Reading Class Hidden Markov Models and Dynamical Systems

Target Participants:
This course is a Reading Class and is primarily addressed to the curricula of all students in the PhD-Programs of the Research Training Group Scientific Computation (PaSCo GK) and of the International Graduate School Dynamic Intelligent Systems. Other participants like e. g. Master students are welcome.

Topic Description:
Hidden Markov Models (HMMs) have been established as a fundamental mathematical modeling technique with applications ranging from speech recognition through bioinformatics to spam filters. Mathematically they can be understood (in the simplest case) as discrete-time discrete-state stochastic dynamical systems. This view is instructive in particular when it comes to understanding the behaviour of the fundamental algorithms for HMMs. In this course we are going to read the book Hidden Markov Models and Dynamical Systems by A. M. Fraser, 2008, and possibly further related texts.

Dates:
All on Wednesdays in room D3.230.

• 06.05.2009, 09:15: Introduction (Christian Horenkamp)
• 13.05.2009, 09:15: Basic Algorithms I (Kathrin Flaßkamp)
• 27.05.2009, 09:15: Basic Algorithms II (Robert Timmermann)
  
• 10.06.2009, 09:15: Variants and Generalizations (Mirko Hessel-von Molo)
• 24.06.2009, 09:15: Continuous States and Observations and Kalman Filtering (Mirko Hessel-von Molo)
  
• 22.07.2009, 09:00: Performance Bounds and a Toy Problem (Christoforos Raptopoulos)
  
• 22.07.2009, 11:00: Obstructive Sleep Apnea (Robert Preis)
  

Organization:
The first meeting including the distribution of the topics took place on Wednesday, 22.04.2009 at 09:15 in room E2.304.

Prerequisites:
Basic courses in probability theory, linear algebra, differential equations. Somewhat more experience with dynamical systems theory and information theory will be helpful, but not strictly necessary.

Literature:
We will read the book Hidden Markov Models and Dynamical Systems by A. M. Fraser, 2008. You can find the table of contents here.