Teaching, Tutorials, Notes

Courses & Tutorials
Readings
Lecture notes

Verzeichnis meiner Lehrveranstaltungen im KVV des Fachbereichs Informatik.

Courses & Tutorials

SS 12 - Machine Learning

Summer term course at FU Berlin. See the course web page.

WS 11/12 - Robotics

Winter term course at FU Berlin. See the course web page.

SS 11 - Machine Learning

Summer term course at FU Berlin. See the course web page.

ICML 2011 Tutorial on Machine Learning & Robotics

Machine Learning tutorial at the Interdisciplinary College 2011

The 3 basic lectures target an interdiciplinary audience (students from Computer Sci, Cog Science, Neuroscience, Psychology), covering basics in ML, Bayesian Modelling, and RL:

1. Introduction
2. Linear Models (non-linear features, regularization, cross-validation, `linear/polynomial/kernel Ridge/Lasso regression/logistic classification')
3. Bayesian Modelling (Bayes, examples, regularization & prior, error & likelihood, MAP view on Ridge/Lasso regression, EM, Bayes Nets)
4. Reinforcement Learning (Markov Decision Process, values, temporal difference, model-free vs. model-based, planning by probabilistic inference)

BCCN lecture Computational models of goal-directed behavior

slides exercise.

WS 10/11 - Robotics

Winter term course at FU Berlin. See the course web page.

RLSS 09 - Inference & Planning

Lectures given at the Robot Learning Summer School (Lisbon, July 20-24 2009).
Slides: part 1, part 2

Part 1: Introduction to probabilistic inference \& learning
: -- probabilities, joint distributions, graphical models -- inference, message passing -- learning, Expectation Maximization
Part 2: Planning by Inference: -- general idea of inference by planning -- Markov Decision Processes revisited -- Stochastic Optimal Control revisited
Summary & further reading: -- brief summary -- further reading -- food for thought

SS 09 - Introduction to Graphical Models

Summer term course at TU Berlin. See the course web page.

ICML 08 tutorial - Stochastic Optimal Control

Tutorial, held together with Bert Kappen on Saturday July 5 2008 in Helsinki, Finland as part of the 25th International Conference on Machine Learning (ICML 2008). See the tutorial web page.

Interesting Readings

Anil Ananthaswamy: I, algorithm: A new dawn for artificial intelligence. A popular science article in NewScientist, 2011.

Pat Langley: The changing science of machine learning. Editorial in Machine Learning 82, 275-279, 2011.

Thomas G. Dietterich et al.: Structured machine learning: the next ten years. Machine Learning, 73, 3-23, 2008.

Yoshua Bengio & Yann LeCun: Scaling learning algorithms towards AI. Large-Scale Kernel Machines, 34, 2007.

Rodney Douglas, Terry Sejnowski & others: Future Challenges for the Science and Engineering of Learning. Report of an NSF workshop, 2007.

Tom Mitchell: The Discipline of Machine Learning. Report CMU-ML-06-108, Carnegie Mellon University, 2006.

Leo Breiman: Statistical modeling: The two cultures. Statistical Science, 2001.

Lecture Notes

These notes are meant to be as brief (and concise) as possible. They are not full tutorials or lecture scripts.

Some notes on gradient descent.: (Gradient descent, monotonicity & stepsize adaptation, covariant & natural gradient, co- and contra-variance, relation to Newton step, Rprop)
Factor graphs and belief propagation.: (Graphical models, probabilistic inference, message passing algorithms, loopy BP)
Gaussian identities.: (Normal and canonical representation, product of Gaussians, linear transformation, marginals & conditionals, entropy, Kullback-Leibler divergence, mixture of Gaussians, collapsing)
Basic 3D geometry (for robotics).: (Rotation representations, transformations (static, dynamic, affine, contra-/co-variant), kinematic chains, Jacobian & Hessian)
Markov Decision Processes.: (definition, Bellman optimality equation, Q-function, computing value functions, value iteration, direct solution, policy iteration, Q-learning, TD(lambda), eligibility traces)
Influence Diagrams.: (brief definition, inference methods in influence diagrams, relation to MDPs)
Stochastic Optimal Control.: (discrete time formulation, linear-quadratic-Gaussian case, Riccati equations, message passing formulation, classical cost formulation)