Weight

Appears in 1 paper

A number attached to an input connection in a neural network.

As used in Paper 02 — The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain →

A number attached to an input connection in a neural network. It determines how much influence that input has on the neuron's output. High positive weight = strong positive influence. Negative weight = the input pushes the neuron toward not firing. Weights are the things that get learned during training.