BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning Hierarchical Translation Structure with Linguistic Annota
 tions - Markos Mylonakis\,  University of Amsterdam
DTSTART:20111011T120000Z
DTEND:20111011T133000Z
UID:TALK32771@talks.cam.ac.uk
CONTACT:Bill Byrne
DESCRIPTION:While it is generally accepted that many translation phenomena
  are correlated with linguistic structures\, employing linguistic syntax f
 or translation has proven a highly non-trivial task. The key assumption be
 hind many approaches is that translation is guided by the source and/or ta
 rget language parse\, employing rules extracted from the parse tree or per
 forming tree transformations. These approaches enforce strict constraints 
 and might overlook important translation phenomena that cross linguistic c
 onstituents. We propose a novel flexible modelling approach to introduce l
 inguistic information of varying granularity from the source side. Our met
 hod induces joint probability synchronous grammars and estimates their par
 ameters\, by selecting and weighing together linguistically motivated rule
 s according to an objective function directly targeting generalisation ove
 r future data. We obtain statistically significant improvements across 4 d
 ifferent language pairs with English as source\, mounting up to +1.92 BLEU
  for Chinese as target.
LOCATION: Cambridge University Engineering Department\,  Room LR10
END:VEVENT
END:VCALENDAR