Oct 22
midterm projects
First order of business : discuss where you are
with your midterm projects and what you're working on.
Bayes
I would like to work through chapters 14 & 15,
at least for the big picture stuff if not
all the details.
14 is about Bayes Networks, what they are,
what the equations that the model are,
and how to simulate them - explicitly
or by estimating the probabilities
with sampling methods.
15 adds the notion of time, and covers
Markov Chains and dynamic systems.
The robot location algorithm (kalman filters) is particularly cool, IMHO.
After that we'll move to chapter 18, machine learning.
In class :
- discuss the general ides from chapter 14
- browse the slides for the earthquake model
- run through the monte-carlo ideas
- try some of the problems together in class (I have solutions we can look at)
- 14.11 nuclear power plan
- 14.12 telescopes
- 14.18 rain/sun monte-carlo sampling
Your mission : read chapter 14 & 15 ... I'll have a homework assignment
and/or some projects due the week after next or so.
What happened: after discussing the big ideas of chapter 14 (networks, the probability rules, markov chains, Gibbs Sampling) we did about half of 4.12.
