Difference between revisions of "Apertium separable/report2017"

From Apertium
Jump to navigation Jump to search
Line 26: Line 26:
==Future work==
==Future work==
See [[Lsx_module#Future_work]].
* 10:53 firespeaker: pektii: if we offload multiwords from the transducers to lsx, (1) how do we do N N compounds with lsx? (2) how does translation *to* a multiword work?
* recycling dictionaries and/or paradigms? lsx-dictionaries are packaged in language pairs. the eng-spa lsx-dictionary can mostly be reaped by eng-cat. could we make use of the similarity?
* Support for language pairs: we haven't gotten much extensive beta testing. The following are language pairs that have packaged the lsx-module:
** eng-cat
** eng-deu (?)

Revision as of 22:04, 31 August 2017

Project description

The purpose of this project is to allow Apertium language-pair developers to better translate "seperable" or "discontiguous" multiwords. We do this by re-ordering word tokens before translation occurs. For example, "take something out" becomes "take out something" so that "take out" can be translated as a single unit.

To do this, a finite-state transducer was used. The transducer accepted certain patterns of words (paradigms), such as adj-noun or det-adj-noun, that could separate the multiword. If the pattern was accepted, then the transducer would output the re-ordered words for better translation quality.

Work done

  • all spacing, punctuation, and superblanks were preserved
  • support for the "plus thing":
echo "^абай<adj>$ ^бол<v><iv><neg><aor><p2><sg>+ма<qst>$^.<sent>$" | lsx-proc kaz-kir.autoseq.bin
^абай бол<v><iv><neg><aor><p2><sg>+ма<qst>$^.<sent>$
  • (for language developers: have the language-data writer write it explicitly in the .lsx file)

Future work

See Lsx_module#Future_work.