Universität Wien

340234 UE Maschinelle Translation (2022S)

3.00 ECTS (2.00 SWS), SPL 34 - Translationswissenschaft
Prüfungsimmanente Lehrveranstaltung
GEMISCHT
Moodle

An/Abmeldung

Hinweis: Ihr Anmeldezeitpunkt innerhalb der Frist hat keine Auswirkungen auf die Platzvergabe (kein "first come, first served").

Details

max. 25 Teilnehmer*innen
Sprache: Deutsch, Englisch

Lehrende

Termine (iCal) - nächster Termin ist mit N markiert

  • Dienstag 08.03. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 15.03. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 22.03. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 29.03. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 05.04. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 03.05. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 10.05. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 17.05. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 24.05. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 31.05. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 14.06. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 21.06. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG
  • Dienstag 28.06. 15:00 - 16:30 Medienlabor II ZfT Gymnasiumstraße 50 4.OG

Information

Ziele, Inhalte und Methode der Lehrveranstaltung

UE Maschinelle Translation (English Version):

Goals:
Students will apply elements of the three main machine translation (MT) paradigms to a range of tasks, ranging from automatic translation between closely-related languages, language variants or dialects to building a new machine translation engine from scratch.

Using state-of-the-art technologies, students will learn to design a machine translation engine, gather training data, clean the data, train at least one engine, evaluate it, and improve it.

In addition, students will also use controlled language and technical writing principles and techniques in order to author or edit content for the purpose of improving the quality of raw MT output. Post-editing MT (PEMT) standards and best practices will be applied to practical PEMT tasks, which will give students the opportunity to fine-tune their skills identifying and annotating subtle MT errors using popular industry annotation frameworks.

Content:
• Rule-based (RBMT), statistical (SMT) and neural machine translation (NMT): current applications
• NMT architectures
• MT evaluation metrics
• MT quality estimation (QE) practices
• MT in professional workflows
• MT in more complex systems (e.g. speech-to-speech translation)
• Controlled language; pre-editing
• Post-editing machine translation (PEMT) standards and best practices
• Ethics of using MT and impact of MT on freelance linguists

Didactic approach:
Students will need to complete assignments involving a wide range of technologies involved in the process of building MT engines, estimating the quality of their output, and comparing the results of these automatic estimations to other MT evaluation techniques.

Students will also gain experience of post-editing MT output.
Depending on the lecturer available, the seminar will be held in German or in English. If it is held in English, it will have, whenever possible, simultaneous (but automatic, machine-generated) translation into German which, although not perfect, should still give students broad access to the live discussions and so that they have the live experience of machine translation applied in the
course itself.

Art der Leistungskontrolle und erlaubte Hilfsmittel

Continuous evaluation:
- attendance, seminar presentation and weekly reflections count for 20% of the mark.
- student portfolio counts for 40% of the mark.
- final reflective essay counts for 40% of the mark.

Mindestanforderungen und Beurteilungsmaßstab

MT UE pass mark
In order to pass this module, a student needs to reach the threshold of 4.
MT UE marking map
excellent - sehr gut (1)
good - gut (2)
average - befriedigend (3)
sufficient - genügend (4)
insufficient - nicht genügend (5)

Prüfungsstoff

Continuous evaluation

Literatur

- Carstensen, K-U. 2017. Sprachtechnologie - Ein Überblick. http://kai-uwe-carstensen.de/
Publikationen/Sprachtechnologie.pdf
- Chan, Sin-Wai. Ed. 2015. Routledge encyclopedia of translation technology Abingdon, Oxon : Routledge.
- Depraetere, I. Ed. 2011. Perspectives on translation quality. Berlin: de Gruyter Mouton
- Hausser, Roland. 2000. Grundlagen der Computerlinguistik - Mensch-Maschine-Kommunikation in natürlicher Sprache (mit 772 – Übungen). Springer.
- Kockaert, H. J. and Steurs, F. Eds. 2015. Handbook of terminology. Amsterdam; Philadelphia: John Benjamins Publishing Company.
- Koehn, P. 2020. Neural Machine Translation. Cambridge University Press
- Munday, J. 2012. Evaluation in translation: critical points of translator decision-making: Routledge.
- O'Hagan, M. Ed. 2019. The Routledge Handbook of Translation and Technology. Abingdon: Routledge
- Waibel, A. 2015. Sprachbarrieren durchbrechen: Traum oder Wirklichkeit? Nova Acta Leopoldina NF 122, Nr. 410, 101–123. https://isl.anthropomatik.kit.edu/downloads/
NAL_Bd122_Nr410_101-124_Waibel_low_res.pdf
- Wright, S. E. and Budin, G. 1997/2001. The Handbook of Terminology Management. Two volumes. Amsterdam/Philadelphia: John Benjamins Publishing Company.
- BS EN ISO 17100:2015: Translation Services. Requirements for translation services
- ISO/DIS 18587 Translation services - Post-editing of machine translation output – Requirements
- ASTM F2575 Standard Guide for Quality Assurance in Translation

Zuordnung im Vorlesungsverzeichnis

Letzte Änderung: Mo 07.03.2022 13:09