# Switch Net 4 - linking multiple tiny neural layers with a connectionist device

**URL:** <https://discourse.numenta.org/t/switch-net-4-linking-multiple-tiny-neural-layers-with-a-connectionist-device/9095>\
**Category:** Lounge\
**Created:** [November 2, 2021, 5:38am UTC](https://discourse.numenta.org/t/switch-net-4-linking-multiple-tiny-neural-layers-with-a-connectionist-device/9095 "2021-11-02T05:38:49Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![SeanOConnor](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/seanoconnor/32/3441_2.png) [@SeanOConnor](https://discourse.numenta.org/u/SeanOConnor)\
**Post date:** [November 2, 2021, 5:38am UTC](https://discourse.numenta.org/t/switch-net-4-linking-multiple-tiny-neural-layers-with-a-connectionist-device/9095/1 "2021-11-02T05:38:49Z")

</div>

You can link multiple small neural network layers together using the fast Walsh Hadamard transform as a connectionist device.  
Java type code:  
[https://discourse.processing.org/t/switch-net-4-neural-network/33220](https://discourse.processing.org/t/switch-net-4-neural-network/33220)  
JavaScript code:  
[https://editor.p5js.org/seanhaddps/sketches/ERaMwFcej](https://editor.p5js.org/seanhaddps/sketches/ERaMwFcej)  
Full Screen: [https://editor.p5js.org/seanhaddps/full/ERaMwFcej](https://editor.p5js.org/seanhaddps/full/ERaMwFcej)  
In the code I used width 4 neural layers, yeh, with the Beyond ReLU idea I’ve mentioned before.  
You can use different width non-linear layers 2,8,1,16… or just use parametric non-linear functions.  
The WHT is a connectionist device because it connects each input to all the outputs coequally.

 ![wht1](https://canada1.discourse-cdn.com/flex030/uploads/numenta/original/2X/7/7cd2dfb7cbfbc043fd0a43b1cdb62f7a9d42b5da.jpeg)  
Allowing you to turn sparse neural networks into pseudo-fully-connected networks for example.