back-propagation algorithm

computing

Written by

The Information Architects of Encyclopaedia Britannica

Last updated

July 16, 2026

Quick Summary

The back-propagation algorithm is a method used in training artificial neural networks. These networks are computer programs inspired by the human brain, designed for tasks like problem solving and machine learning.

Back-propagation allows a neural network to learn by adjusting the connections between its artificial neurons. When the network produces an output, the algorithm compares it to the desired output. If there's a difference, it calculates the error and "propagates" it backward through the network to modify the connection weights. This process helps the network improve its accuracy over time by learning from examples. This technique is fundamental to how many modern AI systems learn and improve.123

Summary created from Britannica articles using AI. AI can make mistakes, so verify using Britannica articles.

3 Britannica Sources

Ask Britannica AI about this topic

Britannica doesn't have a full article on this topic, so we've curated a list of our relevant coverage.

neural network
…feedback mechanism, known as a back-propagation algorithm, that enables it to adjust the connection weights back through the network, training it in response to representative examples. Second, recurrent neural networks can be developed, involving signals that proceed in both directions as well as within and between layers, and these networks…
Read More
…feedback mechanism, known as a back-propagation algorithm, that enables it to adjust the connection weights back through the network, training it in response to representative examples. Second, recurrent neural networks can be developed, involving signals that proceed in both directions as well as within and between layers, and these networks…
Read More