Intensivwoche 2014

Datum: Montag, 9. Juni bis Freitag, 13. Juni 2014

Ort: Centre of Computational Linguistics, Vytautas Magnus Universität, Kaunas, Litauen

Thema: Mehrwortausdrücke in der Sprachtechnologie

Teilnehmende: Masterstudierende der Multilingualen Textanalyse und der Computerlinguistik bzw. Computerlinguistik und Sprachtechnologie

Leitung: Prof. Martin Volk und Magdalena Plamada, Universität Zürich

Programm:

9. Juni

10. Juni

11. Juni

12. Juni

13. Juni

09:30- 10:45

Introduction of the Institutes in Kaunas, Zurich & Stockholm

Student presentations 1

Student presentations 2

Student presentations 3

Company Visit in Vilnius: Tilde

11:00- 12:15

Talk: MWEs in Lithuanian (Jolanta Kovalevskaitė and Loic Boizou)

Talk: MWEs in MT (Martin Volk)

Talk: Introduction to MWEs and Construction Grammar (Gintarė Grigonytė)

Talk: NLP in IR (Tomas Krilavičius)

Lunch

14:00- 16:00

Group Task 1: Manual Annotation of MWEs on small Parallel Treebanks (Fabienne Leuenberger)

Group Task 2: Automatic Annotation of MWEs

Final Round of Discussions

- open -

16:30- 17:30

Reports by the groups

Departure to Vilnius

- open -

Student presentations 1: MWEs in MT

Dolores Batinic: Evaluation of English idiomatic expressions in online MT systems
Julia Luchsinger:Evaluation of German idiomatic expressions in online MT systems
Hernani Marques: The handling of multiword expressions (French - Portuguese) in the rule - based MT system Apertium

Student presentations 2: Searching and Visualizing MWEs

Justina Mandravickaitė: Visualization of multiword idiomatic expressions (clichés) in the Lithuanian news portal Delfi.lt headlines
Danica Pajovic: Search and annotation of French idiomatic expressions in the Text+Berg corpus

Student presentations 3: MWEs in various Languages and Applications

Nicolas Fischer: A comparison of multiword expressions in a parallel corpus of Old - Russian vs. Modern - Russian
Reto Baumgartner: A comparison of verb particles in participles of Icelandic vs. Swedish
Kornélia Papp: Automatic filtering of user generated text