Scribd
Upload
84 views

Chapter 1. Introduction To Neural Network

The document discusses artificial neural networks including biological neural networks, artificial neural networks, and applications of artificial neural networks. It describes the components and functioning of biological neural networks. It then explains single layer and multi-layer artificial neural networks and provides examples of computations. It concludes by listing several applications of artificial neural networks in fields such as pattern recognition, computer vision, and time series prediction.

Uploaded by

tranduykhiem1603
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
84 views

Chapter 1. Introduction To Neural Network

The document discusses artificial neural networks including biological neural networks, artificial neural networks, and applications of artificial neural networks. It describes the components and functioning of biological neural networks. It then explains single layer and multi-layer artificial neural networks and provides examples of computations. It concludes by listing several applications of artificial neural networks in fields such as pattern recognition, computer vision, and time series prediction.

Uploaded by

tranduykhiem1603
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 34

APPLIED

ARTIFICIAL
NEURAL
NETWORK

© Faculty of Electronic Technology IUH


Chapter 1
Introduction to neural network
1.1 Biological neural network

1.2 Artificial neural network

1.3 Applications of artificial neural networks

© Faculty of Electronic Technology IUH 2


REFERENCES

Charu C. Aggarwal, Neural Networks and Deep

Learning: A Textbook, Springer, 2018

Michael Nielsen, Neural Networks and Deep Learning,

Google Book, 2019

Ahmed Gad, Fatima Jarmouni, Introduction to Deep

Learning and Neural Networks with Python, Elsevier,

2020
© Faculty of Electronic Technology IUH 3
1.1 Biological neural network

© Faculty of Electronic Technology IUH 4


1.1 Biological neural network

⮚ The human nervous system is a complex network of

cells, which are referred to as neurons.

⮚ The human brain is a complex,

nonlinear, and parallel

information processing system,

capable of learning, memory and generalization.

⮚ The human brain is made up of about 1011 neurons,

which linked together into a network


© Faculty of Electronic Technology IUH 5
1.1 Biological neural network

⮚ A typical neuron consists of a cell body (soma), dendrites,

and a single axon.

Image source here

© Faculty of Electronic Technology IUH 6


1.1 Biological neural network
How BNN work
⮚ Each neuron receives electrical signals (impulses) from
neighboring neurons via fibers called dendrites.
⮚ Cell body (soma) has the function of synthesizing
incoming signals.
⮚ When the total sum of incoming signals exceeds some
threshold, the neuron “fires” output signal via long fibers
called axons.
⮚ The junction between two neurons is called a synapse.

© Faculty of Electronic Technology IUH 7


1.1 Biological neural network

How BNN work

© Faculty of Electronic Technology IUH 8


1.2 Artificial neural network

© Faculty of Electronic Technology IUH 9


1.2 Artificial neural network
What is ANN
⮚ Artificial neural network (ANN) simulate the human–brain
processes by mathematic.
Biological Neural
Artificial Neural Network
Network
Dendrites Inputs

Cell nucleus Nodes

Synapse Weights

Axon Output

© Faculty of Electronic Technology IUH 10


1.2 Artificial neural network
1.2.1 Single layer neural network, perceptron
⮚ Neural network computation consists of two processes:

© Faculty of Electronic Technology IUH 11


1.2 Artificial neural network
Input computation
⮚ The equation is expressed in summation form as.

or in vector form as

Where are m-dimension column vector; w0 – bias

- vector of weights - input vector


with m-features

© Faculty of Electronic Technology IUH 12


1.2 Artificial neural network
Input computation
⮚ Expanding the weight vector and input vector

The computation can be rewritten as follows

Where are m+1 - dimension column vector.


© Faculty of Electronic Technology IUH 13
1.2 Artificial neural network
Output computation
⮚ The output of the neural, denoted by y, is obtained by
passing f (or net) through the active function a.

⮚ Some active functions:


1. Step function:

2. Sign function:

© Faculty of Electronic Technology IUH 14


1.2 Artificial neural network
3. Linear function:

4. Rectifier Linear Unit (ReLU):

5. Sigmoid Function
15

© Faculty of Electronic Technology IUH


1.2 Artificial neural network
6. Hyperbolic tangent Function

⮚ The simplest single layer neural network is referred to as


the perceptron.
⮚ The perceptron contains a single input layer and an
output node.
⮚ Output of perceptron is a binary value.
16

© Faculty of Electronic Technology IUH


1.2 Artificial neural network

The perceptron model

⮚ In perceptron, the active function is sign function or step


function. 17

© Faculty of Electronic Technology IUH


1.2 Artificial neural network

Single neuron in action. Source


18
here

© Faculty of Electronic Technology IUH


1.2 Artificial neural network
Example:
⮚ Computation the output of the neural y, with active function:
a. Step
b. Linear
c. Sigmoid
d. Tanh

19

© Faculty of Electronic Technology IUH


1.2 Artificial neural network

1.2.2 Multi-layer neural network

⮚ The neurons are linked together to form a neural network.


