tailieunhanh - Báo cáo khoa học: "Using Parse Features for Preposition Selection and Error Detection"

We evaluate the effect of adding parse features to a leading model of preposition usage. Results show a significant improvement in the preposition selection task on native speaker text and a modest increment in precision and recall in an ESL error detection task. Analysis of the parser output indicates that it is robust enough in the face of noisy non-native writing to extract useful information. | Using Parse Features for Preposition Selection and Error Detection Joel Tetreault Educational Testing Service Princeton NJ USA JTetreault@ Jennifer Foster NCLT Dublin City University Ireland jfoster@ Martin Chodorow Hunter College of CUNY New York NY USA @ Abstract We evaluate the effect of adding parse features to a leading model of preposition usage. Results show a significant improvement in the preposition selection task on native speaker text and a modest increment in precision and recall in an ESL error detection task. Analysis of the parser output indicates that it is robust enough in the face of noisy non-native writing to extract useful information. 1 Introduction The task of preposition error detection has received a considerable amount of attention in recent years because selecting an appropriate preposition poses a particularly difficult challenge to learners of English as a second language ESL . It is not only ESL learners that struggle with English preposition usage automatically detecting preposition errors made by ESL speakers is a challenging task for NLP systems. Recent state-of-the-art systems have precision ranging from 50 to 80 and recall as low as 10 to 20 . To date the conventional wisdom in the error detection community has been to avoid the use of statistical parsers under the belief that a WSJ-trained parser s performance would degrade too much on noisy learner texts and that the traditionally hard problem of prepositional phrase attachment would be even harder when parsing ESL writing. However there has been little substantial research to support or challenge this view. In this paper we investigate the following research question Are parser output features helpful in modeling preposition usage in well-formed text and learner text We recreate a state-of-the-art preposition usage system Tetreault and Chodorow 2008 henceforth T C08 originally trained with lexical features and augment it .

TỪ KHÓA LIÊN QUAN