== Block-Seminar ``Representations in Machine Learning" ==

Vorbesprechung am 20.10.2009 um 14:00 Uhr im Raum FR 6046. 

=== Language ===

 English

=== Topics ===

 * Grundlagen zu verschiedenen Konzepten des Lernens von Repraesentationenen und 'Features'
 * Artikel der aktuellen Forschung zB zu deep, learning autoencoders, convolutional nets, etc
 * MDL, Kernel PCA, etc

Das Blockseminar wird Grundlagen-Artikel und Artikel der aktuellen Forschung besprechen, die sich mit Alternativen zu standard graphischen Modellen, und insbesondere mit dem Lernen von Repraesentationen.


=== Voraussetzungen ===

 * basics in Machine Learning, probabilistic & graphical models

=== Papers ===

 * Weston et al: Deep Learning via Semi-Supervised Embedding (ICML 2008) [[http://www.kyb.tuebingen.mpg.de/bs/people/weston/papers/deep_embed.pdf]]
 * Hinton et al: A Fast Learning Algorithm for Deep Belief Nets (Neural Computation) [[http://www.cs.toronto.edu/~hinton/absps/ncfast.pdf]]
 * Tishby et al: The Information Bottleneck Method [[http://www.cis.upenn.edu/~pereira/papers/allerton.pdf]]
 * Honton et al: Reducing the Dimensionality of Data with Neural Networks (Science) [[http://www.cs.toronto.edu/~hinton/science.pdf]]
 * Schoelkopf et al: Kernel Principal Component Analysis [[http://www.eecs.berkeley.edu/~wainwrig/stat241b/scholkopf_kernel.pdf]]
 * Gruenwald: A Tutorial Introduction to the Minimum Description Length Principle [[http://homepages.cwi.nl/~pdg/ftp/mdlintro.pdf]]
 * DeBie et al: Eigenproblems in Pattern Recognition [[http://www.meduniwien.ac.at/user/roman.rosipal/Papers/eig_book04.pdf]]
 * Welling: Herding Dynamical Weights to Learn (ICML 2009) [[http://www.cs.mcgill.ca/~icml2009/papers/447.pdf]]
 * LeCun et al: Gradient-Based Learning Applied to Document Recognition[[http://yann.lecun.com/exdb/publis/pdf/lecun-98.pdf]]

=== Further reading ===
 * Domingos: Structured Machine Learning: Ten Problems for the Next Ten Years [[http://www.cs.washington.edu/homes/pedrod/papers/ilp07.pdf]]
 * Bengio et al: Scaling Learning Algorithms towards AI. [[http://yann.lecun.com/exdb/publis/pdf/bengio-lecun-07.pdf]]