Parser-independent semantic tree alignment

We describe an approach for training a semantic role labeler through cross-lingual projection between different types of parse trees, with the purpose of enhancing tree alignment on the level of syntactic translation divergences. After applying an existing semantic role labeler to parse trees in a resource-rich language (English), we partially project the semantic information to the parse trees of the corresponding target sentences, based on word alignment. After this precision-oriented projection, we apply a method for training a semantic role labeler which consists in determining a large set of features describing target predicates, roles and predicate-role connections, independently from the type of tree annotation (phrase structure or dependencies). These features describe tree paths starting at or connecting nodes. The semantic role labeling method does not require any knowledge of the parser nor manual intervention. We evaluated the performance of the cross-lingual projection and semantic role labeling using an English parser assigning PropBank labels and Dutch manually annotated parses, and are currently studying ways to use the predicted semantic information for enhancing tree alignment.status: publishe