BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Unsupervised Multilingual Learning for Morphological Segmentation 
 - Tom Lippincott
DTSTART:20100202T123000Z
DTEND:20100202T133000Z
UID:TALK23048@talks.cam.ac.uk
CONTACT:Diarmuid Ó Séaghdha
DESCRIPTION:At this session of the NLIP Reading Group we’ll be discussin
 g the following paper:\n\nBenjamin Snyder and Regina Barzilay. 2008. "Unsu
 pervised Multilingual Learning for Morphological Segmentation":http://aclw
 eb.org/anthology-new/P/P08/P08-1084.pdf. In Proceedings of ACL-08.\n\n*Abs
 tract:*\n For centuries\, the deep connection between languages has brough
 t about major discoveries about human communication. In this paper we inve
 stigate how this powerful source of information can be exploited for unsup
 ervised language learning. In particular\, we study the task of morphologi
 cal segmentation of multiple languages. We present a nonparametric Bayesia
 n model that jointly induces morpheme segmentations of each language under
  consideration and at the same time identifies cross-lingual morpheme patt
 erns\, or abstract morphemes. We apply our model to three Semitic language
 s: Arabic\, Hebrew\, Aramaic\, as well as to English. Our results demonstr
 ate that learning morphological models in tandem reduces error by up to 24
 % relative to monolingual models. Furthermore\, we provide evidence that o
 ur joint model achieves better performance when applied to languages from 
 the same family.
LOCATION:GS15\, Computer Laboratory
END:VEVENT
END:VCALENDAR