Neural Network is a mathematical
model of what goes in our mind to
perform particular task or function.
It is implemented using electronic
components or simulated software.
It is a system with many elements
4. Biological Neural Network
Soma: body of the cell that houses
the nucleus, in which the neuron's
main genetic information can be
zones that receive
5. Biological Neural Network (Cont,)
Synapses are elementary structural
and functional units that mediate the
interactions between neurons.
Axon: transmission line
that sends messages.
6. Artificial Neural Network
An artificial neural network (ANN) is a
system composed of many simple
processing elements (Neurons)
operating in parallel whose function is
determined by network structure,
connection strengths (Weights), and
the processing performed at
computing element or nodes.
7. Difference Between Human Behavior and
Neural Network Behavior
Human Behavior Neural Network Behavior
remember certain things completely, Once we create a neural network, we
partially depend on capacity for train it to become expert in an area.
If we do not practice what we learned, Once fully trained, a neural net will not
we start to forget. forget.
The first 10 processes may be If the results are repeatable it will be
accurate, but later we may start to accurate.
make mistakes in the process.
Faster inprocessing data and
Nonlinearity: Perform operations that linear
Fault tolerance: When one element fail NN
continue without reduce parallelism.
Adaptivity: NN has a capability to adapt their
synaptic weights to changes in the surrounding
Contextual information: Every neuron in the
network is potentially affected by the global
activity of all other neurons
9. Neuron Model
Neuron is a simple processing unit.
The sole purpose of a Neuron is to receive
electrical signals, accumulate them and see
further if they are strong enough to pass
Single neuron is useless. It is the complex
connection between them (weights) which
makes brains capable of thinking and having a
sense of consciousness.
10. Neuron Model (Cont,)
Neuron Consist of:
Inputs (Synapses): input signal.
Weights (Dendrites): determines the importance
of incoming value.
Output (Axon): output to other neuron or of NN.
11. Neuron Model (Cont,)
Output: is calculated by inputs which are
multiplied by weights, and then computed by a
mathematical function which determines the
activation of the neuron.
12. Neuron Model (Cont,)
Output when single perceptron (neuron) is used.
The dark blue dots
represents values of true
and the light blue dot
represents a value of
14. Multilayer Neural Network
This network is feed-forward, means the values are propagated in
one direction only.
Input layer: takes the inputs and forwards it to hidden layer
Middle layer: Without this layer,
network would not be capable of
solving complex problems.
Output layer: This layer
consists of neurons which output
Weights: for every neuron there
Are weights that for every input
updating network architecture and connection weights so
that network can efficiently perform a task.
Basic Learning Procedure
run an input pattern through the function
calculate the error (desired value – actual value)
update the weights according to learning rate and error
move onto next pattern
Occure when NN memorize patterns and loose the ability
of generalization. Problem is when to stop learning
16. Learning Paradigm
Supervised The correct answer is provided for the
network for every input pattern Weights are adjusted
regarding the correct answer.
Unsupervised Does not need the correct output the
system itself recognize the correlation and organize
patterns into categories accordingly.
Hybrid A combination of supervised and unsupervised,
Some of the weights are provided with correct output while
the others are automatically corrected.
Neural Network is a modeling for human brain
Neuron is the basic unit of NN
To adapt NN to perform operation we want, it has to be
Most practical form of NN is the one that has multilayer
Try to avoid overfitting