New Media and Language Technologies

Part of "New Media and eScience" MSc Programme
Jožef Stefan International Postgraduate School
Winter 2008 / Spring 2009

URL http://nl.ijs.si/et/teach/mps08-hlt/

Lecturers

• Tomaž Erjavec
• Sašo Džeroski
• and visiting lecturers

Course timetable and materials

December 3 2008 15:15 - 19:00, MPŠ

1. Introduction to LT [PPT] [PDF]
2. Processing words [PPT], [PDF]

December 10, 2008, 15:15 - 19:00, MPŠ

1. Tomaž Erjavec: Computer corpora and Morphosyntactic tagging [PPT] [PDF]
2. Nina Ledinek: Syntactic analysis and the Slovene Dependency Treebank [PPT] [PDF]
   a sample from SDT is in here
3. Darja Fišer: Semantic lexica and Slovene WordNet [PDF]

March 4, 2009, 15:15 - 19:00, MPŠ

1. Language Resources and Machine Learning [PPT, PDF]

April 1, 2009, 11:15 - 15:00, E8 Orange room (Note changed date and location!)

1. Jerneja Žganec Gros, Speech Technologies:
   Introduction, Applications, Text-to-Speech

April 8, 2009, 15:15 - 19:00, MPŠ

1. Presentation and discussion of seminar work by students

Assessment

Exam dates:

• 13.5.2009, 17.00-19.00
• 10.6.2009, 16.00-18.00
• 24.6.2008, 15.00-17.00

Suggestions for seminar topics

1. Train and test the Brill tagger on the JOS corpus
2. Make and analysis of the JOS treebank (an example is here) and try to train and test MALT parser on it.
3. Use the Slovene WordNet for various tasks.

Literature list

1. The main textbook for the field is:

   Daniel Jurafsky, James H. Martin. Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics and Speech Recognition. Prentice-Hall, 2000.
   Contents: I. Words, II. Syntax, III. Semantics, IV. Pragmatics, V. Multilingual Processing.

2. All slides accompanying the lectures are available on the Web (links next to the lectures above)
3. Supplementary reading for the course topics are the following papers:
   • A Machine Learning Approach to Automatic Functor Assignment in the Prague Dependency Treebank. Zdenek Žabokrtsky, Petr Sgall, Saso Džeroski. In Proceedings of the Third International Conference on Language Resources and Evaluation, LREC'02.
   • Machine Learning of Morphosyntactic Structure: Lemmatising Unknown Slovene Words. Tomaž Erjavec and Sašo Džeroski. Applied Artificial Intelligence, 18(1), pp. 17-40, 2004.
   • Slovenian Text-to-Speech Synthesis for Speech User Interfaces. Jerneja Žganec Gros, Aleš Mihelič, Nikola Pavešič, Mario Žganec, Stanislav Gruden. Proceedings of the Third World Enformatika Conference, WEC 2005.
   • The VoiceTRAN Speech-to-Speech Communicator. Jerneja Žganec Gros, France Mihelič, Tomaž Erjavec, and Špela Vintar. Proceedings of the 8th International Conference on Text, Speech and Dialogue, TSD 2005. (Lecture notes in computer science, Lecture notes in artificial intelligence, 3658. Berlin: Springer)
   • Digitisation of Literary Heritage Using Open Standards. Tomaž Erjavec, Matija Ogrin. In Proceedings of eChallenges 2005, 19 - 21 October 2005, Ljubljana.
   • MULTEXT-East Version 3: Multilingual Morphosyntactic Specifications, Lexicons and Corpora. Tomaž Erjavec. In Proceedings of the Fourth International Conference on Language Resources and Evaluation, LREC'04.
4. The following books are also available:
   • The Oxford Handbook of Computational Linguistics. Ruslan Mitkov (ed.) Oxford University Press, 2003. [Sample]
   • Foundations of Statistical Natural Language Processing. Christopher D. Manning, Hinrich Schutze. MIT Press. 1999
   • ..and many other books and papers of the JSI library (room S10)

• MULTEXT-East (Slovene) corpus and lexicon
• IJS-ELAN corpus
• SVEZ-IJS corpus
• Slovene Dependency Treebank
• Slovene WordNet

Last updated 2009-04-08, et