SlideShare a Scribd company logo
1 of 53
Introduction: Artificial Neural Network
Adri Jovin J J, M.Tech., Ph.D.
UITE221- SOFT COMPUTING
Soft Computing
โ€ข Introduced by Lotfi A. Zadeh, University of California, Berkley
โ€ข Collection of computational methods
โ€ข Includes Fuzzy Systems, Neural Networks and Evolutionary Algorithms
โ€ข Deployment of soft computing for the solution of machine learning problems has led to high Machine Intelligence
Quotient
UITE221 SOFT COMPUTING 2
Image Credit: Electrical Engineering and Computer Sciences, UC, Berkeley
โ€œSoft computing differs from hard computing (conventional computing) in its tolerance to
imprecision, uncertainty and partial truthโ€
-Lotfi A. Zadeh
Soft Computing (Contdโ€ฆ)
Fuzzy Systems
Neural
Networks
Evolutionary
Algorithms
UITE221 SOFT COMPUTING 3
Fuzzy-evolutionary hybrids Neuro-fuzzy hybrids
Neuro-evolutionary hybrids
Neuro-fuzzy-evolutionary hybrids
Neural Networks
โ€ข Simplified models of the biological nervous system
โ€ข Processing elements called neurons โ€“ inspired by the brain
โ€ข Parallel distributed processing
โ€ข Characteristics:
โ€“ mapping capabilities or pattern association
โ€“ robustness
โ€“ fault tolerance
โ€“ parallel and high speed information processing
โ€“ nonlinearity
โ€“ adaptivity
UITE221 SOFT COMPUTING 4
This Photo by Unknown Author is licensed under CC BY-SA
Sensory
inputs
Dendrite
Axon
Soma
Synapse
Nucleus
Terminology Relationship
Biological Neuron Artificial Neuron
Cell Neuron
Dendrites Weights or Interconnections
Soma Net input
Axon Output
UITE221 SOFT COMPUTING 5
Simple Model of Artificial Neuron
UITE221 SOFT COMPUTING 6
ฮฃ ฦ’
w1
w2
wn
.
.
.
xn
x2
x1
Inputs
Weights
Summation
unit
Summation of
weighted inputs
Thresholding
unit
Thresholding
output
Output
Simple Model of Artificial Neuron
Let ๐ผ be the total input received by the soma of the artificial neuron
๐ผ = ๐‘ค1๐‘ฅ1 + ๐‘ค2๐‘ฅ2+. . . +๐‘ค๐‘›๐‘ฅ๐‘›
๐ผ =
๐‘–=1
๐‘›
๐‘ค๐‘–๐‘ฅ๐‘–
To generate the output ๐‘ฆ, the sum ๐ผ is passed on to a non-linear filter ๐œ™ called the Activation function or Transfer function
or Squash Function
๐‘ฆ = ๐œ™ ๐ผ
UITE221 SOFT COMPUTING 7
Activation Functions: Heaviside function
Very commonly used activation function: Thresholding function
The sum is compared with a threshold value ๐œƒ. If ๐ผ > ๐œƒ, then the output is 1 else it is 0
๐‘ฆ = ๐œ™
๐‘–=1
๐‘›
๐‘ค๐‘–๐‘ฅ๐‘– โˆ’ ๐œƒ
where, ๐œ™ is the step function known as Heaviside function and is such that
๐œ™ ๐ผ =
1, ๐ผ > 0
0, ๐ผ โ‰ค 0
UITE221 SOFT COMPUTING 8
Input
I
1
0 ๐œƒ
Threshold
๐œ™(๐ผ)
Output
Activation Functions: Signum function
Also known as Quantizer function
๐œ™ ๐ผ =
+1, ๐ผ > 0
โˆ’1, ๐ผ โ‰ค 0
UITE221 SOFT COMPUTING 9
Input
I
+1
0
Threshold
-1 ๐œƒ
๐œ™(๐ผ)
Output
Activation Functions: Sigmoidal function
UITE221 SOFT COMPUTING 10
This Photo by Unknown Author is licensed under CC BY-SA
Varies gradually between the asymptotic values 0 and 1 or -1 and +1
๐œ™ ๐ผ =
1
1 + ๐‘’โˆ’๐›ผ๐ผ
where, ๐›ผ is the slope parameter
The function is differentiable
Prone to vanishing gradient problem
When gradient reaches 0, the network do not learn
Activation Functions: Hyperbolic tangent function
Also known as tanh function
๐œ™ ๐ผ = tanh ๐ผ
Scaled version of sigmoid function
Leads to vanishing gradient problem in very deep neural networks
UITE221 SOFT COMPUTING 11
This Photo by Unknown Author is licensed under CC BY-SA
Other popular activation functions: ReLU and Softmax
UITE221 SOFT COMPUTING 12
This Photo by Unknown Author is licensed under CC BY-SA
โ€ข Most widely used
โ€ข Does not activate all neurons at the same time
โ€ข If input is negative the neuron will not get activated
โ€ข Overcomes the vanishing gradient problem
โ€ข Suited for hidden layers
Softmax Function
Softmax is a type of sigmoid function
Used in handling
Ideally used in output layer of the classification
