Lecture notes:

PDF, opens an external URL in a new window (20MB, 2014-03-02)

Slides:

Part1, opens an external URL in a new window (2MB)

Part2, opens an external URL in a new window (15MB)

Part3, opens an external URL in a new window (16MB)

Motivation:

Machine learning is concerned inferring models/relationships by learning from data. Machine learning methods are gaining importance in various fields, such as, process modeling, speech and image processing, and so forth. In recent years, bioinformatics has become one of the most prominent application areas of machine learning methods: The massive data amounts produced by recent and currently emerging high-throughput biotechnologies provide unprecedented potentials, but also pose yet unseen computational challenges in the analysis of biological data.

This course focuses on so-called unsupervised machine learning techniques, that is, methods aiming at inferring structure/models in data without an explicit target. The students should aquire skills to choose, use, and adapt methods for clustering, data projection, and data reduction for tasks in science and engineering. The students should particularly understand the underlying mathematical objectives and principles of unsupervised machine learning methods. Topics:

• Error models
• Maximum likelihood and the expectation maximization algorithm
• Maximum entropy methods
• Basic clustering methods, hierarchical clustering, and affinity propagation
• Mixture models
• Principal component analysis, independent component analysis, and other projection methods
• Factor analysis
• Matrix factorization
• Auto-associator networks and attractor networks
• Boltzmann and Helmholtz machines
• Hidden Markov models
• Belief networks
• Factor graphs

(Practical course Machine Learning: Unsupervised Techniques (1UE))