[mrp-users] clarification on rules of participation: third-party data

[Stephan Oepen]

dear colleagues,

there is one aspect of participating in the MRP 2019 task that we feel
we have not sufficiently impressed on you so far: for increased
comparability, the task design puts some constraints on which
additional data or pre-trained models can be used in system
development.

the general idea is to collectively maintain a list of legitimate
additional resources.   only resources on the MRP ‘white-list’ can be
used in preparing your system submissions; but we invite your inputs
on which data or models might be beneficial to meaning representation
parsing and should be white-listed.

we have expanded the discussion of this ‘closed track’ assumption on
the task web site and have provided an initial white-list.  also, we
have extended the deadline for additional resource nominations to
monday, june 3.  please see:

  http://mrp.nlpl.eu/index.php?page=4#companion
  http://svn.nlpl.eu/mrp/2019/public/resources.txt

in case you anticipate you might want to use additional data or
pre-trained models in system development, please make contact with us
as quickly as possible, and no later than monday, a little more than a
week from today.

we look forward to hearing from you!

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