[Noisebridge-discuss] ML 11/28 @ 7:00pm: Sampling in Python and Ising Models

[Mike Schachter]

Hey there,

On Monday evening at 7:00pm *sharp* (note the time!) I'd like
to present a short workshop on sampling from probability
distributions in Python. We'll cover the very basic stuff,
like generating random numbers and plotting their histograms
using matplotlib.

Then we'll step it up a bit and talk about Gibb's sampling for
a fun thing called an Ising model:

http://en.wikipedia.org/wiki/Ising_model

We'll talk about how to sample from an Ising model and
potentially generate really cool correlated TV noise images.

Also, the restricted boltzmann machine is a type of Ising
model that's used for deep belief nets, so learning how to
do Gibb's sampling for an Ising model would be quite
helpful for anyone interested in deep nets.

Before you show up, please have the following Python
packages installed:

Numpy
Scipy
NetworkX
Matplotlib

See you then! Since the HTML/CSS/Javascript people are
going to meet up around 8pm, I'm hoping we can run through
things somewhat quickly and not interfere much with their stuff.

  mike