Description
Perceptron Models
The perceptron is a kind of binary classifier that maps its input x (a vector of type Real) to an output value f(x) (a scalar of type Real) calculated as
f(x) = + b
where w is a vector of real-valued weights and is the dot product(which computes a weighted sum). b is the 'bias', a constant term that does not depend on any input value.(x)
Presentation Transcript
Perceptron Models : Perceptron Models The perceptron is a kind of binary classifier that maps its input x (a vector of type Real) to an output value f(x) (a scalar of type Real) calculated as
f(x) = + b
where w is a vector of real-valued weights and is the dot product(which computes a weighted sum). b is the 'bias', a constant term that does not depend on any input value.(x)
Slide3 :
x(j) denotes the j-th item in the input vector
w(j) denotes the j-th item in the weight vector
y denotes the output from the neuron
δ denotes the expected output
α is a constant and 0 < α < 1
the appropriate weights are applied to the inputs that passed to a function which produces the output y
The weights are updated after each input according to the update rule below:
w(j)' = w(j) + α(δ − y)x(j)
Famous Minsky and Papert Book:Perceptrons (1969) : Famous Minsky and Papert Book:Perceptrons (1969) Showed that Perceptrons couldn’t solve general simple nonseperable problems (eg. XOR)