Data-driven parsing using probabilistic linear context-free rewriting systems
Computational Linguistics
| Hi-index | 0.00 |
The idea of organizing this workshop was sparked following very interesting discussions that occurred during EACL09 among various researchers working on statistical parsing of different types of languages. Indeed, an opportunity to discuss the issues that we were all experiencing was much needed, and it seemed such a good idea that we decided to take advantage of IWPT'09, which was held that year in Paris, to organize a panel on this topic. We planned to have presentations on the various issues faced by this small emerging community, which would allow us to share our sometimes similar solutions for parsing different languages. Inspired by the idea of organizing such a meeting, but without knowing quite yet if there was any sense in comparing, for example, Modern Hebrew and French parsing issues, Deirdre Hogan (Dublin City University) suggested that - should the panel be successful - we ought to organize a real workshop. She was right. We had an extremely successful and animated panel discussion. We were surprised to see the extent to which the IWPT'09 audience chose to contribute to the discussion instead of taking a break from the long presentation sessions. This encouraged us to pursue these attempts at providing a forum for discussing such matters even further, and to create a new community of shared interests. This workshop is the result of our common will to do so. We believe that the issues faced by researchers involved in statistical parsing of morphologically rich languages are not always well known outside of this small community, and that the kind of challenges that we all face require a more thorough introduction than we could possibly provide in this foreword. Therefore, we decided to include here an elaborated preface which presents the current state-of-affairs with respect to parsing MRLs and frames the various contributions to our workshop in relation to it. The overview should act as a primer for those who are not experienced in the subject and yet wish to participate in the discussion. All in all, we are proud to have 11 very nice papers presented in our proceedings that will help advance the state of the art in parsing MRLs. In order to obtain sufficient presentation slots, we asked our authors to choose between different modes of presentation, we are glad the authors involved in 3 papers accepted to present them as posters.