[mrp-users] start of the evaluation period

[Stephan Oepen]

dear colleagues,

(0) the LDC confirmed earlier today that all registered participants
who have previously obtained the MRP 2019 training data now should
have access to the evaluation data package.  please login to the LDC
‘My Account’ pages and consult the available downloads:

  https://catalog.ldc.upenn.edu/login

(1) the CodaLab competition page for the task is open and appears
functional: we have already received ten or so registrations and a
couple of trial submissions.  to make sure everything is smooth once
you are ready to submit parser outputs for evaluation, please register
for the task in the next few days:

  https://competitions.codalab.org/competitions/23524#participate

(2) several of you have commented that you find the efficiency of the
MRP scorer inconvenient (to say the least).  a substantially faster
version is available (thanks to code optimizations by hiroaki ozaki)
in a separate ‘dev’ branch, which we encourage participants to try
out.  as of yesterday, this version also supports parallelization of
MRP scoring; please see:

  https://github.com/cfmrp/mtool/issues/55

(3) there have been some conversations on the Microsoft GitHub
repository for mtool that could be of wider interest.  in case you
have not been watching the repository closely, please consider looking
at its issue tracker (incuding closed conversations), for example:

  https://github.com/cfmrp/mtool/issues/35
  https://github.com/cfmrp/mtool/issues/44
  https://github.com/cfmrp/mtool/issues/57

as always, please do make contact with us (by email, or on GitHub for
mtool-related business) as questions arise or in case you have any
suggestions for us!

happy parsing!

stephan oepen, omri abend, jan hajič, daniel hershcovich,
marco kuhlmann, tim o'gorman, and nianwen xue