Deep Learning Architektures
Fakultät für Informatik und Mathematik ©
Name Deep Learning Architektures
Verantwortlich Prof. Dr. Rainer Schmidt
SWS 4
ECTS 5
Sprache(n) Englisch
Lehrform Seminar
Angebot nach Ankündigung
Aufwand

Attendance: about 60 Hrs., Independent Study: about 90 Hrs.

Voraussetzungen

In order to pass successfully this class, you should have basics of linear algebra, statistics, software development as well as good English language skills. The course is open to International students as well as students from other departments who qualify as stated above (Courses in English).

Ziele

Learning objectives are the understanding of deep neural networks and how they apply to different application fields. The students will be able to train, optimize and enhance performance in their deep models, getting a flavor for different frameworks and practical approaches to deep learning.

Inhalt

After an introduction in the learning theory and in the machine learning, we are going to explore different deep learning architectures, reading and understanding recent notable papers in this area.

Medien und Methoden

Books, Internet (General Information), Slides, Journals, .pdfs

Literatur

Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. MIT Press (2016).

Schmidhuber, J.: Deep learning in neural networks: An overview. Neural Netw. 61, 85–117 (2015).

Nielsen, M.A.: Neural Networks and Deep Learning. (2015).

Abu-Mostafa, Y.S., Magdon-Ismail, M., Lin, H.-T.: Learning from data : a short course. [United States] : AMLBook.com (2012).

Online classes:

  • http://cs231n.stanford.edu/

  • http://info.usherbrooke.ca/hlarochelle/neural_networks/content.html

  • http://cs229.stanford.edu/

And selected Papers.

Zuordnungen Curricula
SPO Fachgruppe Code ab Semester Prüfungsleistungen
IG Version 2010 CGBV: Fachliche u. persönliche Profilbildung 1 benotete Seminararbeit (60%)
benotetes Referat (40%)
IG Version 2010 EC: Fachliche u. persönliche Profilbildung 1 benotete Seminararbeit (60%)
benotetes Referat (40%)
IG Version 2010 SWE: Schwerpunkt 1 benotete Seminararbeit (60%)
benotetes Referat (40%)