Advanced Topics in Machine Learning WS 2009/2010
Dozent:
- Prof. Dr. Andreas Schilling
- Dr.-Ing. Martin Giese
- Dr.-Ing. Cristobal Curio
- Dr. Dominik M. Endres
Sprechstunde: n.V.
Zeit: Mi 10-12
Umfang: 2+2
Beginn: Mi. 21. Oktober 2009
Ort: Sand 14, Seminarraum C424
Turnus: 2-semestrig
Prüfungsfach: Praktische Informatik, Theoretische Informatik
Voraussetzungen: Vorlesung Grundlagen des Maschinellen Lernens. Beachten Sie, dass die Vorlesung in Englisch gehalten wird.
Beschreibung
Advanced Topics in Machine Learning (with many practical exercises)
1. Bayesian statistics - introduction
- Two kinds of probability
- Why bother going Bayesian? The theorems of Cox and Fine
- Alternative probability theories
2. Important probability distributions
- Discrete
- Continuous
3. Graphical models
- Bayesian networks
- Conditional independence and D-separation
- Markov Random Fields
- Relationship between directed and undirected graphical models
4. Approximate Inference
- Variational inference
- Exponential family distributions
- Local variational methods
- Expectation propagation
5. Sampling Methods
- Basis Sampling Algorithms
- MCMC Sampling
- Gibbs and Slice Sampling
- Deep Believe Nets
6. Continuous Latent Variables
- Revisiting Principal Component Analysis (PCA)
- Probabilistic PCA and its variants
7. Kernel Methods
- Dual representations
- Radial Basis Function Networks
- Gaussian Processes
8. Sparse Kernel Machines
- Maximum Margin Classifiers
- Relevance Vector Machines
9. Sequential Data
- Markov Models
- Hidden Markov Models
- Linear Dynamical Systems
Literature:
Bishop, C.: Pattern recognition and machine learning, Springer 2006, ISBN 978-0-387-31073-2.
Übung
Tutoren:
Zeit: n.V.
Beginn: n.V.
Ort: siehe Vorlesung
Beschreibung:
Hier oder über den Link "Übungen" kann man die Übungsseite erreichen.