#AI demystified: a decompiler
To prove that any "artificial neural network" is just a statistically programmed (virtual) machine whose model software is a derivative work of the source dataset used during its "training", we provide a small suite of tools to assemble and program such machines and a decompiler that reconstruct the source dataset from the cryptic matrices that constitute the software executed by them.
Finally we test the suite on the classic #MNIST dataset and compare the decompiled dataset with the original one.
I hope that one day people will really understand and embrace Wirth's legacy.
We will have a better cybernetic world then.
Ok, the article is just a bit superficial, but the cover image is basically me every damn day of my life.
The most important program in my whole career.
Think about it....
... a little more...
But that's not the most interesting lesson you can teach them.
Through this DRM violation you can show them how fragile is #Capitalism. 😉
And it's even more general: there are always ways to subvert an oppressive system, usually provided by the oppressors themselves.
QOTO: Question Others to Teach Ourselves. A STEM-oriented instance.
An inclusive free speech instance.
All cultures and opinions welcome.
Explicit hate speech and harassment strictly forbidden.
We federate with all servers: we don't block any servers.