UC-B1

Revision as of 20:14, 5 September 2012

The CorpusA2 is an experimental corpus used to refine the initial versions of the grammar for sentence-based UNLization and NLization, using IAN and EUGENE, respectively. It comprises 5 very short texts in English and their corresponding graphs in UNL.

The Corpus

The CorpusA2 consists of 5 texts: the first one (Text 1) will be used during the workshop; the other four will be used for evaluation and certification. The five texts are translations of Aesop's fables to English. Most of them have been derived from the standard version by George Fyler Townsend (available at The Project Gutenberg), but they have suffered slight changes in order to become more suitable for natural language processing.

*To be manually translated to your target language in order to be used as the input for UNLization (IAN)
**To be used as the input for NLization (EUGENE)

Pre-workshop tasks

The pre-workshop activities involve only Text 1. Participants in the workshop are expected to:

• Translate the Text 1 into their own native languages
• Provide the natural language analysis dictionary corresponding to the translated version of Text 1
• Provide the natural language generation dictionary corresponding to the UNL version of the Text 1

These activities should be concluded before September 30th 2012. The corresponding files must be sent by e-mail to r.martins@undlfoundation.org

Goals

1. To provide the dictionary and grammars necessary to UNLize your translated version of the CorpusA2 (with IAN)
2. To provide the dictionary and grammars necessary to NLize, to your target language, the UNL version of the CorpusA2 (with EUGENE)

Methodology

In order to prepare the dictionaries and grammars to deal with the CorpusA2, follow the steps below for each text:

UNLization

1. Corpus
   1. Translate (manually) the sentences of the CorpusA2 from English into your native language. Be as close as possible to the original, and provide one single translation for each sentence. This will be your input document file, and your goal will be to provide (automatically, through IAN) the UNL graphs for each sentence
   2. Save the translated texts (without the English original) in a plain text (.txt) file with UTF-8 encoding and upload it to UNLWEB>UNLDEV>IAN>NL FILES.
2. Dictionary
   1. Create the analysis dictionary corresponding to the word forms appearing in your translated version of the corpus. The dictionary must be created in compliance with the Dictionary Specs and may contain only tags available at the tagset. For an example of the English dictionary, see English Dictionary.
   2. Save the NL-UNL dictionary in a plain text (.txt) file with UTF-8 encoding and upload it to UNLWEB>UNLDEV>IAN>DICTIONARIES.
3. Grammar
   1. Create the NL-UNL (analysis) transformation grammar necessary to analyze, in UNL, the natural language sentences of the translated corpus. This grammar is the most difficult (and the actual goal) of the whole analysis task. In order to prepare the grammar, study the Grammar Specs. Next, take a look at the structure of the English Grammar for a detailed example. In many cases, it would be simpler just to localize the English grammar to your own locale rather than creating a whole grammar from the scratch. See the instructions at Localization. The grammar is normally developed inside IAN, because we have to test it every time. But you may also prepare a draft grammar, save it in a plain text (.txt) file with UTF-8 encoding and upload it to UNLWEB>UNLDEV>IAN>T-RULES.
   2. Create the NL-UNL (analysis) disambiguation grammar in order to improve the results of the transformation and to control the process of tokenization. Again, this grammar is normally developed inside IAN (at the D-RULES tab), because we have to test it every time. Consult the English Grammar for a detailed example of a disambiguation grammar.
   3. Test the grammar against the corpus and provide the necessary changes.

NLization

1. Corpus
   1. Upload the file corpusA1_unl.txt to UNLWEB>UNLDEV>EUGENE>UNL DOCUMENT. Do not do any change to this file.
2. Dictionary
   1. Create the generation dictionary corresponding to the word forms appearing in your translated version of the corpus. The dictionary must be created in compliance with the Dictionary Specs and may contain only tags available at the tagset. For an example of the English dictionary, see English Dictionary.
   2. Save the UNL-NL dictionary in a plain text (.txt) file with UTF-8 encoding and upload it to UNLWEB>UNLDEV>PROJECTS>EUGENE>DICTIONARIES.
3. Grammar
   1. Export the inflectional grammar of your language from UNLARIUM>GRAMMAR>[YOUR LOCALE]>EXPORT. If the grammar of your language is not available yet, you may:
      1. Provide it through the UNLarium; or
      2. Create the inflectional paradigms only for the inflected forms appearing in the UNL-NL dictionary. In that case, follow the model available at English Inflectional Grammar. The documentation of the English grammar is available at English Inflectional Grammar (only for reference). For further information, see Inflectional paradigms.
   2. Save the inflectional grammar in a plain text (.txt) file with UTF-8 encoding and upload it to UNLWEB>UNLDEV>PROJECTS>EUGENE>RULES.
   3. Provide the UNL-NL (generation) T-grammar and D-grammar necessary to generate natural language sentences from the UNL corpus. Follow the sample instructions available for the analysis grammar.

Samples and Examples

To be provided

Recommended Readings

Before starting the activity, and in order to fully understand what is expected to be done, it is important for you to be acquainted with the following documentation:

• Tagset, because you are expected to use only the tags included in the tagset
• UNL Dictionary Specs, which is essential to understand the dictionary structure
• UNL Grammar Specs, which is essential to understand the grammar structure

It is also interesting to make a test drive with IAN and EUGENE.