⮚ The neural network consists of 3 layers: input layer, hidden
layer and output layer.
⮚ The neural network with a single hidden layer is called
shallow neural network, and refer to network with two or
more hidden layers as a deep neural network.

© Faculty of Electronic Technology IUH 20


1.2 Artificial neural network

1.2.2 Multi-layer neural network

© Faculty of Electronic Technology IUH 21


1.2 Artificial neural network

1.2.2 Multi-layer neural network


⮚ Feed-Forward Networks - A multi-layered neural network
where the information moves from left to right, or in other
words, in a forward direction. The input values pass through
a series of hidden layers on their way to the output layer.

© Faculty of Electronic Technology IUH 22


1.2 Artificial neural network

1.2.2 Multi-layer neural network


⮚ Recurrent neural networks, also known as RNNs, are a
class of neural networks that allow previous outputs to be
used as inputs while having hidden states.

© Faculty of Electronic Technology IUH 23


1.2 Artificial neural network
NEURAL NETWORK COMPUTATION
⮚ The outputs of the layer 1 are the components of input
vector x:

The matrix form:

⮚ The matrix weights in layer l :

© Faculty of Electronic Technology IUH 24


1.2 Artificial neural network
NEURAL NETWORK COMPUTATION
⮚ Obtain all the sum-of-products computations, neti(l), for layer l
simultaneously:

⮚ Because the activation function is applied to each net input


independently of the others, the outputs of the network at
any layer can be expressed in vector form as:

© Faculty of Electronic Technology IUH 25


1.2 Artificial neural network
Example

© Faculty of Electronic Technology IUH 26


1.3 Applications of Artificial neural
network

© Faculty of Electronic Technology IUH 27


1.3 Applications of Artificial
neural network
Artificial neural networks (ANNs) have a wide range of
applications across various fields due to their ability to learn
from data and perform complex tasks. Some notable
applications of artificial neural networks include:

1. Perceptron

⮚ Classification.
⮚ Encode Database (Multilayer Perceptron).
⮚ Monitor Access Data (Multilayer Perceptron).

© Faculty of Electronic Technology IUH 28


1.3 Applications of Artificial
neural network
2. Feed Forward (FF)

⮚ Data Compression.
⮚ Pattern Recognition.
⮚ Computer Vision.
⮚ Sonar Target Recognition.
⮚ Speech Recognition.
⮚ Handwritten Characters Recognition.

© Faculty of Electronic Technology IUH 29


1.3 Applications of Artificial
neural network
3. Radial Basis Network (RBN)
⮚ Function Approximation.
⮚ Timeseries Prediction.
⮚ Classification.
⮚ System Control.

4. Long / Short Term Memory (LSTM)

⮚ Speech Recognition.
⮚ Writing Recognition.

© Faculty of Electronic Technology IUH 30


1.3 Applications of Artificial
neural network
5. Recurrent Neural Network (RNN)
⮚ Machine Translation.
⮚ Robot Control.
⮚ Time Series Prediction.
⮚ Speech Recognition.
⮚ Speech Synthesis.
⮚ Time Series Anomaly Detection.
⮚ Rhythm Learning.
⮚ Music Composition.

© Faculty of Electronic Technology IUH 31


1.3 Applications of Artificial
neural network
6. Markov Chain (MC)
⮚ Speech Recognition.
⮚ Information And Communication System.
⮚ Queuing Theory.
⮚ Statistics.

© Faculty of Electronic Technology IUH 32


1.3 Applications of Artificial
neural network
7. Convolutional Neural Networks
⮚ Identify Faces, Street Signs, Tumors.
⮚ Image Recognition.
⮚ Video Analysis.
⮚ NLP.
⮚ Anomaly Detection.
⮚ Drug Discovery.
⮚ Checkers Game.
⮚ Time Series Forecasting.

© Faculty of Electronic Technology IUH 33


1.3 Applications of Artificial
neural network
8. Support Vector Machines (SVM)

⮚ Face Detection.
⮚ Text Categorization.
⮚ Classification.
⮚ Bioinformatics.
⮚ Handwriting recognition.

© Faculty of Electronic Technology IUH 34

You might also like