Chapter 05 · Section I · 16 min read
Neural network basics
A neuron is a function. A network is many functions glued together. Everything else — ChatGPT, image generators, modern OCR — is detail on top.
A single artificial “neuron” is just a mathematical function. It takes some inputs, multiplies each by a weight, adds them up, and squishes the result into a small range. That’s it.
A neural network is a large collection of these neurons, organised in layers. The output of one layer becomes the input of the next. Modern systems — ChatGPT, the image generators that drew this course’s illustrations, the OCR on your phone — are neural networks with billions of these neurons.
The “learning” is the process of finding good values for the weights, given examples.
This section is a stub. The full version will build one neuron from scratch using Khalti transaction features as inputs.