๐ผ๐‘› =
๐‘’๐‘ง๐‘›
๐‘˜=1
๐‘š
๐‘’๐‘ง๐‘˜
Broader classification
UITE221 SOFT COMPUTING 13
Neural Networks
Single layer
feedforward networks
Multilayer
feedforward networks
Recurrent
networks
Neural Network Architectures
UITE221 SOFT COMPUTING 14
This Photo by Unknown Author is licensed under CC BY-SA This Photo by Unknown Author is licensed under CC BY-SA
This Photo by Unknown Author is licensed under CC BY-SA
Single layer feedforward Network Multilayer feedforward Network
Recurrent Networks
Characteristics of Neural Networks
โ€ข Exhibit mapping capabilities
โ€ข Learn by examples
โ€ข Possess the capability to generalize
โ€ข Robust and Fault-tolerant
โ€ข Can process information in parallel, at a high speed and in a distributed manner.
UITE221 SOFT COMPUTING 15
Learning Methods
UITE221 SOFT COMPUTING 16
Learning Methods
Supervised Unsupervised Reinforced
Gradient Descent Stochastic Hebbian Competitive
Least Mean
Square
Backpropagation
Supervised Learning
โ€ข Every input pattern that is used to train the network is associated with an output pattern, which is the target or the
desired pattern
โ€ข A teacher is assumed to be present to make comparison between computed output and expected output, to determine
error
โ€ข The error can be used to change network parameters which can be used to improve the performance of the network
UITE221 SOFT COMPUTING 17
Unsupervised learning
โ€ข Target output is not presented to the network
โ€ข System learns of its own by discovering and adapting to structural features in the input pattern
UITE221 SOFT COMPUTING 18
Reinforced Learning
โ€ข Though teacher will be present, does not present the expected answer but will indicate whether the computed output
is correct or not
โ€ข Reward is given for correct answer and penalty for a wrong answer
โ€ข Not a widely used method of learning
UITE221 SOFT COMPUTING 19
Hebbian Learning
โ€ข Proposed by Hebb (1949)
โ€ข Based on correlative weight adjustment
๐‘Š =
๐‘–=1
๐‘›
๐‘‹๐‘–๐‘Œ๐‘–
๐‘‡
where (๐‘‹๐‘–, ๐‘Œ๐‘–) is the input-output pattern pair
UITE221 SOFT COMPUTING 20
Gradient descent learning
โ€ข Based on minimization of error defined in terms of weights and activation function of the network
โ€ข Activation function deployed whould be differentiable since the weight update depends on the gradient of the error
โ€ข If ฮ”๐‘Š๐‘–๐‘— is the weight update of the link connecting the ith and jth neuron of two neighboring layers, then
ฮ”๐‘Š๐‘–๐‘— = ๐œ‚
๐œ•๐ธ
๐œ•๐‘Š๐‘–๐‘—
where ๐œ‚ is the learning rate parameter and
๐œ•๐ธ
๐œ•๐‘Š๐‘–๐‘—
is the error gradient with reference to the weight ๐‘Š๐‘–๐‘—
UITE221 SOFT COMPUTING 21
Competitive Learning
โ€ข Those neurons which respond strongly to input stimuli have their weights updated
โ€ข When an input pattern is presented, all neurons in the layer compete and the winning neuron undergoes weight
adjustment
โ€ข โ€œWinner-takes-allโ€ strategy
UITE221 SOFT COMPUTING 22
Stochastic learning
โ€ข Weights are adjusted in a probabilistic fashion
โ€ข e.g.: simulated annealing deployed by Boltzmann and Cauchy machines
UITE221 SOFT COMPUTING 23
Taxonomy of Neural Network Architectures
Adaptive Linear Neural Element (ADALINE)
Adaptive Resonance Theory (ART)
Associative Memory (AM)
Bidirectional Associative Memory (BAM)
Boltzmann Machine
Brain-State-in-a-Box (BSB)
Cascade Correlation (CCN)
Cauchy Machine
Counter Propagation Network (CPN)
Hamming Network
UITE221 SOFT COMPUTING 24
Hopfield Network
Learning Vector Quantization (LVQ)
Many ADALINE (MADALINE)
Multilayer Feedforward Network (MLFF)
Neocognitron
Perceptron
Radial Basis Function (RBF)
Recurrent Neural Network (RNN)
Self-organizing Feature Map (SOFM)
Evolution of Neural Networks
Year Neural Network Designer
1943 McCulloch and Pitts Neuron McCulloch and Pitts
1949 Hebb Network Hebb
1958, 1959, 1962, 1988 Perceptron Frank Rosenblatt, Block, Minsky and Papert
