Learning Dynamical Systems


This page is under construction.
This document provides a high-level overview and table of contents for the tutorial. If you intend to follow the course, you might add the URL for this document to your hot list. The following paragraphs describe the sections of the tutorial and provide links to the corresponding html documents.

Dynamical Systems

An introduction to dynamical systems including some frequently asked questions and pointers to supplementary readings.

Graphical Models

It will be convenient at times to represent the qualitative structure of dynamical systems using appropriate graphical notation. In this section, we introduce such a notation and use it to characterize various learning problems.

Learning Techniques

The document corresponding to this section is very much like the current document. It consists of an introduction followed by short summaries and links to other documents. The documents linked to this page provide much of the content on learning in this tutorial.

Miscellaneous Techniques

There is a lot of mathematics associated with machine learning techniques; much more than we could possibly survey. However, there are a few techniques that appear often enough that it will help to have some passing knowledge of. Currently, we are planning to cover signal processing and function approximation applications of the discrete Fourier transform and singular value decomposition. In addition, we provide a quick introduction to Gibbs sampling which plays an important role in dealing with incomplete information.

Vocabulary

Eventually this document will contain a glossary of terms to supplement the readings. As with most of the other pages associated with the tutorial, this page will grow and change with feedback from participants.

References

This document provides a listing of the readings in bib format with links to on-line sources where available.

Implementation

There are Mathematica packages and notebooks associated with various sections of the tutorial. This document describes the structure of those packages and notebooks and provides information about installing the software on your system.

Searching On Your Own

Try searching the web for ``dynamical systems'' or ``time series'' using your favorite search engine, (e.g., Yahoo, or Lycos) or look in Wolfram Research's on-line Mathematica sources for useful packages and interesting notebooks.
Back to Tutorial