Lab 8 (due 2/28 11:59 pm)

Overview

• Refine/extend your list of MMT translations, as appropriate.
• Choose one more phenomenon illustrated in those examples but not handled by your grammar to fix. Post to Canvas by midday Tuesday for pointers on how to fix it. (I'll likely direct you to lab instructions from previous years.)
• Add a few further examples to your testsuite for that phenomenon (as appropriate).
• Run initial testsuite & testcorpus profile.
• Fix the one phenomenon you chose.
• Make sure your grammar can generate. Edit as necessary until it can. Post many questions to Canvas!
• Get the variable property mapping set up working.
• Translate one sentence from English to your language with the MT system.
• Run final testsuite & testcorpus profile.
• Write it up!

Back to top

Refine/extend MMT sentences

If you don't already have a complete set of MMT translations and you don't have additional phenomena to work on in the ones you do have, try to get more of the, translated. See Lab 7 for instructions.

Note that contrary to last week's instructions, the set up we're using this year expects these files to have one sentence pre line (no blank space in between).

Choose one phenomenon to develop

By midday Tuesday: Choose one phenomenon represented by the MMT sentences not already covered by your grammar to work on this week. Post to Canvas with the IGT for the relevant examples and the phenomena you intend to work on. I will reply with pointers to instructionsn for those phenomena/develop some if necessary.

Extend your testsuite for that phenomenon

Create additional positive/negative testsuite examples that illustrate your chosen phenomena and add them to the testsuite in the usual fashion. This testsuite should now include the MMT sentences plus ~2-6 more examples and should be called lab8.

Initial testsuite run

1. Create and run initial testsuite instances for both the linguist-provided data and your testsuite, using the initial grammar.

   Note If your tsdb/ directory is inside a shared folder on VirtualBox, it will not work.

2. For each of these, explore the results, collect the following information to provide in your write up:
   • How many items parsed?
   • What is the average number of parses per parsed item?
   • How many parses did the most ambiguous item receive?

Back to top

Develop analyses and test your grammar

Based on the instructions I've pointed you to through Canvas + answers to the many questions I hope you will ask, develop analyses for your one additional phenomenon.

For the MMT sentences specifically, you can test your MRSs by looking at the output of this English grammar for the corresponding examples. We don't expect an exact match, but if things are different you should have a clear idea of why. And, of course, you are always welcome to post lots of questions!

Test generation

Test generation with both lkb & ace. Can you generate from short sentences? What about longer ones? To receive full credit on this lab, you need a grammar that can generate from simple transitive sentences and you need to have tested what happens with longer ones (e.g. sentences with clausal modifiers or clausal complements). See Lab 6 for detailed instructions on generation. And, of course, post lots of questions to Canvas.

Variable property mapping

The file semi.vpm provides a mapping between grammar-external features of indices (referential indices and events) and their values, and grammar-internal ones. For background on VPM, see the DELPH-IN wiki. As soon as you start using a VPM file, then only variable properties (features on indices) that are handled in the file are actually preserved.

1. You should already have a semi.vpm file provided by the customization system. Open it up and see which variable properties are there, and then look in your grammar to see what is missing. In general, we'd expect to see all of the features of the types event and ref-ind represented in a mature semi.vpm file.
2. Your script file tells the lkb to load the semi.vpm file with the following line:

   (mt:read-vpm (lkb-pathname (parent-directory) "semi.vpm") :semi)
   

3. The semi.vpm files currently generated by the customization system are not fully ready for MT. In particular, the feature paths should include PNG. and E. only on the internal (left-hand) side. Thus you'll want to edit blocks like this:

   E.TENSE : E.TENSE
     present <> present
     past <> past
     future <> future
     * <> *
   

   To look like this:

   E.TENSE : TENSE
     present <> present
     past <> past
     future <> future
     * <> *
   

4. If your grammar uses a PERNUM feature, you'll need to map separate PER and NUM features from the external (right-hand side) of the VPM to a single PRENUM feature on the internal (left-hand side). See the example under "Properties: An Example" on the DELPH-IN wiki page.
5. If you have any other features you have added on indices, you will need to provide VPM entries for them as well. (If you added them through the customization system, they may be there already.)
6. If your language has aspect marked in some sentences but other forms that are just underspecified for aspect, you'll want to have the default aspect be "no-aspect". Define this as a subtype of aspect in your grammar, but don't have anything other than the semi.vpm mention it otherwise. In semi.vpm, replace * <> * under aspect with the following:

   * >> no-aspect
   no-aspect << [e]
   

7. If all forms in your language are marked for aspect, you may still want some default value in cases where the input MRS has no aspect, so that you don't get all the forms. If you wanted perfective as your defaul aspect, replace * <> * with:

   perfective << [e]