Final Thoughts

Abstract

Our aim in this book was to provide an in-depth account of the current practices and trends in dependency parsing.We saw these as data-driven parsers,which includes transition-based and graphbased parsing, grammar-based parsers, which includes context-free and constraint-based formalisms, and large-scale multi-lingual parser evaluation. We complemented this with an account of the commonalities and differences between each kind of parser in terms of their algorithmic makeup, theoretical properties, and empirical performance when appropriate studies were available. The approaches covered within these pages certainly does not represent the entire spectrum of research on dependency parsing, but we hope that the further reading sections provided in each chapter as well as the appendix of resources and the extensive bibliography that follow will fill out the remaining gaps. Furthermore, this is a burgeoning research area with a number of new studies being published every year.We have attempted to incorporate the latest in the field, but the reader is of course encouraged to check the latest conference proceedings and journal issues to stay abreast of the developments.