Nov 17
First: discuss homework. Particle filter results?
Second: talk about what's what with the rest of the term :
Tue Thu
Nov 17 19 this week
24 thanksgiving
Dec 1 3 penultimate
8 final project due Friday
So it's time to get cracking on final projects.
I don't want to assign too much in the way of
homework over these last few weeks, so that
you have time to work up something.
You all know the drill: pick something that expands
on a topic we've looked at this term. Please do
something different than your first project.
Reasonable topics include :
- a neural net (what we're doing next)
- a data science exploration (similar, machine learning & classification)
- a Bayesian network (what we just finished)
- A particle filter e.g. robot navigation (a numerical approach to a Bayes network)
- text classifier using naive bayes (e.g. spam filter)
- a logic system (if you didn't do that for the first project)
- something else ... please discuss with me.
The neural net material in our textbook is in Chapter 18,
giving an overview of machine learning including
- what sorts of problems / approaches are used
- its connection with curve fitting
- decision trees (which are a precursor to the "random forests" I've seen lately)
- neural nets
It's worth browsing through that, but I would rather use
a different source as our shared reading for these last few weeks :
Warning : there's lots of math and numerical stuff here.
How much we dive in is in large part up to you.
For today, I'll wave my hands for a bit of the "big picture",
and see where we go from there.
Other related resources :
tensorflow.org
Coursera : Neural Networks for Machine Learning (Geoffrey Hinton)
wikipedia articles
another neural net tutorial
Brain - javascript supervised machine learning ; neural net & bayes classifier (
github source)
Kagle - machine learning competitions for big bucks
Random forests in python
Data Science (not quite what I plan to cover now, but related)
- http://learnds.com/ - ipython notebooks looking at Linear Regression, Logistic Regression, Random Forests, & K-Means Clustering
Dylan's particle filtering references
