ARTIFICIAL NEURAL NETWORKS:

A TUTORIAL

BY:
NeginYousefpour
PhD Student
Civil Engineering Department

TEXASA&M UNIVERSITY
 Contents
 Introduction
 Origin Of Neural Network
 Biological Neural Networks
 ANN Overview
 Learning
 Different NN Networks
 Challenging Problems
 Summery
INTRODUCTION
● Artificial Neural Network (ANN) or Neural Network(NN)
has provide an exciting alternative method for solving a
variety of problems in different fields of science and
engineering.
● This article is trying to give the readers a :
- Whole idea aboutANN
- Motivation for ANN development
- Network architecture and learning models
- Outline some of the important use of ANN
 Origin of Neural Network
● Human brain has many incredible characteristics such as massive
parallelism, distributed representation and computation, learning
ability, generalization ability, adaptivity, which seems simple but is
really complicated.
● It has been always a dream for computer scientist to create a
computer which could solve complex perceptual problems this fast.
● ANN models was an effort to apply the same method as human brain
uses to solve perceptual problems.
● Three periods of development for ANN:
- 1940:Mcculloch and Pitts: Initial works
- 1960: Rosenblatt: perceptron convergence theorem
Minsky and Papert: work showing the limitations of a simple
perceptron
- 1980: Hopfield/Werbos and Rumelhart: Hopfield's energy
approach/back-propagation learning algorithm
Biological Neural Network
 Biological Neural Network
● When a signal reaches a synapse: Certain
chemicals called neurotransmitters are
released.
● Process of learning:The synapse
effectiveness can be adjusted by signal
passing through.
● Cerebral cortex :a large flat sheet of
neurons about 2 to 3 mm thick and 2200
cm , 10^11 neurons
● Duration of impulses between neurons:
milliseconds and the amount of
information sent is also small(few bits)
● Critical information are not transmitted
directly , but stored ininterconnections
● The term Connectionist model initiated
from this idea.
ANN Overview: COMPTIONAL MODEL FOR ARTIFICIAL NEURON
The McCullogh-Pitts model

n
z   wixi ; y  H (z)
i1

Wires : axon &dendrites

Feed-
Recurrent
Forward
 Single layer
perceptron

Multilayer
Perceptron
Feed-
Forward Radial Basis
Connection Functions
Competetive
patterns
Networks
Recurrent
Kohonen’s
SOM

