Teaching

Machine Learning: Kernel-based Methods
(Introduction to the Mathematics of Supervised Learning)

Contents

This lecture provides a thorough introduction to kernel-based learning methods.

Prerequisites

Basic knowledge in statistics, analysis, and linear algebra is helpful.

Time

summer semesters (since 2004), Friday, 10.15 s.t.

Handouts (new)

Unit 1: The Learning Problem
Unit 2: Linear Regression and Discrimination
Unit 3: The Perceptron Revisited
Unit 4: Kernels and Reproducing Kernel Hilbert Spaces
Unit 5: Basic Kernel-based Learning Methods
Unit 6: Introduction to Constraint Optimization lecture notes
Unit 7: Support Vector Machines
Unit 8: Solving the SVM Optimization Problem
Unit 9: Rademacher Complexity

Script

Machine Learning: Kernel-based Methods

Laboratory Course

information about the laboratory course

Suggested reading

O. Bousquet, S.Boucheron and G. Lugosi. Introduction to Statistical Learning Theory. In Advanced Lectures in Machine Learning, LNAI 3176, Springer-Verlag, 2004
J. Shawe-Taylor and N. Cristianini. Kernel Methods for Pattern Analysis. Cambridge University Press, 2004.
B. Schölkopf and A. J. Smola. Learning with Kernels. MIT Press, 2002.
N. Cristianini and J. Shawe-Taylor. An Introduction to Support Vector Machines (and other kernel based learning methods). Cambridge University Press, 2000.
K. Lang. Optimization. Springer Texts in Statistics, Springer-Verlag, 2004
T. Poggio and S. Smale. The Mathematics of Learning: Dealing with Data. Notices of the American Mathematical Society (AMS), 50(5), 2003
T. Hastie, R. Tibshirani, and J. Friedman. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer-Verlag, 2001
C. M. Bishop. Pattern Recognition and Machine Learning. Springer-Verlag, 2006


Adaptive Systems Seminar

Description

Seminar about recent findings in the domain of optimization of adaptive systems for diploma and doctoral students.

Time

winter/summer semesters (since 2003/2004), Friday, 14.15 s.t.

Reinforcement Learning

Description

The lecture provides an introduction to reinforcement learning (RL). It covers the basics of RL, including Markov decision processes, dynamic programming, and temporal-difference learning. Additionally, special topics such as policy gradient methods as well as the connection to neural networks and evolutionary computation are presented.

Prerequisites

Very basic knowledge in statistics and analysis is required, basic knowledge about neural networks is helpful.

Time

winter semesters (since 2003/2004), Friday, 10.15 s.t.

Slides

Unit 1: Introduction
Unit 2: The Reinforcement Learning Problem gridworld example
Unit 3: Dynamic Programming
Unit 4: Monte Carlo Methods
Unit 5: Temporal Difference Learning
Unit 6: Eligibility Traces forward/backward proof
Unit 7: Generalization and Function Approximation CMAC details
Unit 8: Planning and Learning
Unit 9: Least-squares Temporal Difference Learning
Unit 10: Policy Gradient Methods
Unit 11: Direct Polcy Search

Laboratory Course

information about the laboratory course

Additional course material

R. Sutton and A. Barto. Reinforcement Learning: An Introduction, MIT Press, 1998.
V. Heidrich-Meisner, M. Lauer, C. Igel, and M. Riedmiller. Reinforcement Learning in a Nutshell. In M. Verleysen, editor, 15th European Symposium on Artificial Neural Networks (ESANN 2007), Belgium: d-side publications, pp. 277-288, 2007
R. J. Williams. Simple statistical gradient-following algorithms for connectionist reinforcement learning, Machine Learning, 8, 229-256, 1992
J. Boyan. Technical Update: Least-Squares Temporal Difference Learning. Machine Learning, 49(2-3), pp. 233-246, 2002
D. E. Moriarty, A. C. Schultz, and J. J. Grefenstette. Evolutionary Algorithms for Reinforcement Learning, Journal of Artificial Intelligence Research, 11, 199-229, 1999
R. S. Sutton, D. McAllester, S. Singh, and Y. Mansour. Policy Gradient Methods for Reinforcement Learning with Function Approximation. In S. A. Solla and T. K. Leen and K.-R. Müller, eds.: Advances in Neural Information Processing Systems 12, pp. 1057-1063, MIT Press, 2000


Machine Learning: Unsupervised Learning

with Prof. Dr. Laurenz Wiskott, see his pages on the course for details

Description

This course is mainly given by Laurenz Wiskott. It covers a variety of unsupervised methods from machine learning such as principal component analysis, independent component analysis, vector quantization, clustering, self-organizing maps, growing neural gas, Bayesian theory and graphical models, deep-belief networks, and Markov random fields.

Time

winter semesters, starting 2009/2010

lecture: Tuesday, 14.15 s.t.
analytical tutorial: Tuesday, 12.15 s.t.

Literature and Lecture Notes

For many topics a script will be available, other literature will be mentioned in the lecture. German lecture notes on Markov random fields can be found here.

Prerequisites

Good command of linear algebra and calculus.

Neurocomputing Methods

with Prof. Dr. Gregor Schöner

Description

A seminar about fundamental methods in neurocomputing.

Time

just in summer semester 2004