1960 ADALINE Widrow and Hoff
1972 Kohonen self organizing feature
map
Kohonen
1982, 1984, 1985,
1986, 1987
Hopfield Network John Hopfield and Tank
1986 Back Propagation Network Rumelhart, Hinton and Williams
1988 Counter-Propagation network Grossberg
1987-1990 Adaptive Resonance Theory (ART) Carpenter and Grossberg
1988 Radial basis function network Broomhead and Lowe
1988 Neo cognitron Fukushima
UITE221 SOFT COMPUTING 25
Basic Models
Models of ANN are specified by three basic entities namely
1. Synaptic interconnections of the models
2. Training or learning rules adopted for updating and adjusting the connection weights
3. Activation functions
UITE221 SOFT COMPUTING 26
Connections
Five basic types of neuron connection architectures
1. Single-layer feed-forward network
2. Multilayer feed-forward network
3. Single node with its own feedback
4. Single-layer recurrent network
5. Multilayer recurrent network
UITE221 SOFT COMPUTING 27
Basic Models: Connection: Single-layer feed-forward network
UITE221 SOFT COMPUTING 28
x1
x2
xn
y1
y2
ym
w11
w21
wn1
w12
w22
wn2
w1m
w2m
wnm
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Output
Neurons
Input
Neurons
Output
layer
Input
layer
Basic Models: Connection: Multilayer feed-forward network
UITE221 SOFT COMPUTING 29
x1
x2
xn
|
|
|
|
|
|
|
z1
z2
zk
|
|
|
|
|
|
|
R1
R2
Rq
|
|
|
|
|
|
|
y1
y2
ym
|
|
|
|
|
|
|
Output
Neurons
Input
Neurons
Input
layer
Output
layer
Basic Models: Connection: Single node with its own feedback
UITE221 SOFT COMPUTING 30
A1
Ai
Am
Ak
-๐œ€
-๐œ€
-๐œ€
-๐œ€
-๐œ€
-๐œ€
Input
Output
Feedback
Competitive Nets
Basic Models: Connection: Single-layer recurrent network
UITE221 SOFT COMPUTING 31
x1
x2
xn
|
|
|
|
|
|
|
Input
layer
y1
y2
ym
|
|
|
|
|
|
|
Output
layer
w11
w22
wnm
Basic Models: Connection: Multilayer recurrent network
UITE221 SOFT COMPUTING 32
x1
x2
xn
|
|
|
|
|
|
|
z1
z2
zk
|
|
|
|
|
|
|
y1
y2
ym
|
|
|
|
|
|
|
Input
layer
Output
layer
w11
w22
wn2
v11
v22
vkm
Basic Models: Connection: Lateral inhibition structure
UITE221 SOFT COMPUTING 33
Basic Models: Learning
Two kinds of learning
1. Parameter Learning: updates the connecting weights in neural network
2. Structure learning: focuses on the change in network structure (no. of processing elements, connection types)
Three categories of learning
1. Supervised Learning
2. Unsupervised Learning
3. Reinforcement Learning
UITE221 SOFT COMPUTING 34
Basic Models: Learning: Supervised Learning
UITE221 SOFT COMPUTING 35
Neural Network
W
Error Signal
Generator
Y
(Actual Output)
D
(Desired Output)
X
(Input)
Error
(D-Y)
signals
Basic Models: Learning: Unsupervised Learning
UITE221 SOFT COMPUTING 36
Artificial Neural
Network
W
Y
(Actual Output)
X
(Input)
Basic Models: Learning: Reinforcement Learning
UITE221 SOFT COMPUTING 37
Neural Network
W
Error Signal
Generator
Y
(Actual Output)
R
(Reinforcement
Signal)
X
(Input)
Error
signals
Basic Models: Activation Functions
1. Identity function
2. Binary step function
3. Bipolar step function
4. Sigmoidal function
i. Binary sigmoid function
ii. Bipolar sigmoid function
5. Ramp function
UITE221 SOFT COMPUTING 38
Basic Models: Activation Functions: Identity function
๐‘“ ๐‘ฅ = ๐‘ฅ for all ๐‘ฅ
UITE221 SOFT COMPUTING 39
This Photo by Unknown Author is licensed under CC BY
Basic Models: Activation Functions: Binary step function
๐‘“ ๐‘ฅ =
1 ๐‘–๐‘“ ๐‘ฅ โ‰ฅ ๐œƒ
0 ๐‘–๐‘“ ๐‘ฅ < ๐œƒ
๐œƒ represents the threshold values
โ€ข Widely used in single-layer nets to convert the net input to an output
that is a binary (1 or 0)
โ€ข Also known as Heaviside function (Refer slide 8)
UITE221 SOFT COMPUTING 40
Input
I
1
0 ๐œƒ
Threshold
๐œ™(๐ผ)
Output
Basic Models: Activation Functions: Bipolar step function
๐‘“ ๐‘ฅ =
1 ๐‘–๐‘“ ๐‘ฅ โ‰ฅ ๐œƒ
โˆ’1 ๐‘–๐‘“ ๐‘ฅ < ๐œƒ
๐œƒ represents the threshold values
