Proposal: End-to-end use case simulation

Here I wanted to propose a way to potentially explain, how the network (math) will work to everyone that can read a bit of code. The idea I had is to create a jupyter notebook in Google Colab. If you do not know what that is, it looks like this:


Basically it is a python script where you can use markdown to comment in between script executions.

Use case
The use case I have in mind is one from the Q&A document. The password of a user is stored inside Nillion. Later the user can authenticate by comparing the password that is store with one he typed into an app or something.

Structure
The purpose of the notebook is to show the inner workings of the network. From the pre-processing stage, OTM, calculation to the result reconstruction. At each point the user has the possibility to check and verify the computed outputs. As it is simply a python script he could easily try to break the masking or temper with the system in some way.

Further ideas
This is only one use case that can be shown. Additional helpful notebook ideas would be:

  • A simulation with active adversaries
  • Performance improvements in the preprocessing stage
  • Performance simulation

@Jambutties Is it possible for the team to implement something like that? Or at least review it? Of course we have to wait for the TWP to be released first, but while it is reviewed something like this could be created in parallel :wink:

1 Like

I love Jupyter Notebooks, I use them everyday for my science stuff! I only use code editors for packaging and delivery otherwise all the good things happen in Jupyter! Love the idea of educational material in notebooks!