Institute

Studying

Research


 

Nina Seemann

Misses  Dipl.-Linguistin
Nina Seemann

Nina Seemann
Phone 0049 711 685-84580
Room01.009
E-Mail
Address
Universität Stuttgart
Institut für Maschinelle Sprachverarbeitung
Pfaffenwaldring 5b
70569 Stuttgart
Deutschland

Consultation

Thursdays from 10:00h to 11:00h

I'm a PhD student in Andreas Maletti's project Tree Transducers in Machine Translation. In this project, we integrated the support for shallow local multi bottom-up tree transducers (MBOT) in the open source toolkit Moses. Furthermore, we developed various settings for MBOT to improve translation quality. Download and usage information can be found here.

Additionally, I worked in the  Morphosyntax project where we evaluated the integration of a classifier in string-to-tree rule selection.


Projects

 

  • Tree Transducers in Machine Translation where we investigate tree-based formalisms for statistical machine translation. We have implemented various training procedures for the extraction of shallow local multi bottom-up tree transducers and intregrated our approach into the Moses Open Source Toolkit.
Teaching

winter term 2014/2015

  • lecture with exercise Parsing I

summer term 2014

  • seminar Formal Models in NLP (Weighted and Unweighted Finite-State Automata)

winter term 2013/2014

  • lecture with exercise Parsing I

summer term 2013

  • seminar Formal Models in NLP (Weighted and Unweighted Finite-State Automata)
  • course Statistical Machine Translation (IMB model 1 and practical Moses Lab)

 summer term 2012

  • seminar Formal Models in NLP (Unweighted Finite-State Automata)

 

Publications

 Conferences:

 Workshops:

  • Nina Seemann & Andreas Maletti. Discontinuous Statistical Machine Translation with Target-Side Dependency Syntax. Proc. Workshop on Machine Translation, Lisboa 2015.
  • Fabienne Braune, Nina Seemann, Daniel Quernheim & Andreas Maletti. Machine Translation with Multi Bottom-up Tree Transducers. 35. Jahrestagung der Gesellschaft für Deutsche Sprache, Potsdam 2013.
  • Nina Seemann, Daniel Quernheim, Fabienne Braune & Andreas Maletti. Preservation of Recognizability for Weighted Linear Exended Top-Down Tree Transducers. Proc. EACL 2012 Workshop on Applications of Tree Automata Techniques in Natural Language Processing, France 2012.
  • Arndt Riester, Kerstin Eckart, Katrin Schweitzer & Nina Seemann. Pitch Accents on Short Referential Expressions in German Radio News. The Prosody-Discourse Interface (IDP). University of Salford, UK 2011.
  • Kerstin Eckart, Arndt Riester, Katrin Schweitzer & Nina Seemann. Querying Information Status and Prosody Annotations: The Problem of Deviating Primary Data. 33. Jahrestagung der Gesellschaft für Deutsche Sprache, Göttingen 2011.

 Journals:

  

Supervision

completed:

  • Olga Podushko, M.Sc., Master thesis
  • Irina Popova, M.Sc., Master thesis
  • Robin Kurtz, M.Sc., Master thesis