Course Coordinator: 
Kenji Doya
Adaptive Systems

This course aims to provide common mathematical frameworks for adaptation at different scales and to link them with biological reality of control, learning, and evolution. We will look at different classes of adaptation problems using real-world examples of robot control, web searching, gene analysis, imaging, and visual receptive fields.

Introduction to machine learning algorithms and their application to modeling and analysis of biological systems.
Detailed Syllabus: 
  1. Introduction: variety of learning and adaptation
  2. Probability theory: entropy, information, Bayes theorem
  3. Pattern classification
  4. Function approximation
  5. Kernel methods
  6. Clustering, Mixture Gaussian, EM algorithm
  7. Principal Component Analysis, Self-organizing map
  8. Graphical models, Belief propagation
  9. Sampling methods, Genetic algorithms
  10. Kalman filter, Particle filter
  11. Reinforcement learning, Dynamic programming
  12. Decision theory, Game theory
  13. Multiple agents, Evolutionary stable strategies
  14. Communication and cooperation
  15. Presentation and discussion
Course Type: 
Midterm Reports 60% (2 x 30%), Final Exam 40%.
Text Book: 
Pattern Recognition and Machine Learning. Bishop (2006) Springer, New York
Reference Book: 
Matlab for Neuroscientists: An Introduction to Scientific Computing in Matlab I, by Wallisch et al. (2008) Academic Press
Prior Knowledge: 

Assumes good knowledge of statistics and ability to look at biological problems in a mathematical way.

OIST courses to complete beforehand:  B07 Statistical Methods