โ€ข Used in single-layer nets to convert the net input to an output that is
a bipolar (+1 or -1)
โ€ข Also known as signum function (Refer slide 9)
UITE221 SOFT COMPUTING 41
Input
I
+1
0
Threshold
-1 ๐œƒ
๐œ™(๐ผ)
Output
Basic Models: Activation Functions: Sigmoidal function
Binary sigmoid function
โ€ข Also known as logistic sigmoid function or unipolar sigmoid function
โ€ข Range is from 0 to 1
๐‘“ ๐‘ฅ =
1
1 + ๐‘’โˆ’๐œ†๐‘ฅ
where ๐œ† is the steepness parameter
Derivative of this function is
๐‘“โ€ฒ
๐‘ฅ = ๐œ†๐‘“ ๐‘ฅ [1 โˆ’ ๐‘“(๐‘ฅ)]
UITE221 SOFT COMPUTING 42
This Photo by Unknown Author is licensed under CC BY-SA
Basic Models: Activation Functions: Sigmoid function (Contdโ€ฆ)
Bipolar sigmoid function
๐‘“ ๐‘ฅ =
2
1 + ๐‘’โˆ’๐œ†๐‘ฅ
โˆ’ 1 =
1 โˆ’ ๐‘’โˆ’๐œ†๐‘ฅ
1 + ๐‘’โˆ’๐œ†๐‘ฅ
where ๐œ† is the steepness parameter
The derivative of this function is
๐‘“โ€ฒ
๐‘ฅ =
๐œ†
2
1 + ๐‘“ ๐‘ฅ 1 โˆ’ ๐‘“ ๐‘ฅ
This function is closely related to the hyperbolic tangent function
UITE221 SOFT COMPUTING 43
This Photo by Unknown Author is licensed under CC BY-SA
Basic Models: Activation Functions: Ramp Function
๐‘“ ๐‘ฅ =
1 ๐‘–๐‘“ ๐‘ฅ > 1
๐‘ฅ ๐‘–๐‘“ 0 โ‰ค ๐‘ฅ โ‰ค 1
0 ๐‘–๐‘“ ๐‘ฅ < 0
UITE221 SOFT COMPUTING 44
Input
1
1
0 ๐œƒ
Threshold
๐œ™(๐ผ)
Output
Important terminologies
โ€ข Weights
โ€ข Bias
โ€ข Threshold
โ€ข Learning Rate
โ€ข Momentum Factor
โ€ข Vigilance Parameter
UITE221 SOFT COMPUTING 45
Weights
โ€ข Contains information about the input signal which is used to solve a problem
โ€ข Can be represented in terms of matrix
โ€ข Also known as connection matrix
โ€ข Weights encode long-term memory [LTM] and the activation states short-term memory [STM]
โ€ข Assume โ€œnโ€ processing elements and each element has โ€œmโ€ adaptive weights, weight matrix W is defined by
๐‘Š =
๐‘ค1
๐‘‡
๐‘ค2
๐‘‡
โ‹ฎ
๐‘ค๐‘›
๐‘‡
=
๐‘ค11 ๐‘ค12 โ€ฆ
๐‘ค21 ๐‘ค22 โ€ฆ
โ‹ฎ โ‹ฎ โ€ฆ
๐‘ค๐‘›1 ๐‘ค๐‘›2 โ€ฆ
๐‘ค1๐‘š
๐‘ค2๐‘š
โ‹ฎ
๐‘ค๐‘›๐‘š
UITE221 SOFT COMPUTING 46
Bias
โ€ข Almost like another weight, say ๐‘ค0๐‘— = ๐‘๐‘—
๐‘ฆ๐‘–๐‘›๐‘— = ๐‘๐‘— +
๐‘–=1
๐‘›
๐‘ฅ๐‘–๐‘ค๐‘–๐‘—
โ€ข Consider the line equation ๐‘ฆ = ๐‘š๐‘ฅ + ๐‘, ๐‘ may be considered as a bias
โ€ข Two types: positive bias and negative bias
UITE221 SOFT COMPUTING 47
Threshold
โ€ข A value upon which the final output of a network may be calculates
โ€ข Used in activation function
โ€ข Based on the threshold value, the activation functions are defined and the output is calculated
๐‘“ ๐‘›๐‘’๐‘ก =
1 ๐‘–๐‘“ ๐‘›๐‘’๐‘ก โ‰ฅ ๐œƒ
โˆ’1 ๐‘–๐‘“ ๐‘›๐‘’๐‘ก < ๐œƒ
where ๐œƒ is the threshold
UITE221 SOFT COMPUTING 48
Learning Rate
โ€ข Denoted by ๐›ผ
โ€ข Used to control the amount of weight adjustment at each step of training
โ€ข Ranges from 0 to 1 and determines the rate of learning at each time step
UITE221 SOFT COMPUTING 49
Momentum Factor
โ€ข Faster convergence if momentum factor is added to the weight update process
โ€ข Generally done in back propagation network
โ€ข Weights from one or more previous training patterns must be saved to use momentum
UITE221 SOFT COMPUTING 50
Vigilance Parameter
โ€ข Denoted by ๐œŒ
โ€ข Generally used in adaptive resonance theory (ART) network
โ€ข Used to control the degree of similarity required for patterns to be assigned to same cluster
โ€ข Ranges approximately from 0.7 to 1 to perform useful work in controlling the number of clusters
UITE221 SOFT COMPUTING 51
Applications
โ€ข Pattern recognition / Image processing
โ€ข Optimization / constraint satisfaction
โ€ข Forecasting and risk assessment
โ€ข Control Systems
UITE221 SOFT COMPUTING 52
References
Rajasekaran, S., & Pai, G. V. (2017). Neural Networks, Fuzzy Systems and Evolutionary Algorithms: Synthesis and
Applications. PHI Learning Pvt. Ltd..
Haykin, S. (2010). Neural Networks and Learning Machines, 3/E. Pearson Education India.
Sivanandam, S. N., & Deepa, S. N. (2007). Principles of soft computing. John Wiley & Sons.
UITE221 SOFT COMPUTING 53

