First Joint Workshop on Statistical Parsing of Morphologically Rich Languages and Syntactic Analysis of Non-Canonical Languages

Sunday August 24, 2014
9:00	Opening
9:05	Invited Talk: Universal Dependency Parsing (Joakim Nivre)  (slides)
	SPMRL
10:00	Parsing German: How Much Morphology Do We Need? Wolfgang Maier, Sandra Kübler, Daniel Dakota and Daniel Whyatt
10:30	(coffee break)
11:00	Joint Ensemble Model for POS Tagging and Dependency Parsing Iliana Simova, Dimitar Vasilev, Alexander Popov, Kiril Simov and Petya Osenova
11:30	Improving the parsing of French coordination through annotation standards and targeted features Assaf Urieli
12:00	Experiments with Easy-first nonprojective constituent parsing Yannick Versley
12:25	(lunch)
	SANCL
14:00	Exploring Options for Fast Domain Adaptation of Dependency Parsers Viktor Pekar, Juntao Yu, Mohab El-karef and Bernd Bohnet
14:30	Self-Training for Parsing Learner Text Aoife Cahill, Binod Gyawali and James Bruno
14:50	The effect of disfluencies and learner errors on the parsing of spoken learner language Andrew Caines and Paula Buttery
15:10	Poster Teasers
15:30	Coffee Break + Poster session (SPMRL short and Shared Task papers)
	SPMRL short papers
	Initial Explorations in Two-phase Turkish Dependency Parsing by Incorporating Constituents İlknur Durgar El-Kahlout, Ahmet Afşın Akın and Ertugrul Yılmaz
	Experiments for Dependency Parsing of Greek Prokopis Prokopidis and Haris Papageorgiou
	Shared Task Papers
New upload!	Combining Clustering Approaches for Parsing: the BASQUE TEAM system in the SPRML'2014 Shared Task Iakes Goenaga, Koldo Gojenola
New upload!	Multilingual discriminative shift reduce phrase structure parsing for the SPMRL 2014 shared task Benoit Crabbé and Djamé Seddah
New upload!	Incorporating Semi-supervised Features into Discontinuous Easy-first Constituent Parsing Yannick Versley
New upload!	Semi-supervised experiments at LORIA for the SPMRL 2014 Shared Task Christophe Cerisara
New upload!	Parsing Morphologically Rich Languages with (Mostly) Off-The-Shelf Software and Word Vectors Arne Köhn, U Chun Lao, Amir Ali B. Zadeh and Kenji Sagae
16:30	Shared Task session
New upload!	The IMS-Wrocław-Szeged-CIS Entry at the SPMRL 2014 Shared Task: Reranking and Morphosyntax Meet Unlabeled Data Anders Björkelund, Özlem Çetinoğlu, Agnieszka Fale\'nska, Richárd Farkas, Thomas Mueller, Wolfgang Seeker and Zsolt Szántó
16:55	Overview of the SPMRL 2014 Shared Task on Parsing Morphologically-rich Languages Djamé Seddah, Reut Tsarfaty, Sandra Kübler and and Marie Candito, Jinho Choi, Matthieu Constant, Richárd Farkas, Marie Candito, Jinho D. Choi, Matthieu Constant and , Richàrd Farkas and Iakes Goenaga, Koldo Gojenola, Yoav Goldberg, Spence Green, Nizar Habash, Marco Kuhlmann, Wolfgang Maier, Joakim Nivre, Adam Przepiorkowski, Ryan Roth, Wolfgang Seeker, Yannick Versley, Veronika Vincze, Marcin Wolinski, Alina Wroblewska and Eric Villemonte de la Clérgerie.
17:20	Open discussion panel : What's next for spmrl? Going deeper, further, wider?
18:15	Concluding remarks

Universal Dependency Parsing

Joakim Nivre, Uppsala University

Statistical parsers show wide variation in accuracy when applied to typologically different languages, suggesting that current parsing models make assumptions about syntactic structure that do not hold universally. As a long-term goal for parsing research I propose that we should seek to construct a universal parser, where knowledge of linguistic universals is built into the model and typological parameters are learned from data, inspired by the way humans acquire language. In this talk, I will present three pieces of work that are in different ways motivated by this long-term vision. First, I will review a recent study of joint morphological and syntactic disambiguation for morphologically rich languages. Second, I will present an ongoing project aiming to develop a universal standard for dependency annotation. Finally, I will speculate on how this universal annotation can be used to advance the state of the art in dependency parsing towards the goal of realizing a truly universal parser.