Content:

The dramatic growth in practical applications for machine learning over the last ten years has been accompanied by many important developments in the underlying algorithms and techniques. In this course, basic machine learning techniques (so-called ''neural'' techniques, or simply neural networks) and recently introduced advanced techniques such as Deep Belief Networks and Gaussian Processes will be discussed. The student will gain a deep understanding of these techniques and learn to apply them for practical applications. We will adopt a modern view on neural networks based on Bayesian methods.

The following topics will be discussed:

• Linear models for regression / classification
• Deep neural networks
• Backpropagation and Second order methods
• Convolutional Neural Networks
• Deep Belief Networks
• Deep Boltzmann Machines
• and more...