More Related Content

What's hot

Introduction to CNN
Introduction to CNNIntroduction to CNN
Introduction to CNNShuai Zhang
ย 
Neural Networks
Neural NetworksNeural Networks
Neural NetworksAdri Jovin
ย 
Perceptron & Neural Networks
Perceptron & Neural NetworksPerceptron & Neural Networks
Perceptron & Neural NetworksNAGUR SHAREEF SHAIK
ย 
5 Introduction to neural networks
5 Introduction to neural networks5 Introduction to neural networks
5 Introduction to neural networksDmytro Fishman
ย 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural NetworkPrakash K
ย 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural NetworkAtul Krishna
ย 
Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...
Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...
Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...Simplilearn
ย 
Training Neural Networks
Training Neural NetworksTraining Neural Networks
Training Neural NetworksDatabricks
ย 
1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...
1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...
1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...PyData
ย 
MNIST and machine learning - presentation
MNIST and machine learning - presentationMNIST and machine learning - presentation
MNIST and machine learning - presentationSteve Dias da Cruz
ย 
Artificial Intelligence with Python | Edureka
Artificial Intelligence with Python | EdurekaArtificial Intelligence with Python | Edureka
Artificial Intelligence with Python | EdurekaEdureka!
ย 
Perceptron
PerceptronPerceptron
PerceptronNagarajan
ย 
Deep Feed Forward Neural Networks and Regularization
Deep Feed Forward Neural Networks and RegularizationDeep Feed Forward Neural Networks and Regularization
Deep Feed Forward Neural Networks and RegularizationYan Xu
ย 
Perceptron algorithm
Perceptron algorithmPerceptron algorithm
Perceptron algorithmZul Kawsar
ย 
Introduction Of Artificial neural network
Introduction Of Artificial neural networkIntroduction Of Artificial neural network
Introduction Of Artificial neural networkNagarajan
ย 
Neural networks.ppt
Neural networks.pptNeural networks.ppt
Neural networks.pptSrinivashR3
ย 
Neural network & its applications
Neural network & its applications Neural network & its applications
Neural network & its applications Ahmed_hashmi
ย 
Convolutional neural network
Convolutional neural networkConvolutional neural network
Convolutional neural networkMojammilHusain
ย 
Autoencoders
AutoencodersAutoencoders
AutoencodersCloudxLab
ย 

What's hot (20)

Introduction to CNN
Introduction to CNNIntroduction to CNN
Introduction to CNN
ย 
Neural Networks
Neural NetworksNeural Networks
Neural Networks
ย 
Perceptron & Neural Networks
Perceptron & Neural NetworksPerceptron & Neural Networks
Perceptron & Neural Networks
ย 
5 Introduction to neural networks
5 Introduction to neural networks5 Introduction to neural networks
5 Introduction to neural networks
ย 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural Network
ย 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural Network
ย 
Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...
Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...
Artificial Neural Network | Deep Neural Network Explained | Artificial Neural...
ย 
Training Neural Networks
Training Neural NetworksTraining Neural Networks
Training Neural Networks
ย 
1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...
1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...
1D Convolutional Neural Networks for Time Series Modeling - Nathan Janos, Jef...
ย 
MNIST and machine learning - presentation
MNIST and machine learning - presentationMNIST and machine learning - presentation
MNIST and machine learning - presentation
ย 
Artificial Intelligence with Python | Edureka
Artificial Intelligence with Python | EdurekaArtificial Intelligence with Python | Edureka
Artificial Intelligence with Python | Edureka
ย 
Perceptron
PerceptronPerceptron
Perceptron
ย 
Deep Feed Forward Neural Networks and Regularization
Deep Feed Forward Neural Networks and RegularizationDeep Feed Forward Neural Networks and Regularization
Deep Feed Forward Neural Networks and Regularization
ย 
Perceptron algorithm
Perceptron algorithmPerceptron algorithm
Perceptron algorithm
ย 
Introduction Of Artificial neural network
Introduction Of Artificial neural networkIntroduction Of Artificial neural network
Introduction Of Artificial neural network
ย 
Neural networks.ppt
Neural networks.pptNeural networks.ppt
Neural networks.ppt
ย 
Neural network & its applications
Neural network & its applications Neural network & its applications
Neural network & its applications
ย 
Neural Networks
Neural NetworksNeural Networks
Neural Networks
ย 
Convolutional neural network
Convolutional neural networkConvolutional neural network
Convolutional neural network
ย 
Autoencoders
AutoencodersAutoencoders
Autoencoders
ย 

Similar to Introduction to Artificial Neural Networks

artificial-neural-networks-rev.ppt
artificial-neural-networks-rev.pptartificial-neural-networks-rev.ppt
artificial-neural-networks-rev.pptRINUSATHYAN
ย 
artificial-neural-networks-rev.ppt
artificial-neural-networks-rev.pptartificial-neural-networks-rev.ppt
artificial-neural-networks-rev.pptSanaMateen7
ย 
Web Spam Classification Using Supervised Artificial Neural Network Algorithms
Web Spam Classification Using Supervised Artificial Neural Network AlgorithmsWeb Spam Classification Using Supervised Artificial Neural Network Algorithms
Web Spam Classification Using Supervised Artificial Neural Network Algorithmsaciijournal
ย 
Web spam classification using supervised artificial neural network algorithms
Web spam classification using supervised artificial neural network algorithmsWeb spam classification using supervised artificial neural network algorithms
Web spam classification using supervised artificial neural network algorithmsaciijournal
ย 
UNIT 5-ANN.ppt
UNIT 5-ANN.pptUNIT 5-ANN.ppt
UNIT 5-ANN.pptSivam Chinna
ย 
Artificial neural network
Artificial neural networkArtificial neural network
Artificial neural networkIshaneeSharma
ย 
Artificial Neural Networks for NIU session 2016 17
Artificial Neural Networks for NIU session 2016 17 Artificial Neural Networks for NIU session 2016 17
Artificial Neural Networks for NIU session 2016 17 Prof. Neeta Awasthy
ย 
Autoencoders for image_classification
Autoencoders for image_classificationAutoencoders for image_classification
Autoencoders for image_classificationCenk BircanoฤŸlu
ย 
08 neural networks
08 neural networks08 neural networks
08 neural networksankit_ppt
ย 
Artificial neural network by arpit_sharma
Artificial neural network by arpit_sharmaArtificial neural network by arpit_sharma
Artificial neural network by arpit_sharmaEr. Arpit Sharma
ย 
2.5 backpropagation
2.5 backpropagation2.5 backpropagation
2.5 backpropagationKrish_ver2
ย 
Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)Gaurav Mittal
ย 
Artificial Neural Network Seminar Report
Artificial Neural Network Seminar ReportArtificial Neural Network Seminar Report
Artificial Neural Network Seminar ReportTodd Turner
ย 
Deep learning
Deep learningDeep learning
Deep learningKuppusamy P
ย 
Feedforward neural network
Feedforward neural networkFeedforward neural network
Feedforward neural networkSopheaktra YONG
ย 

