Neural Network

A neural network is a type of computer program inspired by how the human brain works. Your brain has billions of neurons connected together, passing signals back and forth. A neural network has digital "neurons" (usually called nodes) arranged in layers and connected to each other, passing numbers back and forth.

A basic neural network has three parts: an input layer that receives data (like the pixels of an image), one or more hidden layers that process the data and find patterns, and an output layer that gives you the result (like "this is a cat"). During training, the connections between nodes get stronger or weaker depending on whether the network got the answer right or wrong. Over thousands of examples, these connections adjust until the network becomes accurate.

You can think of each node as making a tiny decision: "Based on the signals I am getting, should I pass a strong signal forward or a weak one?" On its own, one node is useless. But millions of them working together can do remarkable things -- like understand the meaning of a sentence, generate a painting, or predict tomorrow's weather.

Neural networks come in different shapes and sizes. Simple ones might have just a few layers. The most powerful ones, like those behind GPT and Gemini, have billions of connections across hundreds of layers. These are the "deep" neural networks that power deep learning.