CoNLL 2009: Semantic role labeling (Spanish)

The CoNLL-2009 shared task combined dependency parsing and the task of identifying and labeling semantic arguments of verbs (and other parts of speech whenever available). Participants had to choose from two tasks: joint task (syntactic dependency parsing and semantic role labeling), or  semantic role labeling only task. The semantic role labeling task consists of three steps: disambiguating the sense of the predicate, finding the arguments of the predicate and assigning a semantic role to the arguments. For this task syntactic dependency parses are provided by the organizers, which were obtained using state-of-the art parsers. In the test data predicates are marked, but predicate disambiguation is still part of the task (determining the sense of the predicate).