Determine what data visualisations can be derived from a neural model

[StefanKennedy]

Update:
Niall and Stefan have contributed to this issue by highlighting some libraries and visualizations that can be performed on Keras and Tensorflow models and the like. Thus far, no effort on the frontend to implement neural visualizations.

[StefanKennedy]

A way to visualize convolutional neural networks: https://medium.com/@awjuliani/visualizing-neural-network-layer-activation-tensorflow-tutorial-d45f8bf7bbc4

We're not working with images, but might be a lead to separate the layers to test them independently

[StefanKennedy]

Not sure if we'll end up with Keras, but this is interesting: https://keras.io/callbacks/

Tensorboard may also show some interesting visualisations

[Deniall]

Here's an interesting JS library, that acts as a wrapper on Tensorflow or Keras models for visualization purposes, you can see what happens in the network when a new sample is entered:

https://medium.freecodecamp.org/tensorspace-js-a-way-to-3d-visualize-neural-networks-in-browsers-2c0afd7648a8