Neural Network

Definition

A computing system loosely inspired by biological brains, consisting of interconnected nodes (neurons) organized in layers. Each connection has a weight that's adjusted during training. Simple in concept but capable of learning incredibly complex patterns when scaled up. The foundation of modern AI.

How It Works

A neural network is a computing system loosely inspired by biological brains, built from layers of interconnected nodes (neurons) that process information. Each connection has a weight that determines its strength, and the network learns by adjusting these weights during training. Modern AI — from image recognition to language models — is built on neural networks.

The architecture varies by task. Convolutional neural networks (CNNs) excel at image processing by detecting spatial patterns. Recurrent neural networks (RNNs) were used for sequential data like text before transformers took over. Transformers, the architecture behind modern LLMs, use attention mechanisms to process sequences in parallel. All of these are neural networks — they just wire their neurons differently.

What makes neural networks powerful is their ability to learn complex patterns from data without being explicitly programmed. You don't tell a neural network the rules for recognizing cats — you show it millions of cat photos and it figures out the patterns. This learning ability, combined with massive scale and enormous datasets, is what produced the AI revolution we're living through. The networks used today have billions or trillions of parameters, far beyond anything researchers imagined when the field started decades ago.