Parallel corpus pruning - Revision history

Francis Tyers at 13:47, 18 March 2009

2009-03-18T13:47:27Z

Francis Tyers at 13:28, 18 March 2009

2009-03-18T13:28:40Z

Francis Tyers at 13:27, 18 March 2009

2009-03-18T13:27:25Z

Francis Tyers at 13:18, 18 March 2009

2009-03-18T13:18:14Z

Francis Tyers at 13:17, 18 March 2009

2009-03-18T13:17:47Z

Francis Tyers: /* Possible ways of pruning */

2009-03-18T13:16:09Z

Possible ways of pruning

Francis Tyers at 13:13, 18 March 2009

2009-03-18T13:13:37Z

Francis Tyers at 11:26, 18 March 2009

2009-03-18T11:26:44Z

Francis Tyers at 11:17, 18 March 2009

2009-03-18T11:17:50Z

Francis Tyers at 14:21, 5 March 2009

2009-03-05T14:21:28Z

@@ Line 23: / Line 23: @@
 The subject "Guð" is moved, the Icelandic uses relatives, where the Faroese doesn't.
+An equally valid, but more literal translation in Icelandic would be:
+* is: 7  '''Guð gjörði þá festinguna og greindi vötnin undir festingunni frá vötnum yfir henni. Og svo varð.'''
+Which is exactly the same word order as the Faroese.
 ==Possible ways of pruning==

@@ Line 14: / Line 14: @@
 Is good, there is a one-to-one- correspondence and the words are in the same order. If stuff is moved around for stylistic reasons, for example
-* fo: 7  Gud gjørdi tá hvølvið og skilti vatnið undir hvølvinum frá vatninum yvir hvølvinum. Og so varð.
+* fo: 7  '''Gud gjørdi tá hvølvið og skilti vatnið undir hvølvinum frá vatninum yvir hvølvinum. Og so varð.'''
 ::God made then firmament-DEF and divided waters-DEF under firmament-DEF from waters-DEF above firmament-DEF: and so was.
-* is: 7  Þá gjörði Guð festinguna og greindi vötnin sem voru undir festingunni frá þeim vötnum sem voru yfir henni. Og það varð svo.
+* is: 7  '''Þá gjörði Guð festinguna og greindi vötnin sem voru undir festingunni frá þeim vötnum sem voru yfir henni. Og það varð svo.'''
 ::Then made God firmament-DEF and divided waters-DEF which were under firmament-DEF from waters-DEF which were above firmament-DEF: And so was so.

@@ Line 15: / Line 15: @@
 * fo: 7  Gud gjørdi tá hvølvið og skilti vatnið undir hvølvinum frá vatninum yvir hvølvinum. Og so varð.
+::God made then firmament-DEF and divided waters-DEF under firmament-DEF from waters-DEF above firmament-DEF: and so was.
 * is: 7  Þá gjörði Guð festinguna og greindi vötnin sem voru undir festingunni frá þeim vötnum sem voru yfir henni. Og það varð svo.
+::Then made God firmament-DEF and divided waters-DEF which were under firmament-DEF from waters-DEF which were above firmament-DEF: And so was so.
-where the subject "Guð" is moved (among other differences) it isn't so good.
+The subject "Guð" is moved, the Icelandic uses relatives, where the Faroese doesn't.
 ==Possible ways of pruning==

@@ Line 10: / Line 10: @@
 * is: 1  Í upphafi skapaði Guð himin og jörð.
-:"In beginning created God heavens and earth"
+::"In beginning-DEF created God heavens and earth"
 Is good, there is a one-to-one- correspondence and the words are in the same order. If stuff is moved around for stylistic reasons, for example

@@ Line 9: / Line 9: @@
 * fo: 1  Í upphavi skapti Gud himmal og jørð.
 * is: 1  Í upphafi skapaði Guð himin og jörð.
+:"In beginning created God heavens and earth"
 Is good, there is a one-to-one- correspondence and the words are in the same order. If stuff is moved around for stylistic reasons, for example

@@ Line 21: / Line 21: @@
 * One of the ways would be to discard phrases which can't be produced by the MT system (presumably in terms of lemma matches).
 * Look at the ratio of unaligned words. The higher the ratio, the more likely the translation is "freer"
+* Look at the fertility (between close languages lower)
 * ...

@@ Line 1: / Line 1: @@
 This page describes some ideas of how to prune parallel corpora to retrieve the most '''literally''' translated aligned segments. Ony of the obstacles to inducing effective transfer rules from parallel corpora is lack of literalness in translation. For example, for a given aligned segment, it may not have been translated between languages, but rather both from a single source language.
+* <u>Una de las personas</u> que recientemente han asesinado en Sri Lanka ha sido al Sr . Kumar Ponnambalam , <u>quien hace pocos meses visitó el Parlamento Europeo .</u>
-<pre>
+* <u>Una delle vittime</u> più recenti è stato Kumar Ponnambalam , che qualche mese fa era venuto in visita qui al Parlamento europeo .
-Una de las personas que recientemente han asesinado en Sri Lanka ha sido al Sr . Kumar Ponnambalam , quien hace pocos meses visitó el Parlamento Europeo .
+* <u>Uma das pessoas</u> recentemente assassinadas foi o senhor Kumar Ponnambalam,  <u>que ainda há poucos meses visitara o Parlamento Europeu</u>.
-Una delle vittime più recenti è stato Kumar Ponnambalam , che qualche mese fa era venuto in visita qui al Parlamento europeo .
-Uma das pessoas recentemente assassinadas foi o senhor Kumar Ponnambalam, que ainda há poucos meses visitara o Parlamento Europeu.
-</pre>
 When making rules, we're more interested in "literal" translations, so for example:
@@ Line 19: / Line 15: @@
 * is: 7  Þá gjörði Guð festinguna og greindi vötnin sem voru undir festingunni frá þeim vötnum sem voru yfir henni. Og það varð svo.
-where the subject "Guð" is moved it isn't so good.
+where the subject "Guð" is moved (among other differences) it isn't so good.
 ==Possible ways of pruning==
@@ Line 25: / Line 21: @@
 * One of the ways would be to discard phrases which can't be produced by the MT system (presumably in terms of lemma matches).
 * Look at the ratio of unaligned words. The higher the ratio, the more likely the translation is "freer"
-*
+* ...
 [[Category:Discussions]]

← Older revision		Revision as of 14:21, 5 March 2009
Line 10:		Line 10:


	[[Category:~~Discussion~~]]		[[Category:Discussions]]