Similar to Introduction to Artificial Neural Networks (20)

Unit iii update
Unit iii updateUnit iii update
Unit iii update
ย 
artificial-neural-networks-rev.ppt
artificial-neural-networks-rev.pptartificial-neural-networks-rev.ppt
artificial-neural-networks-rev.ppt
ย 
artificial-neural-networks-rev.ppt
artificial-neural-networks-rev.pptartificial-neural-networks-rev.ppt
artificial-neural-networks-rev.ppt
ย 
Neural
NeuralNeural
Neural
ย 
Web Spam Classification Using Supervised Artificial Neural Network Algorithms
Web Spam Classification Using Supervised Artificial Neural Network AlgorithmsWeb Spam Classification Using Supervised Artificial Neural Network Algorithms
Web Spam Classification Using Supervised Artificial Neural Network Algorithms
ย 
Web spam classification using supervised artificial neural network algorithms
Web spam classification using supervised artificial neural network algorithmsWeb spam classification using supervised artificial neural network algorithms
Web spam classification using supervised artificial neural network algorithms
ย 
UNIT 5-ANN.ppt
UNIT 5-ANN.pptUNIT 5-ANN.ppt
UNIT 5-ANN.ppt
ย 
Artificial neural network
Artificial neural networkArtificial neural network
Artificial neural network
ย 
Unit 1
Unit 1Unit 1
Unit 1
ย 
Artificial Neural Networks for NIU session 2016 17
Artificial Neural Networks for NIU session 2016 17 Artificial Neural Networks for NIU session 2016 17
Artificial Neural Networks for NIU session 2016 17
ย 
Autoencoders for image_classification
Autoencoders for image_classificationAutoencoders for image_classification
Autoencoders for image_classification
ย 
MNN
MNNMNN
MNN
ย 
08 neural networks
08 neural networks08 neural networks
08 neural networks
ย 
Artificial neural network by arpit_sharma
Artificial neural network by arpit_sharmaArtificial neural network by arpit_sharma
Artificial neural network by arpit_sharma
ย 
2.5 backpropagation
2.5 backpropagation2.5 backpropagation
2.5 backpropagation
ย 
Multi Layer Network
Multi Layer NetworkMulti Layer Network
Multi Layer Network
ย 
Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)
ย 
Artificial Neural Network Seminar Report
Artificial Neural Network Seminar ReportArtificial Neural Network Seminar Report
Artificial Neural Network Seminar Report
ย 
Deep learning
Deep learningDeep learning
Deep learning
ย 
Feedforward neural network
Feedforward neural networkFeedforward neural network
Feedforward neural network
ย 

More from Adri Jovin

Adri Jovin J J - CV
Adri Jovin J J - CVAdri Jovin J J - CV
Adri Jovin J J - CVAdri Jovin
ย 
Introduction to Relational Database Management Systems
Introduction to Relational Database Management SystemsIntroduction to Relational Database Management Systems
Introduction to Relational Database Management SystemsAdri Jovin
ย 
Introduction to ER Diagrams
Introduction to ER DiagramsIntroduction to ER Diagrams
Introduction to ER DiagramsAdri Jovin
ย 
Introduction to Database Management Systems
Introduction to Database Management SystemsIntroduction to Database Management Systems
Introduction to Database Management SystemsAdri Jovin
ย 
Introduction to Genetic Algorithm
Introduction to Genetic AlgorithmIntroduction to Genetic Algorithm
Introduction to Genetic AlgorithmAdri Jovin
ย 
Introduction to Fuzzy logic
Introduction to Fuzzy logicIntroduction to Fuzzy logic
Introduction to Fuzzy logicAdri Jovin
ย 
Introductory Session on Soft Computing
Introductory Session on Soft ComputingIntroductory Session on Soft Computing
Introductory Session on Soft ComputingAdri Jovin
ย 
Creative Commons
Creative CommonsCreative Commons
Creative CommonsAdri Jovin
ย 
Image based security
Image based securityImage based security
Image based securityAdri Jovin
ย 
Blockchain Technologies
Blockchain TechnologiesBlockchain Technologies
Blockchain TechnologiesAdri Jovin
ย 
Introduction to Cybersecurity
Introduction to CybersecurityIntroduction to Cybersecurity
Introduction to CybersecurityAdri Jovin
ย 
Advanced Encryption System & Block Cipher Modes of Operations
Advanced Encryption System & Block Cipher Modes of OperationsAdvanced Encryption System & Block Cipher Modes of Operations
Advanced Encryption System & Block Cipher Modes of OperationsAdri Jovin
ย 
Heartbleed Bug: A case study
Heartbleed Bug: A case studyHeartbleed Bug: A case study
Heartbleed Bug: A case studyAdri Jovin
ย 
Zoom: Privacy and Security - A case study
Zoom: Privacy and Security - A case studyZoom: Privacy and Security - A case study
Zoom: Privacy and Security - A case studyAdri Jovin
ย 
Elliptic Curve Cryptography
Elliptic Curve CryptographyElliptic Curve Cryptography
Elliptic Curve CryptographyAdri Jovin
ย 
El Gamal Cryptosystem
El Gamal CryptosystemEl Gamal Cryptosystem
El Gamal CryptosystemAdri Jovin
ย 
Data Encryption Standard
Data Encryption StandardData Encryption Standard
Data Encryption StandardAdri Jovin
ย 
Classical cryptographic techniques, Feistel cipher structure
Classical cryptographic techniques, Feistel cipher structureClassical cryptographic techniques, Feistel cipher structure
Classical cryptographic techniques, Feistel cipher structureAdri Jovin
ย 
Mathematical Foundations of Cryptography
Mathematical Foundations of CryptographyMathematical Foundations of Cryptography
Mathematical Foundations of CryptographyAdri Jovin
ย 
Security Models
Security ModelsSecurity Models
Security ModelsAdri Jovin
ย 

