https://github.com/iandanforth/spviewer

A very simple, visual example of the Spatial Pooler algorithm in operation
https://github.com/iandanforth/spviewer

Last synced: about 1 year ago
JSON representation

A very simple, visual example of the Spatial Pooler algorithm in operation

• Host: GitHub
• URL: https://github.com/iandanforth/spviewer
• Owner: iandanforth
• License: mit
• Created: 2013-11-05T15:54:23.000Z (over 12 years ago)
• Default Branch: master
• Last Pushed: 2014-06-06T12:49:00.000Z (about 12 years ago)
• Last Synced: 2024-08-10T14:16:48.224Z (almost 2 years ago)
• Language: Python
• Size: 185 KB
• Stars: 10
• Watchers: 5
• Forks: 3
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
spviewer

========

A very simple, visual example of the Spatial Pooler algorithm in operation.

Modifying the parameters set in exampleView.py can help you explore how

the spatial pooler works.

Caveats

- Currently this uses the new python implementation of the SP, which is not the

  default

- All implementations of the SP add noise to the output of the SP which can make

  things confusing. An item on my TODO list is to add an option to turn this

  noise off.

- The way the SP is used in the example is meant to replicate how convolutional

  neural networks are used. This is not how other examples like HotGym work, but

  works well for image data, and makes the SP comparable to more mainline work.

Requirements

- NuPIC

- PIL

- Pygame

Use:

python exampleView.py