392028 Grundlagen Neuronaler Netze (V) (SoSe 2024)

Contents, comment

Neuronale Netze sind ein aktives Feld der Informatik und haben speziell in der Form von Deep Neural Networks in den letzten Jahren weitreichende neue Möglichkeiten in vielen Bereichen des Machine Learnings eröffnet, z.B. in der Spracherkennung und Computer Vision. Die Vorlesung vermittelt ein Verständnis der grundlegenden Konzepte neuronaler Informationsverarbeitung.

Ausgehend von Modellvorstellungen der Informationsverarbeitung in biologischen Neuronennetzen werden theoretische Grundlagen und Lernverfahren künstlicher neuronaler Netze dargestellt.
In praktischen Übungen wird das Gelernte vertieft und mit Hilfe von Simulationen in Python programmiertechnisch angewandt.

Ziel der Vorlesung: Die Studierenden ...
... verstehen Grundlagen von Lernen und Gedächtnisprozessen,
... kennen verschiedene Netzwerktypen und dazugehörende Lernverfahren,
... sind in der Lage, die Leistungsfähigkeit der besprochenen Verfahren einzuschätzen
... und können diese auf Probleme in Anwendungsdomänen erfolgreich einsetzen.

Inhaltlich führt die Veranstaltung zuerst überwachte Lernverfahren (supervised learning) ein, wie Regression und Klassifikation. Dazu werden verschiedene Neuronen-Modelle und Neuronale Netz-Modelle diskutiert (Perceptron, MLP, Deep Neural Networks, Convolutional Neural Networks). Im zweiten Teil liegt der Fokus auf unüberwachten Lernverfahren, zum Beispiel Self-Organizing Maps und Hebb-Lernen.

Neural networks have become a thriving field of computer science. In particular Deep Neural Network approaches have been quite successful over the last years in areas like computer vision or speech recognition. The lecture discusses basic notions and concepts underlying neural networks and learning of such systems, starting from supervised learning for regression and classification. It further introduces standard neural networks models, the perceptron and the multi-layer perceptron. Turning to unsupervised learning, several algorithms for vector quantization are introduced, Hebb-learning, and Self-Organizing Maps.

Learning goals: Students …
… understand basic principles of learning and memory,
… know different network types and corresponding learning methods,
… are able to assess the efficiency and effectivity of these methods,
… and can compare as well as apply different neural network models and learning approaches.

Requirements for participation, required level

Algorithmen und Datenstrukturen, Vertiefung Mathematik - die Theorie Neuronaler Netze ist ein Thema der angewandten Mathematik (lineare Algebra).

Requirements for the Lecture
Algorithmen und Datenstrukturen, Vertiefung Mathematik. Neural Networks is an applied math topic - make sure that you recall the necessary basic math.

The lecture achieves 2 CP, the corresponding exercises 2 CP, and an oral exam, which is necessary to complete the module, adds 1 CP.

Bibliography

Literatur
- Deep Learning by I. Goodfellow, Y. Bengio and A.Courville, MIT Press, 2016
- C. Bishop, Pattern Recognition and Machine Learning, Springer 2006, in particular Chap. 1, Chap. 3.1-3.3 (linear models für regression), Chap 4.1. /4.2. (für classification) , Chap 5. (feedforward neural networks).
Darüber hinaus englischsprachiges Skript.

Part of the lecture is based on the book on Deep Learning by I. Goodfellow, Y. Bengio and A.Courville, MIT Press, 2016. Part of the lecture will lean on C. Bishop, Pattern Recognition and Machine Learning, Springer 2006, in particular Chap. 1, Chap. 3.1-3.3 (linear models für regression), Chap 4.1. /4.2. (für classification) , Chap 5. (feedforward neural networks).
Furthermore, lecture notes (in English) are available.

Teaching staff

Dates ( Calendar view )

Frequency Weekday Time Format / Place Period  
weekly Di 10-12 H12 08.04.-19.07.2024

Subject assignments

Module Course Requirements  
39-Inf-NN_ver1 Grundlagen Neuronaler Netze Neuronale Netze und Lernen I Ungraded examination
Graded examination
Student information
Neuronale Netze und Lernen I Student information

The binding module descriptions contain further information, including specifications on the "types of assignments" students need to complete. In cases where a module description mentions more than one kind of assignment, the respective member of the teaching staff will decide which task(s) they assign the students.

Degree programme/academic programme Validity Variant Subdivision Status Semester LP  
Studieren ab 50    
Veranstaltungen für Schülerinnen und Schüler   Die Anmeldung zum Schnupperstudium erfolgt über die ZSB per E-Mail an: dop@uni-bielefeld.de  

The lecture achieves 2 CP, the corresponding exercises 2 CP, and an exam, which is necessary to complete the module, adds 1 CP. The detailed schedule for the exercises is published through moodle.

E-Learning Space

A corresponding course offer for this course already exists in the e-learning system. Teaching staff can store materials relating to teaching courses there:

Registered number: 128 (2)
This is the number of students having stored the course in their timetable. In brackets, you see the number of users registered via guest accounts.
Address:
SS2024_392028@ekvv.uni-bielefeld.de
This address can be used by teaching staff, their secretary's offices as well as the individuals in charge of course data maintenance to send emails to the course participants. IMPORTANT: All sent emails must be activated. Wait for the activation email and follow the instructions given there.
If the reference number is used for several courses in the course of the semester, use the following alternative address to reach the participants of exactly this: VST_453213498@ekvv.uni-bielefeld.de
Coverage:
126 Students to be reached directly via email
Notes:
Additional notes on the electronic mailing lists
Last update basic details/teaching staff:
Monday, January 22, 2024 
Last update times:
Thursday, February 1, 2024 
Last update rooms:
Thursday, February 1, 2024 
Type(s) / SWS (hours per week per semester)
V / 2
Department
Faculty of Technology
Questions or corrections?
Questions or correction requests for this course?
Planning support
Clashing dates for this course
Links to this course
If you want to set links to this course page, please use one of the following links. Do not use the link shown in your browser!
The following link includes the course ID and is always unique:
https://ekvv.uni-bielefeld.de/kvv_publ/publ/vd?id=453213498
Send page to mobile
Click to open QR code
Scan QR code: Enlarge QR code
ID
453213498