More from Adri Jovin (20)

Adri Jovin J J - CV
Adri Jovin J J - CVAdri Jovin J J - CV
Adri Jovin J J - CV
ย 
Introduction to Relational Database Management Systems
Introduction to Relational Database Management SystemsIntroduction to Relational Database Management Systems
Introduction to Relational Database Management Systems
ย 
Introduction to ER Diagrams
Introduction to ER DiagramsIntroduction to ER Diagrams
Introduction to ER Diagrams
ย 
Introduction to Database Management Systems
Introduction to Database Management SystemsIntroduction to Database Management Systems
Introduction to Database Management Systems
ย 
Introduction to Genetic Algorithm
Introduction to Genetic AlgorithmIntroduction to Genetic Algorithm
Introduction to Genetic Algorithm
ย 
Introduction to Fuzzy logic
Introduction to Fuzzy logicIntroduction to Fuzzy logic
Introduction to Fuzzy logic
ย 
Introductory Session on Soft Computing
Introductory Session on Soft ComputingIntroductory Session on Soft Computing
Introductory Session on Soft Computing
ย 
Creative Commons
Creative CommonsCreative Commons
Creative Commons
ย 
Image based security
Image based securityImage based security
Image based security
ย 
Blockchain Technologies
Blockchain TechnologiesBlockchain Technologies
Blockchain Technologies
ย 
Introduction to Cybersecurity
Introduction to CybersecurityIntroduction to Cybersecurity
Introduction to Cybersecurity
ย 
Advanced Encryption System & Block Cipher Modes of Operations
Advanced Encryption System & Block Cipher Modes of OperationsAdvanced Encryption System & Block Cipher Modes of Operations
Advanced Encryption System & Block Cipher Modes of Operations
ย 
Heartbleed Bug: A case study
Heartbleed Bug: A case studyHeartbleed Bug: A case study
Heartbleed Bug: A case study
ย 
Zoom: Privacy and Security - A case study
Zoom: Privacy and Security - A case studyZoom: Privacy and Security - A case study
Zoom: Privacy and Security - A case study
ย 
Elliptic Curve Cryptography
Elliptic Curve CryptographyElliptic Curve Cryptography
Elliptic Curve Cryptography
ย 
El Gamal Cryptosystem
El Gamal CryptosystemEl Gamal Cryptosystem
El Gamal Cryptosystem
ย 
Data Encryption Standard
Data Encryption StandardData Encryption Standard
Data Encryption Standard
ย 
Classical cryptographic techniques, Feistel cipher structure
Classical cryptographic techniques, Feistel cipher structureClassical cryptographic techniques, Feistel cipher structure
Classical cryptographic techniques, Feistel cipher structure
ย 
Mathematical Foundations of Cryptography
Mathematical Foundations of CryptographyMathematical Foundations of Cryptography
Mathematical Foundations of Cryptography
ย 
Security Models
Security ModelsSecurity Models
Security Models
ย 

Recently uploaded

call girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธ
call girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธcall girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธ
call girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธ9953056974 Low Rate Call Girls In Saket, Delhi NCR
ย 
โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...Marc Dusseiller Dusjagr
ย 
A Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy ReformA Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy ReformChameera Dedduwage
ย 
Class 11 Legal Studies Ch-1 Concept of State .pdf
Class 11 Legal Studies Ch-1 Concept of State .pdfClass 11 Legal Studies Ch-1 Concept of State .pdf
Class 11 Legal Studies Ch-1 Concept of State .pdfakmcokerachita
ย 
URLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website AppURLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website AppCeline George
ย 
Solving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptxSolving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptxOH TEIK BIN
ย 
18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdf18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdfssuser54595a
ย 
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdfEnzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdfSumit Tiwari
ย 
The basics of sentences session 2pptx copy.pptx
The basics of sentences session 2pptx copy.pptxThe basics of sentences session 2pptx copy.pptx
The basics of sentences session 2pptx copy.pptxheathfieldcps1
ย 
Sanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdfSanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdfsanyamsingh5019
ย 
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptxContemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptxRoyAbrique
ย 
Introduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher EducationIntroduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher Educationpboyjonauth
ย 
Employee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxEmployee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxNirmalaLoungPoorunde1
ย 
Accessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impactAccessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impactdawncurless
ย 
MENTAL STATUS EXAMINATION format.docx
MENTAL     STATUS EXAMINATION format.docxMENTAL     STATUS EXAMINATION format.docx
MENTAL STATUS EXAMINATION format.docxPoojaSen20
ย 
Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...
Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...
Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...EduSkills OECD
ย 

