Assignment 1: Get the ERG up and running

All term projects in this class will involve manipulating the output of the ERG (or another DELPH-IN grammar), so everyone needs to be able to produce (and manipulate) MRSs. The DELPH-IN consortium makes available several processing platforms that can work with the ERG and other DELPH-IN grammars; which you use is up to you:

• ACE developed by Woodley Packard (see also AceTop on the DELPH-IN Wiki)
• PET (as packaged in the LOGON system)
• Agree developed by Glenn Slayden
• RESTful API to the ERG, from the University of Oslo

The ERG itself can be downloaded here, or retrieved as part of the LOGON distribution. If you plan to work with another language, contact me.

If you plan to use python, pyDelphin will likely be useful.

To turn in

Due Monday 4/4, by 11pm.

• A brief description of which processing platform you decided to use, and any issues you encountered in setting it up.
• A sample sentence (from the data set you plan to target, if you have already chosen, if not a sentence of interest to you)
• The MRS assigned by the ERG to that sentence, for the top ranked parse per the parse selection model distributed with the ERG.