Paper: A log-linear model with an n-gram reference distribution for accurate HPSG parsing

ACL ID W07-2208
Title A log-linear model with an n-gram reference distribution for accurate HPSG parsing
Venue Conference on Parsing Technologies
Session Main Conference
Year 2007
Authors

This paper describes a log-linear model with an n-gram reference distribution for accurate probabilistic HPSG parsing. In the model, the n-gram reference distribution is simply defined as the product of the probabilities of selecting lexical entries, which are pro- vided by the discriminative method with ma- chine learning features of word and POS n-gram as defined in the CCG/HPSG/CDG supertagging. Recently, supertagging be- comes well known to drastically improve the parsing accuracy and speed, but su- pertagging techniques were heuristically in- troduced, and hence the probabilistic mod- els for parse trees were not well defined. We introduce the supertagging probabilities as a reference distribution for the log-linear model of the probabilistic HPSG. This is the first model which proper...