Multilayer perceptron
Also: MLP, feedforward neural network, multi-layer perceptron
A feedforward network in which every neuron in one layer connects to every neuron in the next, with a nonlinear activation between layers. With even one hidden layer of enough units it can approximate any continuous function.
h = φ(W₁x + b₁), ŷ = W₂h + b₂
Example
An MLP for 28×28 digit images with layers 784 → 128 → 10 has 784·128 + 128 + 128·10 + 10 = 101,770 parameters, and unlike a single perceptron it can solve XOR.