ajtulloch / hopfield-networks

Hopfield Networks for unsupervised learning in Haskell

Hopfield Networks in Haskell

Hopfield Networks are a simple form of neutral network, that can be understood as a simplified model of memory.

This implementation is based on the description in Information Theory, Inference, and Learning Algorithms by David MacKay, and John Myles White's Julia implementation on Hopfield networks on GitHub.

Installation

cabal install hopfield-networks

Demonstration

If your cabal binary directory is in your $PATH, you can directly run a demonstration.

$ hopfield_demonstration
Training patterns
--------
|X    X|
| X  X |
|  XX  |
|  XX  |
|  XX  |
| X  X |
|X    X|
--------
--------
|XXXXXX|
|X    X|
|X    X|
|X    X|
|X    X|
|X    X|
|XXXXXX|
--------
Validation
("Corruption error",4.8989797)
("Reproduction error",0.0)
"Original"
--------
|X    X|
| X  X |
|  XX  |
|  XX  |
|  XX  |
| X  X |
|X    X|
--------
"Corrupted"
--------
|X   XX|
| XX X |
|  XXX |
|  XX X|
|  XX  |
| X  X |
|XXX  X|
--------
"Reproduction"
--------
|X    X|
| X  X |
|  XX  |
|  XX  |
|  XX  |
| X  X |
|X    X|
--------
("Corruption error",4.8989797)
("Reproduction error",0.0)
"Original"
--------
|XXXXXX|
|X    X|
|X    X|
|X    X|
|X    X|
|X    X|
|XXXXXX|
--------
"Corrupted"
--------
|XX XXX|
|XX   X|
|X   XX|
|XX   X|
|X    X|
|X  X X|
|XXX XX|
--------
"Reproduction"
--------
|XXXXXX|
|X    X|
|X    X|
|X    X|
|X    X|
|X    X|
|XXXXXX|
--------