Literality Based Sample Sorting for Syntax Projection
Bruno Cavestro and Nicola Cancedda
In: Cross-Language Knowledge Induction Workshop, 2-4 Aug 2005, Cluj-Napoca, Romania.
We consider the problem of projecting syntax trees across different sides of a parallel corpus, without using any language dependent feature. To achieve this task we introduce a literality score and use it to sort the bi-sentences of the parallel corpus in different classes. We show how to iteratively train a parser over those classes.