BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Cycling to Enrich and Enhance Dictionary Glosses - Roberto Navigli
 \, University of Rome &quot\;La Sapienza&quot\;
DTSTART:20090528T110000Z
DTEND:20090528T120000Z
UID:TALK18634@talks.cam.ac.uk
CONTACT:Johanna Geiss
DESCRIPTION:Machine readable dictionaries constitute an important source o
 f information\nin many fields of Natural Language Processing. It has been 
 argued that the\nrichness and quality of dictionaries greatly impacts the 
 performance of the\ntasks in which they are employed. In recent years\, mo
 nolingual approaches\nto the disambiguation of dictionary definitions have
  been investigated.\nHowever\, little attention has been paid to the disam
 biguation of bilingual\ndictionaries.\n\nIn this talk\, we present a novel
  algorithm for the disambiguation of\nmonolingual and bilingual dictionary
  glosses. The dictionary is represented\nas a graph and cyclic patterns ar
 e sought in the graph to assign the most\nappropriate sense to each gloss 
 word. We first report on the algorithm's\nperformance on monolingual dicti
 onaries and then focus on a two-fold\nexperimental evaluation of the appro
 ach on the Ragazzini-Biagi\nEnglish-Italian dictionary: first\, we show th
 at disambiguation performance\nachieves very high accuracy\, thus providin
 g a high-quality sense assignment\nfor most ambiguous translations in the 
 dictionary\; second\, we will see how\nthe algorithm's output can be used 
 to enhance the dictionary. Different\nautomatic suggestions can be produce
 d which aim to reduce the degree of\nlexical and semantic inconsistency wi
 thin the dictionary entries\,\nthus improving their quality.
LOCATION:SW01\, Computer Laboratory
END:VEVENT
END:VCALENDAR