index

Statistical Language Processing and Learning Lab.
Part of Language and Computation, ILLC, FNWI, UvA.
Contact: Dr Khalil Sima'an

Our research concentrates on structured language processing with application to morpho-syntactic parsing and hierarchical machine translation. Our work aims at inducing and exploiting the latent structure that represents relevant salient regularities in natural language data (mono- and multi-lingual corpora) for improved language applications. Sample recent topics (see, e.g., publications or the individual pages of the group members):

effective syntax-based models for machine translation, including syntactic language models, syntactic and latent reordering models, synchronous grammar learning (with H. Hassan (DCU), A. Way (DCU), M. Mylonakis and M. Khalilov),
representational and computational issues in parsing morphologically-rich languages (mainly Semitic - with Reut Tsarfaty, Yoad Winter (now U. Utrecht), Roy Bar-Haim (Technion) and Alon Itai (Technion)),
computational aspects (algorithm and complexity) and statistical learning for Data-Oriented Parsing (with R. Scha, R. Bod, A. Zollmann, L. Buratto and D. Prescher)
inducing rich syntactic lexica from a mix of unannotated and annotated data (with T. Deoskar and M. Mylonakis),

Researchers

Khalil Sima'an (ILLC, UvA)		Principal Investigator

Markos Mylonakis (UvA, NWO VIDI project)	2007-11	PhD student
Gideon Wenniger (UvA, Open Competitie project)	2010-14	PhD student
Sophie Arnoult	2011-12	Research assistant

Alumni Ph.D. students

19 Jan. 2012	Markos Mylonakis (UvA, NWO VIDI): graduation 19 January 2012	Xerox Research Centre Europe
Graduated Mar. 2010	Reut Tsarfaty (UvA, NWO MOSAIEK project): graduated 24 March 2010	Uppsala University
Graduated Jan. 2009	Hany Hassan co-supervision together with Andy Way at Dublin City University, Dublin, Ireland.	Microsoft Research

Alumni postdocs

Detlef Prescher (2003-2006 -- Project LeStoGram Consistent Statistical Estimation from Treebanks)
Yoav Seginer -- substituting lecturer for my statistical NLP courses (2007--2008)
Tejaswini Deoskar -- 50% substituting lecturer and 50% postdoc (2008-10, VIDI project, NWO)
Maxim Khalilov (2009-2011) Postdoc on the VIDI project

Ongoing projects:

Machine Translation When Exact Pattern Match Fails (NWO-EW - Open Competitie) (2010-2014), PI=Khalil Sima'an (226k Euro)
Priors for Estimating Probabilistic Grammars from Incomplete Data (NWO-EW - VIDI) [2007-2011]
Innovative Research Incentive, Netherlands Organisation for Scientific Research (NWO) - Exact Sciences (766 kEuro), PI=Khalil Sima'an
Modern Hebrew and Arabic Processing [2000-present] NWO-Mosiaek Reut Tsarfaty, PI=Khalil Sima'an (approx. 200 kEuro)
People involved: Khalil Sima'an (UvA), Reut Tsarfaty (UvA) and Hebrew Project (Technion).
Graduates: Roy-Bar Haim (Technion, Haifa), Saib Mansour (Technion, Haifa)
Collaboration: 2005 Johns Hopkins University (JHU) Summer Workshop 2005 on Parsing Arabic Dialect.
In March 2010, Tsarfaty defended her Ph.D. thesis.

Concluded projects:

Learning Stochastic Tree-Grammars from Treebanks (LeStoGram) NWO-EW Open Competitie [2003-2006], PI=Khalil Sima'an (approx. 200 kEuro)
Project was concluded in October 2006
Beyond Treebank Annotations: Ambiguity Resolution by Similarity-Based Performance Models Personal innovation grant (KNAW Fellowship, Royal Dutch Academy for Arts and Sciences), awarded in 2002. (approx 200 kEuro). PI=Khalil Sima'an.
Project was concluded in 2003