Advanced Message Passing Algorithms

A collection of message passing algorithms. Please read manuscript for the detail of algorithms.

Project tree

.

• bin
  • alpha_compare_mmse.py
  • alpha_compare.py
  • ep.py
  • marginal_loopy.py
  • plot_save.py
  • varying_loopy.py
• manuscript
• README.md
• requirements.txt
• src
  • alphaBP.py
  • factorgraph.py
  • loopy_modules.py
  • maxsum.py
  • modules.py
  • utils.py
  • variationalBP.py

Installation

First clone the repository:

$ git clone https://github.com/FirstHandScientist/expectation_propagation.git

Environment

Create a virtual environment with a python2 interpreter at 'path/to/your/evn/'

$ virtualenv -p python2.7 pyenv27

Then activate your environment:

$ source path/to/your/evn/pyenv27/bin/activate

and install the requirement file:

$ pip install -r requirements.txt

Run Experiment

To run experiment of comparing discussed algorithm in manuscript amp.pdf, at different value of SNR, active the pyenv27 first and run

$ python bin/ep.py

to compare the algorithms discussed in amp.pdf. You may configure the setting in scrips in bin to make experiment configuration.

For experiments in Figure 3 in amp.pdf, run

$ python bin/alpha_compare.py

$ python bin/alpha_compare_mmse.py

plot_save can be used to generate figures by saved experiment results.

python bin/plot_save.py path/to/figure/date

Acknowledge

Code factorgraph is adapted from repository.