Perceptron

Question 1

Start

Answer 1

1958, by Frank Rosenblatt from Cornell Aeronautical Laboratory
Intended to be machine rather than program
First implemented in software for IBM 704
Implemented in custom built hardware as well

Question 2

What it does

Answer 2

It is a binary classifier
>function that can decide whether an input belongs to some specific class
Consists of input values and list of weights
Each weight (Wn) is scalar multiple modified by algorithm to enable perceptron learning
Also has input Xn
Has ‘b’ value bias modified by perceptron to enable it to learn desired behaviour
F(x) =
>1 if wx + b >0
>0 otherwise

Question 3

Improvement

Answer 3

Heavily experimented on in 60s by layering perceptrons by feeding them into each other
In 1980s resurgence of neural network research

Question 4

Artificial Neural Networks (ANNs) or (NNs)

Answer 4

Evolved form of perceptrons
Networks of artificial neurons
>have sets of inputs and weights for each input
>also have bias variable and maybe some others depending on output function
NNs have neurons arranged in layers and output of each neuron in a layer is then given as input to each neuron in following layer