Neural Networks and Applications

11B1595

Bachelor

German

5.0

irregular

1 semester

The technology of neural networks as a subfield of AI has developed into an indispensable and efficient tool for many engineering problems. The methods of neural networks and their possible areas of application (e.g. regression tasks, forecasts and image classifications) are demonstrated and practiced.  

1. motivation and biological basics

2. data analytical basics

3. theory of neural networks (perceptron, multilayer perceptron, learning methods, quality criteria, deep learning, generalization)

4. application-specific neural networks (e.g. regression, image classification, time series prediction)

5. familiarization with software libraries for the implementation and use of neural networks

The total workload for the module is 150 hours (see also "ECTS credit points and grading").

• Homework / Assignment or
• Written examination or
• oral exam

Term paper: 10–15 pages

Written examination: see the applicable study regulations

Oral examination: see the general section of the examination regulations

Basic knowledge of differential and integral calculus as well as knowledge of functions with multiple variables is required.  Knowledge of linear regression is also helpful.

Students who have successfully studied this subject are familiar with common network structures and learning methods. They have theoretical background knowledge and can assess the potential of neural networks. They are familiar with typical applications of neural networks and have learned to create and use neural networks for practice-oriented examples.

Students who have successfully completed this module also have basic knowledge in the areas of data analysis, statistics and other methods of knowledge-based systems. The module repeats and deepens some of the basic knowledge already acquired in the modules Mathematics 1 and 2 and Programming 1 and 2.

Students learn the basic procedures of data analysis. They learn about the potential fields of neural networks and can use them to solve data-based problems. They will learn about and use a number of tools and libraries for creating neural networks.

• Gervens, Theodor

• Gervens, Theodor