Scribd
Upload
54 views13 pages

"Artificial Neural Networks": A Presentation On

The presentation discusses Artificial Neural Networks (ANN), which are machine learning tools inspired by the biological neural networks of the brain. It covers the architectures, learning algorithms, applications, limitations, and future scope of ANNs, highlighting their increasing relevance in machine learning and the potential for integration with other models. Key applications include classification, regression analysis, and image processing, while limitations involve training challenges and the black-box nature of the networks.

Uploaded by

reaz4524
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
54 views13 pages

"Artificial Neural Networks": A Presentation On

The presentation discusses Artificial Neural Networks (ANN), which are machine learning tools inspired by the biological neural networks of the brain. It covers the architectures, learning algorithms, applications, limitations, and future scope of ANNs, highlighting their increasing relevance in machine learning and the potential for integration with other models. Key applications include classification, regression analysis, and image processing, while limitations involve training challenges and the black-box nature of the networks.

Uploaded by

reaz4524
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 13

A Presentation on

“Artificial Neural Networks”

By
ANM Reazul Islam (222-15-6164)
Md.Shaharia Jaman Bijoy (222-15-6317)
Angkur Saha( 222-15-6182)
Shatabda Singha(222-15-6417)
Sultan Mahmud(222-15-6110)
Outline

 Introduction
 Architectures
 Learning
 Applications
 Limitations
 Future Scope
Introduction

 Artificial Neural Network (ANN) is a machine learning tool.


 It is inspired by biological neural network of brain.
 Billion of neurons are connected together in the brain. They receive electrochemical
signals from neighboring neurons, they process it and forward it to the next neighboring
neurons in the network.
 Similar to the biological neurons, artificial neurons are connected together to form the
artificial neural network.
 The network process the data in patterns rather than executing sequential information.
 As information is stored in brain as strengths of synaptic gaps between neurons, similarly
the knowledge is stored in ANN as weights associated with the interconnection between
artificial neurons.
Introduction (Cont….)

Structure of an artificial neuron is shown in the above figure. The net input y in to the
neuron can be given as:
yin = b+x1w1+x2w2+…..+xnwn
Or, yin =

where b is the bias term. The output of the neuron can be given as:
yout = f(yin)
Where f(.) is called the activation or transfer function of the neuron. This function is used
as per the required output.
Architectures

There are two popular architectures of ANN:-


1. Feed-Forward Architecture
 Single Layer Feed-Forward Architecture
 Multilayer Feed-Forward Architecture

2. Feed-Backward Architecture
 Recurrent Networks
Architectures (Cont….)

Single Layer Feed-Forward Architecture


Architectures (Cont….)

Multilayer Feed-Forward Architecture


Architectures (Cont….)

Recurrent Networks
Learning

 The main goal of neural network is to solve the problems through learning which are
complex to solve by human.
 In all the neural net learning methods, weights of the network are adjusted.
 There are a variety if learning algorithms used by neural networks to solve problems.
 Some popular neural network learning algorithms are:
 Backpropagation Learning
 Hebbian learning Supervised Learning
 Perceptron Learning
 Delta/Least Mean Square learning
 Winner-Takes-All Unsupervised Learning
Applications

There are a variety of applications of neural networks in which some important can be listed
as:-
 Classification
 Clustering
 Regression Analysis
 Time Series prediction
 Predictive analytics
 Pattern Recognition
 Image Processing
 Decision Making
Limitations

However ANNs have proven their strong capabilities in problem solving, there are some limitations associated
with them:-
 Issue of Training: They require a lot of training iterations.
 Black-Box: We only concerned with the input and output.
 Over fitting and Under fitting: Size and volume of data matters.
Future Scope

 As the demand of machine learning is being increased, the scope of neural network is also
increased.
 With the introduction deep learning, deep neural networks are trend now a days.
 New features can be added to the neural network model to increase its efficiency in
problem solving.
 Integration of other models with neural networks like Hybrid Neural Networks, has
increased the power of this model. It may be continued further.
 Tuning of network parameters can also increased the correctness in the system.
Thank You

You might also like