jspωiki
Multilayer Perceptron

Overview#

Multilayer Perceptron (MLP) is a class of Feedforward Neural network.

Multilayer Perceptrons consists of at least three layers of nodes. Except for the input nodes, each node is a neuron that uses a nonlinear Activation Function.

Multilayer Perceptrons utilizes a supervised Learning technique called backpropagation for training. Multilayer Perceptrons multiple layers and non-linear activation distinguish MLP from a linear perceptron. Multilayer Perceptrons can distinguish data that is not linearly separable.

Multilayer Perceptron are sometimes colloquially referred to as "vanilla" Artificial Neural networks, especially when they have a single hidden Layer

More Information#

There might be more information for this subject on one of the following: