Universität Wien

052316 VU Deep Learning for Natural Language Processing (2024W)

6.00 ECTS (4.00 SWS), SPL 5 - Informatik und Wirtschaftsinformatik
Prüfungsimmanente Lehrveranstaltung
Moodle
Do 10.10. 09:45-11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02

An/Abmeldung

Hinweis: Ihr Anmeldezeitpunkt innerhalb der Frist hat keine Auswirkungen auf die Platzvergabe (kein "first come, first served").
Für diese LV an-/abmelden

Details

max. 25 Teilnehmer*innen
Sprache: Englisch

Lehrende

Termine (iCal) - nächster Termin ist mit N markiert

  • Donnerstag 03.10. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 03.10. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 10.10. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 17.10. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 17.10. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 24.10. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 24.10. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 31.10. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 31.10. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 07.11. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 07.11. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 14.11. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 14.11. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 21.11. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 21.11. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 28.11. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 28.11. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 05.12. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 05.12. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 12.12. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 12.12. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 09.01. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 09.01. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 16.01. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 16.01. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 23.01. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 23.01. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 30.01. 09:45 - 11:15 PC-Seminarraum 3, Kolingasse 14-16, OG02
  • Donnerstag 30.01. 11:30 - 13:00 PC-Seminarraum 3, Kolingasse 14-16, OG02

Information

Ziele, Inhalte und Methode der Lehrveranstaltung

This course will cover topics related to the application of DL (Deep Learning) techniques for solving NLP (Natural Language Processing) tasks. It will start with a comprehensive introduction of the deep learning paradigms, neural probabilistic language models and word embeddings. Next, it will cover concepts like the encoder-decoder architecture, recurrent neural networks, Attention and contextual embeddings. The course will continue with advanced developments of the recent years such as the Transformer architecture, BERT and its derivatives, T5, GPT-* as well as ChatGPT. An integral part will be the implementation of certain concepts using Python frameworks such as Numpy and Pytorch. There will be 5 advanced and involving programming projects. Consequently, a solid background of Python is highly relevant and necessary.
(DH students who want to take this lecture need to have passed the lecture "Practical Machine Learning for Natural Language Processing" with very good success, or have equivalent previous knowledge in programming and machine learning, for successfully participating in this lecture.)

Art der Leistungskontrolle und erlaubte Hilfsmittel

- Regular assignments throughout the semester in Moodle 10%
- Programming exercises 20%
- Midterm exam 35%
- Final exam 35%

Mindestanforderungen und Beurteilungsmaßstab

The participant must attend at least 75 % of the sessions. The grade is calculated from the total points as follows:

>= 90% very good (1)
>= 80% good (2)
>= 65% satisfactory (3)
>= 50% sufficient (4)
< 50% not sufficient (5)

Prüfungsstoff

Handing in regular assignments throughout the semester in Moodle, solving rogramming exercises. Questions in the exams can be on all topics covered in the lecture and exercise sessions.

Literatur

Jason Brownlee: "Basics of Linear Algebra for Machine Learning - Discover the Mathematical Language of Data in Python"
https://github.com/balban/Books/tree/master/Linear%20Algebra

Yoav Goldberg: "Neural Network Methods for Natural Language Processing", Morgan & Claypool, 2017
https://github.com/Michael2Tang/ML_Doc

Steven Bird, Ewan Klein, Edward Loper: "Natural Language Processing with Python - Analyzing Text with the Natural Language Toolkit"
https://www.nltk.org/book

Ian Goodfellow and Yoshua Bengio and Aaron Courville: "Deep Learning", MIT Press, 2016.
https://www.deeplearningbook.org

Zuordnung im Vorlesungsverzeichnis

Letzte Änderung: Di 24.09.2024 15:25