ARTIFICIAL
NEURAL
NETWORKS
BY: - A D I T YA PA N D
C L A SS : - 9 ’ E ’
ROLL:-01
UID:-SBPS/13-
HOW DOES AN ARTIFICIAL NEURAL
NETWORKS WORKS ??
■ An artificial neural network is originally designed to self
programme and adopt using algorithms to categorise and
understand images,concept and photographs.The neural
system receives inputs from other sources in form of
images and patterns.To be able to recognise the
input,developers design the artificial network into layers
this it operates in parallel.
■The input layer resembles the dendrites on the human
brain while the hidden layer resembles the bod cells and is
placed in between input and output layers and the output
layer resembles the synaptic output of the brain.The
artificial neurons take the inputs from hidden layers
according to synaptic weight.Synaptic weight is the
strength or amplitude of connecton among nodes.The
weighted inputs,then,generated an output to the output
layers.
WHAT ARE THE TYPES OF AN
ARTIFICIAL INTELLIGENCE ??
The artificial network computational model work
similarly to the human nervous system.The
technology is the implemented in various
fields.However,the base of the artificial network
is the mathematical operational or known as
algorithms.Here are some types of artificial
networks.
Feedforward neural network
Radial basis neural network
Kohonen self organising
Recurrent neural network
Convolutional neural network
Modular neural network
WHAT ARE THE MAIN
COMPONENTS OF THE ARTIFICIAL
NEURAL NETWORKS ??
Since an artificial network is designed after
the human brain,it has some main
components which are neurons and bias.
A single artificial network is made of
hundreds or thousands of neurons
completed.
It is quite rare to find single networks since
most developers choose multi-layer networks
or stacked neural networks.
Nodes build the network layers as well as the
place of computation activities.
WHAT IS A MULTILAYER
NEURAL NETWORK??
The multilayer neural network contains some
artificial network layers or nodes.
Each layer has a different design for a
different purpose.In early invention,it is
technologies employ multilayer networks.
For example:- A convolutional neural
network uses
dozens of layers to recognize an image.
WHAT IS A BACK PROPAGATION
ALGORITHM IN A NEURAL
NETWORK??
Continuing the neural
network,backpropagation refers to neural
network training.
It is the kind of method to tune the neural
network according to the error rate from the
previous epoch.
Such a method allows you to minimize the
error rate and improve the model reliability
by increasing the generalization.
In genral,this is method to train the artificial
networks to reduce the loss function of
weights.
WHAT ARE NEURAL NETWORKS
BEST FOR ??
Neural networks have diverse
applications,from handwriting analysis,facial
recognition,to weather forecasting.
Lots of field uses artificial network including
healthcare,finace,automotive,aerospace,robo
tics, telecommunication,and more.
The nural network is simply changing how
the world used to work for the better.
WHAT ARE THE ADVANTAGES OF
ARTIFICIAL NEURAL
NETWORKS ??
In general,the artificial network can learn and
analyse data by themselves.
The output from the network is not limited to only
the input they get.
The technology does not affect data loss since
the input data is not stored in a database but on
the network itself.
The system can recognize fault and give output
even when neuron does not respond or there is
information missing.
The good thing about the artificial network is the
system can perform several tasks at once without
affecting the performance of the system in
general.
DISADVANTAGES OF AN
ARTIFICIAL NEURAL
NETWORK ??
The artificial network does not come without
draw backs.
There are some concerns regarding the
technology such as hardware independence,
unexplained functioning,unknown duration,
difficulty to show the problem and more.
CONCLUSION
The artificial neural network is the future of
artificial intelligence.
The technology has been implemented in
various sectors to solve complex problems.
The neural network applications are getting
wider from simple face recognition to
observing social media users’ behaviour and
finding fault in some applications.
Written,
designed and
directed by:-
9’e’ aditya
Pandey 2234
