Deep Learning Dictionary - Lightweight Crash Course

Perceptrons - Deep Learning Dictionary

A perceptron, also referred to as a single layer perceptron, is an algorithm that accepts an input and maps it to a binary output based on a set of parameters called weights.

The algorithm to determine the output from a perceptron is relatively simple:

1. Calculate the weighted sum of inputs
2. Compare the weighted sum to threshold value to determine binary output

The weights of the perceptron are first randomly initialized, and then during a training process, the perceptron learns the optimal values for these weights that result in the most accurate binary output.

A single layer perceptron is considered the simplest kind of artificial neural network. We'll learn much more about these networks in later entries.