1.0 Neural Network

Neural networks are computing systems loosely inspired by how brains process signals.

In general, a biological neuron has three essential parts:

1. Dendrites
2. Axons
3. Synapse

Purely from a computational perspective a basic equation to emulate how a Neuron thinks goes like this. This is the most basic Mccullough-Pitts Function model of a Neural Network.

$$S=\sum _{i=1}^n{w}_i{x}_i$$

$S$ = The Sum $n$ = The $i$ total number of inputs $i$ = The number of inputs (1) $w_i$ = The Weights (Learned by the model and assigned by the model based on importance) $x_i$ = The inputs (Given to the model by data from external factors)

This equation essentially Sums the weights and the inputs together. The goal for a Neural model is to get a value as high as possible.

There is a baseline value that you can assign to the model, if it passes a certain threshold the model will automatically assign a higher weight and leaves ones with low weight and low input. This is called a Threshold Function.

Therefore, the Neuron is given the simple Mccullough-Pitts Function. We will use $\theta$ as the Threshold Function.

Using a simple example if $S$ is $\ge \theta$, then it passes and goes on to the next step, no change is needed

But if it goes below, then the weight gets adjusted.

As such, Neural Networks are incredibly heavy as it requires potentially billions of calculations per minute. Therefore, a GPU is used to…

A process called 1.5 Quantization is used to..