Neural Network
Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary
A computational model built from layers of connected nodes that learns patterns from data by adjusting the strengths of the connections between them.
What is a Neural Network?
A neural network is a computational model built from layers of connected nodes, each performing a simple calculation and passing the result to the next layer. The network learns by adjusting the strength of those connections until its outputs match the training examples.
The full formal name is artificial neural network – same concept, longer label. Neural networks handle tasks where the rules are too messy to write by hand: recognizing a face, transcribing speech, translating a sentence.
How a Neural Network Works
Data enters through an input layer, flows through one or more hidden layers, and exits through an output layer as a prediction. Each node multiplies its inputs by learned weights, sums them, and passes the result through an activation function that lets the network capture non-linear patterns.
Training works by comparison and correction. The network makes a prediction, a loss function measures how wrong it was, and backpropagation adjusts every weight to shrink that error. Repeat this over thousands of examples and the network’s answers converge on the right ones.
The result is knowledge stored as millions of numbers rather than readable rules. That is why neural networks raise explainability questions that simpler models do not.
Types of Neural Networks
- Feedforward neural networks pass data in one direction, input to output, and serve as the baseline architecture for prediction and classification tasks.
- Convolutional neural networks (CNNs) scan images with small filters that detect edges, textures, and shapes, which makes them the standard for vision tasks.
- Recurrent neural networks (RNNs) process sequences such as text or time series, carrying a memory of earlier inputs forward.
- Generative adversarial networks (GANs) pit two networks against each other to produce new data that resembles the training set, such as synthetic images.
Example of a Neural Network
A phone’s face unlock runs a convolutional neural network. The camera image enters as raw pixels; early layers detect edges and contours; deeper layers combine them into facial features like the spacing of eyes and the shape of the jaw.
The output layer condenses this into a compact signature and compares it against the enrolled face. During training, the network’s weights were tuned on millions of faces, which is why it still works in dim light or at an angle.
Related AI terms: Artificial Neural Network · Deep Learning · Hidden Layer · Backpropagation · Explainability
Did you like the Neural Network gist?
Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.
Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About