https://github.com/johnmyleswhite/HopfieldNets.jl

Hopfield networks in Julia
https://github.com/johnmyleswhite/HopfieldNets.jl

Last synced: 3 months ago
JSON representation

Hopfield networks in Julia

• Host: GitHub
• URL: https://github.com/johnmyleswhite/HopfieldNets.jl
• Owner: johnmyleswhite
• License: other
• Created: 2013-07-28T15:30:39.000Z (almost 11 years ago)
• Default Branch: master
• Last Pushed: 2015-05-05T17:19:39.000Z (about 9 years ago)
• Last Synced: 2024-01-18T11:01:12.125Z (5 months ago)
• Language: Julia
• Size: 215 KB
• Stars: 14
• Watchers: 7
• Forks: 12
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Lists

• alex-mikhalev-awesome-stars - HopfieldNets.jl - Hopfield networks in Julia (Julia)

README

        
HopfieldNets.jl

===============

# NOTICE

**This package is unmaintained. Its reliability is not guaranteed.**

# Introduction

Discrete and continuous Hopfield nets in Julia.

# Usage Example

We'll restore a corrupted representation fo the letter X specified as a vector of -1's and +1's:

	using HopfieldNets

	include(Pkg.dir("HopfieldNets", "demo", "letters.jl"))

	patterns = hcat(X, O)

	n = size(patterns, 1)

	net = DiscreteHopfieldNet(n)

	train!(net, patterns)

	settle!(net, 10, true)

	Xcorrupt = copy(X)

	for i = 2:7

	     Xcorrupt[i] = 1

	end

	Xrestored = associate!(net, Xcorrupt)

	all(Xcorrupt .== X)

	all(Xrestored .== X)

# Sources

The idea for the letters demo is taken from an implementation of [Hopfield nets in Ruby](https://github.com/bartolsthoorn/hopfield-ruby). The code is based on the presentation of the theory in Mackay's book and in Storkey's learning rule paper.