What is a Neural Network?
A Neural Network is a mathematical model inspired by biological neurons, made up of nodes (neurons) and weights (connection strengths). Inputs flow through multiple layers to produce predictions; training adjusts those weights to improve accuracy. Every LLM, Transformer, CNN, and RNN is a different neural network architecture.
Practical understanding: you don’t need calculus to use them — think of a neural network as “a giant function you can learn from data.” Feed it a sentence, get the next token; feed it an image, get “this is a cat.” Claude Opus has hundreds of billions of parameters, each one a weight — that’s what “big model” really means.