David Smith & Zayd Hammoudeh Practice Final Question #9: Give the formal definition of '''perceptron'''. Explain and give an example of a '''feed forward network''' is and what a '''recurrent network''' is. A '''perceptron''' is a neural network where the activation function (''g'') takes a real valued input and returns a real valued output. In a standard perceptron, the activation function is the threshold function while in a sigmoid function the activation function is the logistic function. A '''feed‐forward network''' is a neural networks where the neuron outputs only move in a single direction (i.e. forward). No back lines are allowed in the network so a neuron’s output can never form part of its own input signals. A '''recurrent network''' is a neural networks where the outputs of neurons are looped back to eventually form part of the neuron’s own inputs (either directly or through a predecessor node). Example Networks: [[https://cdn.mediacru.sh/U/UuoZ5ulCik1T.svg]]