Neuromorphic Computing?

I am not really interested in this AI trend, because all trends related to AI, machine learning and deep learning are all depending on software levels. Two years ago, when I was doing a little research about RISC (Reduced Instruction Set Computer) I found something which is called URISC or Ultimate Reduced Instruction Set Computer. The first URISC I discovered was a one instruction set machine called SUBLEQ.

I did more research and I found SUBLEQ (and other OISCs) are implementations of Turing’s machine. Later, in Wikipedia, I found this article :

I asked a friend what the hell is ZISC? and he answered that this type of processors are designed to act as a brain, with a hardware layer neural network implemented. I did more research and surprisingly, I found nothing about a free or open-source ZISC implementation!


OK, I didn’t give up. I searched about “neural network” in logisim (anyway, it’s a great tool for logical simulations!) and I found this video on YouTube :

I asked them for more information and schematics, but they didn’t care. So, I did another search about Logical gates perceptron and I found this :

https://www.robotshop.com/community/forum/t/my-artificial-neuron-approach/4735

But this project is way simpler than what I thought about neuro-computers. So, I decided to study neural networks.

Studying neural networks requires a huge knowledge of mathematics and I found sources who simplified that math (or, they teach you what you need, at least better than stupid teachers). For example, I found a guy on YouTube called Siraj Raval (his github) and he teaches concepts of AI and Machine Learning in the simplest ways possible with code and math!

After I learned about neural networks, I found a feedforward neural network would be easy to implement on hardware layers. But I couldn’t model what I learned in a bunch of boolean functions. So, I asked this in the reddit :

I didn’t get any answer even in LibreCores gitter and Kaggle. So, I decided to write this article here. I hope someone will read this and answer me :))

Thanks for reading my article :)