Dependency Parsing by Inference over High-recall Dependency Predictions

As more and more syntactically-annotated corpora become available for a wide variety of languages, machine learning approaches to parsing gain interest as a means of developing parsers without having to repeat some of the labor-intensive and language-specific activities required for traditional parser development, such as manual grammar engineering, for each new language. The CoNLL-X shared task on multi-lingual dependency parsing (Buchholz et al., 2006) aims to evaluate and advance the state-of-the-art in machine learning-based dependency parsing by providing a standard benchmark set comprising thirteen languages. In this paper, we describe two different machine learning approaches to the CoNLL-X shared task.