Pattern Recognition and Machine Learning I
Prerequisite for the lecture is the knowledge from the mathematics lectures (Stochastics or Discrete Structures, Analysis, Linear Algeba) of a completed Bachelor degree in Computer Science, Electrical Engineering, Mechatronics, Mathematics or similar.
This course (lecture/exercise) provides a basic introduction to an area that deals with the analysis of data, the recognition of regularities in these data and the creation of models from data. These models can be used, for example, to classify data (e.g. categorization during quality control) or to solve a regression problem (e.g. prediction of the performance of a wind turbine).
In this lecture basic methods and procedures are discussed on the basis of a worldwide known standard textbook for pattern recognition and machine learning. The goal is to know them in detail in such a way that they can not only be applied in a particular manner, but also in further developments Among other things, the following topics will be discussed: Fundamentals (e.g. stochastics, model selection, curse of dimensionality, decision and information theory), distributions (e.g. multinomial, dirichlet, gaussian and student distribution, nonparametric estimation of distributions), linear models for regression, linear models for classification, kernel functions and advanced neural networks (e.g. Convolutional Neural Networks, Radial Basis Function Networks), Gaussian processes.
In the exercise the application of different techniques is examined with the help of Jupyter notebooks and suitable Python libraries. Sample data from different application fields are considered. The goal is to learn the safe, systematic and careful use of the mentioned techniques.
The course in creates the prerequisites for further courses such as Laboratory Deep Learning or Lecture/Exercise Autonomous Learning.