Neural network
A computing system made of layers of simple units whose connection strengths are learned from data.
Also: artificial neural network
A neural network takes numbers in, passes them through layers of units, and produces numbers out. Each connection has a weight, and each layer mixes its inputs using those weights and then applies a simple non-linear function. Training adjusts the weights, often billions of them, so that outputs move closer to the desired answers. The idea dates back to the 1940s and 1950s. Deep networks, with many layers, took off in the 2010s once graphics processors and large datasets made training them practical. Every language model on this site is a neural network.