MC1 Making Sense of Sensory Data with Generative Machine Learning


Introduction to probabilistic models of sensory data. From probabilistic models to unsupervised and supervised learning algorithms (mixture models, component analysis, and sparse coding). Advanced models, deep models and current limits. Probabilistic models and sensory systems of humans and animals.


The course aims to provide the audience with a background of how probabilistic models are used in Machine Learning and AI. The principle approach will be presented using some prominent examples to provide intuition. Finally, the general theoretical framework and school of thought is outlined, and current research directions, including deep unsupervised learning, will be outlined.


Pattern Recognition and Machine Learning, C. M. Bishop, Springer 2006.
Theoretical Neuroscience - Computational and Mathematical Modeling of Neural Systems, Laurence F. Abbott and Peter Dayan, 2001.
Information Theory, Inference, and Learning Algorithms, D. MacKay, 2003. (free online)
Machine Learning: A Probabilistic Perspective, K. P. Murphy, MIT Press, 2012.

Course location

Forum 1

Course requirements


Instructor information.

Jörg Lücke


University of Oldenburg


Jörg Lücke received his PhD from the Ruhr-University Bochum, Germany, in 2005, and then joined the Gatsby Computational Neuroscience Unit, UCL, UK, as a senior research fellow. With grants from different funding agencies (2008-2013), he then built up his own research group at the Frankfurt Institute for Advanced Studies, Goethe-University Frankfurt, and later moved the group to the Technical University Berlin. In 2013 he is Professor at the University of Oldenburg, Germany, where he heads the newly established Machine Learning Group. Jörg Lücke conducts research projects on basic Machine Learning, Computational Neuroscience, and on Machine Learning for Computer Hearing and Computer Vision.