Introduction

Belief Nets:

• Represent and reason about uncertainty (probability distributions)

Supervised learning:

• map input to output to match training data.

Can use similar ideas to learn probabilty distributions? Yes. Input:

• Give a list of settings (that's the training data).