Deep Learning Lab
Deep Learning and Grand Challenges of ML Lab
You should have attended a bachelor lecture in Machine Learning (e.g. Soft Computing), knowledge gaps can be filled in online courses on Machine Learning. In addition, it is recommended to have heard "Temporal and Spatial Data Mining" lecture. Basic Python knowledge is essential.
Contents and goals
Artificial Intelligence and especially Deep Learning are key technologies for the realization of numerous application domains, e.g. autonomous driving, speech recognition or automated monitoring of large-scale industrial facilities. In this lab, you will learn how to use deep learning methods and neural networks for time series analysis. During the lab you will learn the essential basics as well as advanced techniques of the current state of the art of Deep Learning, e.g. Generative Adversarial Networks, Recurrent Neural Networks, Variational Autoencoders, Transformers and many more.
In addition to the methodology, the focus is on the practical implementation of common methods of Deep Learning, i.e. you will learn the use of different frameworks, e.g. Scikit-learn, PyTorch, as well as the ability to create, perform and evaluate experiments in a scientific approach.
You will be able to try out the Deep Learning methods you have learned in the context of a project using heterogeneous, large-volume data sets or time series of a high-performance research test bench for electric motors.
The lab is divided into three phases:
1. lecture and exercise phase: teaching of the essential basics
2. seminar phase: independent investigation of state-of-the-art Deep Learning methods
3rd project phase: practical implementation of a given project (if necessary in small groups)