Lecture Mustererkennung (pattern recognition)

(Here you can find formally binding study and examination regulations and module descriptions. )

© Peter Birkholz

Content

In the lecture Pattern Recognition, we deal with the methods of machine learning, such as those used in speech recognition, object recognition in images or large language models. We start with conventional regression methods, linear classifiers, simple statistical classifiers and Support Vector Machines. We then enter the world of artificial neural networks. We first look at the simplest neuron model - the perceptron - and the perceptron learning algorithm. We then look at multi-layer neural networks for deep learning, their variants and training methods. Finally, we will learn about methods that are particularly suitable for processing time series, especially recurrent neural networks and the transformer architecture. The parallel lecture Signal Analysis deals with methods of feature extraction that are required as input by the machine learning methods.

The lecture language can be English or German and will be determined in the first lecture.

The lecture is suitable for

Also suitable for students of other technical disciplines.

Recommended background knowledge: Systemtheorie (system theory) I and II, Signalverarbeitung (signal processing)

The lecture is offered regularly in summer semester.

Please enroll in the corresponding OPAL class for further information regarding organization and content.

Lecture Mustererkennung on learning platform OPAL