Alignment all to contain all alignment points for all

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: not slap inconsistent X inconsistent Phrase §  Extract has such phrase pairs! alignment all to contain all alignment points for all covered words Phrase alignment has to contain at least one alignment point 13 February Phrase Pair Extraction Example Word alignment induced phrase Word alignment induced phra (Maria, Mary), (no, did not), (slap, bofetada bofetada bruja bofetada bruja daba una bofetada), (a la, the), Maria no daba una a verde Maria no daba una a la la verde (bruja, witch), (verde, green) Mary Mary (Maria no, Mary did not), (no daba una bofetada, did not slap), (daba did did una bofetada a la, slap the), (bruja not not verde, green witch) (Maria no daba una bofetada, Mary slap slap did not slap), (no daba una bofetada the the a la, did not slap the), (a la bruja verde, the green witch) " green green (Maria no daba una bofetada a la, witch witch Mary did not slap the), (daba una bofetada a la bruja verde, slap the green witch)" Maria, Mary),no(no, did bofetada (slap, daba una bofetada), (a la, the), (bruja, witc Maria, Mary), (no, una not), (slap, daba una bofetada), (a la, the), (bruja, wit (Maria daba did not), a la bruja verde, Mary did not slap the (Maria no, Mary did not), (no daba una bofetada, did not slap), (daba una bof (Maria greenMary did not), (no daba una bofetada, did not slap), (daba una bo no, witch)" bofetada, did not slap), (daba una bo bo (bruja verde, green witch), (Maria no daba una bofetada, Mary did not slap), (bruja verde, green witch) (no daba una bofetada a la, did not slap the), (a la bruja verde, the green witc green witc (Maria no daba una bofetada a la, Mary did not slap the), (daba una bofetad Phrase Size §  Phrases do help §  But they don’t need to be long §  Why should this be? Bidirectional Alignment Alignment Heuristics Phrase Scoring c(e, f ) g(f, e) = log c(e) g(les chats, cats) = log c(cats, les chats) c(cats) aiment poisson les chats le frais cats like fresh fish . . . §  Learning weights has been tried, several times: §  [Marcu and Wong, 02] §  [DeNero et al, 06] §  … and others §  Seems not to work well, for a variety of partially understood reasons §  Main issue: big chunks get all the weight, obvious priors don’t help §  Though, [DeNero et al 08] Scoring: §  Basic approach, sum up phrase translation scores and a language model §  §  §  §  §  Define y = p1p2…pL to be a translation with phrase pairs pi Define e(y) be the out...
View Full Document

This document was uploaded on 04/04/2014.

Ask a homework question - tutors are online