in their log file and their score has increased slightly.  After consultation with the KAIST-AILAB team, we agree that POMDPX_NUS should receive the sole *1st place* award with KAIST-AILAB now taking the sole *2nd place* award for the Boolean POMDP track.

If you are publishing any analysis from the competition, please ensure that you use the revised results data here:

  http://users.cecs.anu.edu.au/~ssanner/IPPC_2011/IPPC_2011_Data.tgz

The README has been updated, so please check there to see if it answers any questions you have with this data.

Cheers,

Scott

P.S. The next IPPC should run in 2013.  I think it is crucial that competitors who have run in a previous IPPC run the 2013 competition, so that the competition improves with each iteration.  If you are willing to run the next IPPC (or even just one of the tracks), please send an email to Lee McCluskey (t.l.mccluskey [at] hud.ac.uk), to inform him of your interest in doing so.