Universität Wien

122049 PS Proseminar Linguistics 2 (2020S)

Language and Artificial Intelligence

5.00 ECTS (2.00 SWS), SPL 12 - Anglistik
Prüfungsimmanente Lehrveranstaltung
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: Englisch

Lehrende

Termine (iCal) - nächster Termin ist mit N markiert

  • Donnerstag 05.03. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 19.03. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 26.03. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 02.04. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 23.04. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 30.04. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 07.05. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 14.05. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 28.05. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 04.06. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 18.06. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09
  • Donnerstag 25.06. 14:00 - 16:00 Raum 2 Anglistik UniCampus Hof 8 3E-EG-09

Information

Ziele, Inhalte und Methode der Lehrveranstaltung

State-of-the-art intelligent linguistic systems (like the chatbot here: https://ucard.univie.ac.at/studierende/) are capable of ‘understanding’ linguistic meaning to a certain extent. This proseminar is an introduction to the interdisciplinary field of Artificial Intelligence (AI) with a focus on how semantics is modeled in AI approaches to language. After a brief introduction to the history and central questions of AI and linguistics we discuss two prominent – but substantially different – semantic representation models: (i) semantic web and (ii) embedding models.

In the methodological part of the course, we dig deeper into vector representations of words (‘word embeddings’), i.e. numerical encodings of word meaning. These encodings can be used to perform actual computations. For example, they let AI systems perform analogical reasoning like “king : queen = man : X implies that X = woman”. Students learn how to perform computations like that and how to visualize semantic fields using open-access online tools.

The third part of the course is dedicated to group projects focusing on different topics related to embedding methods (e.g. ‘AI and gender bias’, ‘embeddings and semantic change’ or ‘embeddings and emotion’). In these projects, students apply the skills acquired in the second part to their research questions by analyzing embeddings derived from massive text data.
In general, the course is intended to give insights into how AI language systems deal with semantics. This should help to critically reflect on these nowadays widely used digital techniques. The course does not require any previous knowledge in statistics, computer science or programming, nor is it supposed to be an introduction to any of these fields and methods. However, a basic knowledge of high-school mathematics will prove useful (basic calculus, percentages, vectors) to understand the methods discussed in the course.

UPDATE: The first sessions of the course will be conducted via moodle. There will be online tasks and discussions.

Art der Leistungskontrolle und erlaubte Hilfsmittel

Short essay, online quiz, group project (presentation and proseminar paper), participation in class.

Mindestanforderungen und Beurteilungsmaßstab

Students understand basic concepts in semantic modeling in AI systems and the principles of word-embedding techniques as well as different fields of application. Furthermore, they can use online tools to perform simple computations with word embeddings.

Assessment:
Short essay (10%)
Online quiz (15%)
Presentation (30%)
Proseminar paper (30%)
Participation in class (15%)
Pass grade: 60%

Prüfungsstoff

Literatur

Will be provided in class and includes (chapters of):

Erk, K., 2012. Vector space models of word meaning and phrase meaning: A survey. Language and Linguistics Compass, 6(10), pp.635-653.
Frankish, K., & Ramsey, W. M. (Eds.). (2014). The Cambridge handbook of artificial intelligence. Cambridge University Press.
Hofstadter, D. R. (1995). Fluid concepts and creative analogies: Computer models of the fundamental mechanisms of thought. Basic books.
Rumelhart, David E., Bernard Widrow, and Michael A. Lehr. "The basic ideas in neural networks." Communications of the ACM 37.3 (1994): 87-93.
Young, T., Hazarika, D., Poria, S., & Cambria, E. (2018). Recent trends in deep learning based natural language processing. IEEE Computational intelligence magazine, 13(3), 55-75.

Zuordnung im Vorlesungsverzeichnis

Studium: BA 612;
Code/Modul: BA06.1;
Lehrinhalt: 12-2044

Letzte Änderung: Mo 07.09.2020 15:20