BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian Models for Dependency Parsing Using Pitman-Yor Priors - H
 anna Wallach (University of Cambridge)
DTSTART:20080617T101500Z
DTEND:20080617T111500Z
UID:TALK12109@talks.cam.ac.uk
CONTACT:David MacKay
DESCRIPTION:In this talk\, I will introduce a Bayesian dependency parsing 
 model for natural language\, based on the hierarchical Pitman-Yor process.
  This model arises from a Bayesian reinterpretation of a classic dependenc
 y parser. I will show that parsing performance can be substantially improv
 ed by (a) using a hierarchical Pitman-Yor process as a prior over the dist
 ribution over dependents of a word\, and (b) sampling model hyperparameter
 s. Finally\, I will present a second Bayesian dependency model in which la
 tent state variables mediate the relationships between words and their dep
 endents. This model clusters parent-child dependencies into states using a
  similar approach to that employed by Bayesian topic models when clusterin
 g words into topics. Each latent state may be viewed as a sort of speciali
 sed part-of-speech tag or "syntactic topic" that captures the relationship
 s between words and their dependents.
LOCATION:TCM Seminar Room\, Cavendish Laboratory\, Department of Physics
END:VEVENT
END:VCALENDAR