Valency Effects on Dependency Grammar Induction
Speaker: David McClosky
Institution: Brown University
Abstract:
We extend the model of Klein and Manning (2004) to incorporate further valency information. Our extensions achieve significant improvements in the task of unsupervised dependency grammar induction. We use an expanded grammar which tracks higher orders of valency and allows each valency slot to be filled by a separate distribution rather than using one distribution for all slots. Additionally, we show that our performance improves if our grammar restricts the maximum number of attachments in each direction, forcing our system to focus on the common case. Taken together, these techniques constitute a 23.4% error reduction in dependency grammar induction over the model by Klein and Manning (2004) on English.