SpiNNakerCSPs

A Spiking Neural Network Solver for Constraint Satisfaction Problems.

This repository will help you create a spiking neural network whose structure represents a constraint satisfaction problem and whose dynamics implements a stochastic search of its solution.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

In order to use this repository you should have a working installation of the PyNN implementation for SpiNNaker, you can follow the user installation instructions in:

http://spinnakermanchester.github.io/

Note: we currently use sPyNNaker7, but will eventually upgrade to be compatible with sPyNNaker8.

you will also require:

• numpy
• matplotlib
• simplejson

Installing

1. Just clone the project:

    git clone https://github.com/GAFonsecaGuerra/SpiNNakerCSPs.git
    cd SpiNNakerCSPs
   

2. Run some of the examples e.g. escargot.py file to test its working:

    python escargot.py
   

Usage

1. Edit the CSP_template.py file to implement your own SNN solver for a CSP of your interest. It simply creates an instance of the CSP class available in the snn_creator.py file and implements the methods therein. An initial analysis is possible with the tools in analysis.py.

2. You can use different strategies of noise implementation to improve the stochastic search.

Files

• README.md

• CODE_OF_CONDUCT.md

• .gitignore

• spinnaker_csp/

  • init.py
  • snn_creator.py
  • analysis.py
  • translators
    • init.py
    • spin2csp.py
    • sudoku2csp.py
    • world_bordering_countries.py
  • puzzles/
    • init.py
    • sudoku_puzzles.py

• examples/

  • init.py
  • cmp_australia.py
  • cmp_world.py
  • spin_lattice.py
  • sudoku.py
  • run.sh

Authors

This package has been developed by Gabriel Fonseca - GAFonsecaGuerra and Steve Furber - sfurber at The University of Manchester as part of the paper:

"Using Stochastic Spiking Neural Networks on SpiNNaker to Solve Constraint Satisfaction Problems" Submitted to the journal Frontiers in Neuroscience| Neuromorphic Engineering

See also the list of contributors who participated in this project.

Contributing

If you want to contribute to the development of this repository please read CODE_OF_CONDUCT.md for details on our code of conduct, and CONTRIBUTING.md to know the process for submitting pull requests to us.