BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A New Twist on Methodologies for ESL Grammatical Error Detection -
  Joel Tetreault\, Educational Testing Service
DTSTART:20121012T110000Z
DTEND:20121012T120000Z
UID:TALK39568@talks.cam.ac.uk
CONTACT:Ekaterina Kochmar
DESCRIPTION:The long-term goal of our work is to develop a system which de
 tects errors\nin grammar and usage so that appropriate feedback can be giv
 en to\nnon-native English writers\, a large and growing segment of the wor
 ld's\npopulation. Estimates are that in China alone as many as 300 million
  people\nare currently studying English as a second language (ESL). In par
 ticular\,\nusage errors involving prepositions are among the most common t
 ypes seen in\nthe writing of non-native English speakers. For example\, Iz
 umi et al.\,\n(2003) reported error rates for English prepositions that we
 re as high as\n10% in a Japanese learner corpus.\n\nSince prepositions are
  such a nettlesome problem for ESL writers\,\ndeveloping an NLP applicatio
 n that can reliably detect these types of\nerrors will provide an invaluab
 le learning resource to ESL students. In\nthis talk we first review one po
 pular machine learning methodology for\ndetecting preposition and article 
 errors in texts written by ESL writers.\nNext\, we describe a novel approa
 ch to ESL grammatical error detection:\nusing round-trip machine translati
 on to automatically correct errors.\n\nThis is joint work with Nitin Madna
 ni (ETS) and Martin Chodorow (CUNY).\n\nSpeaker Bio:\n\nJoel Tetreault is 
 a Managing Senior Research Scientist specializing in\nComputational Lingui
 stics in the Research & Development Division at\nEducational Testing Servi
 ce in Princeton\, NJ. His research focus is Natural\nLanguage Processing w
 ith specific interests in anaphora\, dialogue and\ndiscourse processing\, 
 machine learning\, and applying these techniques to\nthe analysis of Engli
 sh language learning and automated essay scoring.\nCurrently he is working
  on automated methods for detecting grammatical\nerrors by non-native spea
 kers\, plagiarism detection\, and content scoring\nmethods. Previously\, h
 e was a postdoctoral research scientist at the\nUniversity of Pittsburgh's
  Learning Research and Development Center\n(2004-2007). There he worked on
  developing spoken dialogue tutoring\nsystems. Tetreault received his B.A.
  in Computer Science from Harvard\nUniversity (1998) and his M.S. and Ph.D
 . in Computer Science from the\nUniversity of Rochester (2004).
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR