Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 16

Introduction:-

• Artificial Neural mimic human brain nervous


system which consist with large number of
interconnections of large number of neurons.
• These neurons achieve tasks of perception
and recognition in short time compared to
computers.
• Brain interconnections and network massively
parallel needed for computer .
• Still computer very fast ( nano-second) but human is
milli second speed in counting) but human brain
processing is much faster i.e in face recognition.
• This is because the network in human neurons is
massively parallel network of neurons.(10 billion of
neurons with 60 trillions of interconnections)
• Computer can do small task compared to brain for
interconnections shrotage
Artificial neural net work advantages :-
• In Non-linearity the relation between the input and the
output could not be expressed with simple equations.
• In linearity the relation between input an output can
be expressed with simple equation.
• The artificial neural network can solve nonlinearity
problem which is distributed in nature artificial
intelegent can do input output mapping where output
follow input.
• by adjusting free parameters of the system we can
manipulating our inputs closer desired output.
• Here we use process of learning many times evolving
teacher
Synaptic terminals
• The inputs are the signals need to be
processed The weight is an amount from 0 to
1 affecting the output.
• The neuron where the x input multiplied by
the weights and summed up’

Nonlinear processing unit

• The activation function where the sumed


valuesofare
What we change is strength multiplied
connection by certain
W adjustment equation
to get desired to
respnse
give the final output y.
W0 is the bias between -1 and+1
g is the activation 0 if the out put a is less than zero
1 if the output a is greater than zero
Multilayer composed of input
neuron layer , hidden layer or
layers and output layer

You might also like