Hopefield
Network

ARTmodels
 Learning
● What is the learning process in ANN?
- updating network architecture and connection weights so that
network can efficiently perform a task
● What is the source of learning for ANN?
- Available training patterns
- The ability of ANN to automatically learn from examples or input-out
put relations
● How to design a Learningprocess?
- Knowing about available information
- Having a model from environment: Learning Paradigm
- Figuring out the update process of weights: Learning rules
- Identifying a procedure to adjust weights by learning rules: Learning
algorithm
Learning Paradigm
1. Supervised
● The correct answer is provided for the network for every input pattern
● Weights are adjusted regarding the correct answer
● In reinforcement learning only a critique of correct answer is provided
2. Unsupervised
●Does not need the correct output
●The system itself recognize the correlation and organize patterns into
categories accordingly
3. Hybrid
●Acombination of supervised and unsupervised
●Some of the weights are provided with correct output while the
others are automatically corrected.
Learning Rules
● There are four basic types of learning rules:
- Error correction rules
- Boltzmann
- Hebbian
- Competitive learning
● Each of these can be trained with or with out a teacher
● Have a particular architecture and learningalgorithm
Error Correction Rules
● An error is calculated for the output and is us ed to modify
the connection weights; error gradually reduces
● Perceptron learning rule is based on this error-correction
principle
● A perceptron consists of a single neuron with adjustable
weights and a threshold u.
● If an error occurs the weights get updated by iterating unit
ill reaching an error of zero
● Since the decision boundary is linear so ifthe patterns are
linearly separated, learning process converges w ith an infinite
No. of iteration
 Boltzman Learning
● Used in symmetric recurrent networks(symmetric:Wij=Wji)
● Consist of binary units(+1 for on, -1 for off)
● Neurons are divided into two groups: Hidden &Visible
● Outputs are produced according to Boltzman statistical
mechanics
● Boltzman learning adjust weights until visible units satisfy a
desired probabilistic distribution
● The change in connection weight or the error-correction is
measured between the correlation between two pair of input
and output neuron under clamped and free-operating
condition
Hebbian Rules
● One of the oldest learning rule initiated
form neurobiological experiments
● The basic concept of Hebbian Learning:
when neuron A activates, and then causes
neuron B to activate, then the connection
strength between the two neurons is
increased, and it will be easier for A to
activate Bin the future.
● Learning is done locally, it means weight of
a connection is corrected only with respect
to neurons connected to it.
● Orientation selectivity: occurs due to
Hebbian training of a network
Competitive Learning Rules
● The basis is the “winner take all” originated from biological Neural
network
● All input units are connected together and all units of output are
also connected via inhibitory weights but gets feed back with
excitory weight
● Only one of the unites with largest or smallest input is activated
and its weight becomes adjusted
● As a result of learning process the pattern in the winner unit
(weight) become closer to he input pattern.
● http://www.peltarion.com/blog/img/sog/competitive.gif
Multilayer Perceptron
● The most popular networks with feed-forward system
● Applies the Back Propagationalgorithm
● As a result of having hidden units, Multilayer perceptron can form
arbitrarily complex decision boundaries
● Each unit in the first hidden layer impose a hyperplane in the space
pattern
● Each unit in the second hidden layer impose a hyperregion on outputs of
the first layer
● Output layer combines the hyperregions of all units in second layer
Back propagation Algorithm
 Radial Basis FunctionNetwork(RBF)
● A Radial Basis Function like Gaussian Kernel is applied as an
activation function.
● ARadial Basis Function(also called Kernel Function) is a real-
valued function whose value depends only on the distance
from the origin or anyother center: F(x)=F(|x|)
● RBF network uses a hybrid learning , unsupervised clustering
algorithm and asupervised least square algorithm
● As a comparison to multilayer perceptronNet.:
-The Learning algorithm is faster than back-propegation
- After training the running time is much more slower
Let’s see an example!
Gaussian Function
Kohonen Self-Organizing Maps
● It consists of a two dimensional array of output units connected to all input
nodes
● It works based on the property of Topology preservation
● Nearby input patterns should activate nearby output units on the map
● SOM can be recognized as aspecial competitive learning
● Only the weight vectors of winner and its neighbor units are updated
 Adaptive Resonance Theory Model
● ART models were proposed theories to overcome the
concerns related to stability-plasticity dilemma in
competitive learning.
● Is it likely that learning could corrupt the existing knowledge
in a unite?
● If the input vector is similar enough to one of the stored
prototypes(resonance), learning updates the prototype.
● If not, a new category is defined in an “uncommitted” unit.
● Similarity is controlled by vigilance parameter.
Hopfield Network
● Hopfield designed a network based on an energy function
● As a result of dynamic recurrency, the network total energy
decreases and tends to a minimum value(attractor)
● The dynamic updates take places in two ways: synchronously and
asynchronously
● traveling salesman problem(TSP)
Challenging Problems
● Pattern recognition
- Character recognition
- Speech recognition

● Clustering/Categorization
- Data mining
- Data analysis
 Challenging Problems
● Function approximation
- Engineering &scientific Modeling

● Prediction/Forecasting
- Decision-making
- Weatherforecasting
Challenging Problems
● Optimization
- Traveling salesmanproblem(TSP)
 Summery
● Agreat overview of ANN is presented in this paper, it is very
easy understanding and straightforward
● The different types of learning rules, algorithms and also
different architectures are well explained
● Anumber of Networks were described through simple words
● The popular applications of NN were illustrated
● The author Believes that ANNS brought up both enthusiasm
and criticism.
● Actually except for some special problems there is no evidence
that NN is better working than other alternatives
● More development and better performance in NN requires
the combination of ANN with newtechnologies