Just a question on the submission process.
Must our systems produce results on the same scale as you have in the
training data (a 0 - 100 scale, where 0-25 is low, 38-62 is medium,
and 75-100 is high?).
If that is the case, should we assume that our fine grained scores
will automatically be considered as coarse grained scores, or do we
need to submit those separately?
If that's not the case and we can use our own scoring scale, should we
then submit separate fine and coarse grained files (since you won't
necessarily know what our binning strategy is...) And if we did that,
would each pair of fine/coarse grained scores count as a single
variant (so we'd end up with 3 pairs of scores, or six files...)
The 3 variant limit is based on the following from the test data...
> 3. Submission information
> Please send the output of your participating system to disco201...@gmail.com, no later than
> April 8, 2011. You can submit up to 3 variants for your system.
> Don't forget to submit a 4-page system description on the same date or earlier. Please check
> http://disco2011.fzi.de/ for details.
Thanks!
Ted
--
Ted Pedersen
http://www.d.umn.edu/~tpederse