Neural Networks (NNW)

Dive into how modern AI works: You will develop a deep understanding of neural network methodologies – from feedforward and convolutional networks to transformers, large language models and generative models – including their mathematical foundations and typical application areas. You also learn how to implement, train and apply these methods in practice using a deep learning framework such as PyTorch.

Contents

• Motivation
• Tasks (classification, regression, supervised / unsupervised learning)
• Bayesian classification and probability theory
• Fundamental concepts of neural networks (elements, nomenclature, matrix representation, error/cost functions)
• Deep feedforward networks (characteristics, error backpropagation, problems training deep networks, advanced methods for minimizing cost functions)
• Convolutional neural networks
• Sequential data: Recurrent networks and Transformers, Large Language Models (LLMs), Vision Transformers
• Kernel-based methods (radial basis functions, support vector machines)
• Undirected models, generative models and autoencoder (fundamentals, restricted Boltzmann machines, deep belief networks)
• Evaluation and feature selection

Competencies

Students are able to

• explain the mathematical background and properties of neural networks,
• apply neural networks for classification and regression tasks,
• relate neural networks to other machine learning methods.

Literature

• C.M. Bishop, H. Bishop: Deep Learning. Springer, 2024
• K. Murphy: Probabilistic Machine Learning: An Introduction. MIT Press, 2022
• K. Murphy: Probabilistic Machine Learning: Advanced Topics. MIT Press, 2023
• I. Goodfellow, Y. Bengio, A. Courville: Deep Learning. MIT Press, 2016
• S. Theodoridis, K. Konstantinos: Pattern Recognition. Academic Press, 2009
• C.M. Bishop: Pattern Recognition and Machine Learning. Springer, 2006
• C.M. Bishop: Neural Networks for Pattern Recognition. Oxford University Press, 1996
• D. Yu, L. Deng: Automatic Speech Recognition – A Deep Learning Approach. Springer, 2015
• R.O. Duda et al.: Pattern classification. Wiley-Interscience, 2001

Lecturer

• Prof. Dr. Ivo Wolf

Recommended Previous Knowledge 

• Fundamentals of linear algebra and analysis
• Although not a prerequisite, knowledge of probability theory and statistics will be beneficial.

Module Details

Semester	1/2
Lecture Language	German
Frequency	Not regularly
Credit Points (ECTS)	5
Course Coordinator	Prof. Dr. Ivo Wolf
Duration	1 Semester
Study Focus	Medical Data Science (MDS)
Course Achievement	None
Prerequisite for exam	Compulsory Assignment (PU)
Exam	Written Exam 90 Minutes (K90)

Weekly Hours (SWS) 

Lecture	2 SWS
Exercises	2 SWS
Sum	4 SWS

Work Load 

Lecture	30 h
Lab	30 h
Self Study	30 h
Assignments	30 h
Exam Preparation	30 h
Sum	150 h