Recently uploaded (20)

call girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธ
call girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธcall girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธ
call girls in Kamla Market (DELHI) ๐Ÿ” >เผ’9953330565๐Ÿ” genuine Escort Service ๐Ÿ”โœ”๏ธโœ”๏ธ
ย 
โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
โ€œOh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
ย 
A Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy ReformA Critique of the Proposed National Education Policy Reform
A Critique of the Proposed National Education Policy Reform
ย 
Class 11 Legal Studies Ch-1 Concept of State .pdf
Class 11 Legal Studies Ch-1 Concept of State .pdfClass 11 Legal Studies Ch-1 Concept of State .pdf
Class 11 Legal Studies Ch-1 Concept of State .pdf
ย 
URLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website AppURLs and Routing in the Odoo 17 Website App
URLs and Routing in the Odoo 17 Website App
ย 
Solving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptxSolving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptx
ย 
18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdf18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdf
18-04-UA_REPORT_MEDIALITERAะกY_INDEX-DM_23-1-final-eng.pdf
ย 
Staff of Color (SOC) Retention Efforts DDSD
Staff of Color (SOC) Retention Efforts DDSDStaff of Color (SOC) Retention Efforts DDSD
Staff of Color (SOC) Retention Efforts DDSD
ย 
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdfEnzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
ย 
9953330565 Low Rate Call Girls In Rohini Delhi NCR
9953330565 Low Rate Call Girls In Rohini  Delhi NCR9953330565 Low Rate Call Girls In Rohini  Delhi NCR
9953330565 Low Rate Call Girls In Rohini Delhi NCR
ย 
The basics of sentences session 2pptx copy.pptx
The basics of sentences session 2pptx copy.pptxThe basics of sentences session 2pptx copy.pptx
The basics of sentences session 2pptx copy.pptx
ย 
Sanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdfSanyam Choudhary Chemistry practical.pdf
Sanyam Choudhary Chemistry practical.pdf
ย 
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptxContemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
Contemporary philippine arts from the regions_PPT_Module_12 [Autosaved] (1).pptx
ย 
Introduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher EducationIntroduction to ArtificiaI Intelligence in Higher Education
Introduction to ArtificiaI Intelligence in Higher Education
ย 
TataKelola dan KamSiber Kecerdasan Buatan v022.pdf
TataKelola dan KamSiber Kecerdasan Buatan v022.pdfTataKelola dan KamSiber Kecerdasan Buatan v022.pdf
TataKelola dan KamSiber Kecerdasan Buatan v022.pdf
ย 
Employee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxEmployee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptx
ย 
Accessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impactAccessible design: Minimum effort, maximum impact
Accessible design: Minimum effort, maximum impact
ย 
Model Call Girl in Bikash Puri Delhi reach out to us at ๐Ÿ”9953056974๐Ÿ”
Model Call Girl in Bikash Puri  Delhi reach out to us at ๐Ÿ”9953056974๐Ÿ”Model Call Girl in Bikash Puri  Delhi reach out to us at ๐Ÿ”9953056974๐Ÿ”
Model Call Girl in Bikash Puri Delhi reach out to us at ๐Ÿ”9953056974๐Ÿ”
ย 
MENTAL STATUS EXAMINATION format.docx
MENTAL     STATUS EXAMINATION format.docxMENTAL     STATUS EXAMINATION format.docx
MENTAL STATUS EXAMINATION format.docx
ย 
Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...
Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...
Presentation by Andreas Schleicher Tackling the School Absenteeism Crisis 30 ...
ย 

Introduction to Artificial Neural Networks

  • 1. Introduction: Artificial Neural Network Adri Jovin J J, M.Tech., Ph.D. UITE221- SOFT COMPUTING
  • 2. Soft Computing โ€ข Introduced by Lotfi A. Zadeh, University of California, Berkley โ€ข Collection of computational methods โ€ข Includes Fuzzy Systems, Neural Networks and Evolutionary Algorithms โ€ข Deployment of soft computing for the solution of machine learning problems has led to high Machine Intelligence Quotient UITE221 SOFT COMPUTING 2 Image Credit: Electrical Engineering and Computer Sciences, UC, Berkeley โ€œSoft computing differs from hard computing (conventional computing) in its tolerance to imprecision, uncertainty and partial truthโ€ -Lotfi A. Zadeh
  • 3. Soft Computing (Contdโ€ฆ) Fuzzy Systems Neural Networks Evolutionary Algorithms UITE221 SOFT COMPUTING 3 Fuzzy-evolutionary hybrids Neuro-fuzzy hybrids Neuro-evolutionary hybrids Neuro-fuzzy-evolutionary hybrids
  • 4. Neural Networks โ€ข Simplified models of the biological nervous system โ€ข Processing elements called neurons โ€“ inspired by the brain โ€ข Parallel distributed processing โ€ข Characteristics: โ€“ mapping capabilities or pattern association โ€“ robustness โ€“ fault tolerance โ€“ parallel and high speed information processing โ€“ nonlinearity โ€“ adaptivity UITE221 SOFT COMPUTING 4 This Photo by Unknown Author is licensed under CC BY-SA Sensory inputs Dendrite Axon Soma Synapse Nucleus
  • 5. Terminology Relationship Biological Neuron Artificial Neuron Cell Neuron Dendrites Weights or Interconnections Soma Net input Axon Output UITE221 SOFT COMPUTING 5
  • 6. Simple Model of Artificial Neuron UITE221 SOFT COMPUTING 6 ฮฃ ฦ’ w1 w2 wn . . . xn x2 x1 Inputs Weights Summation unit Summation of weighted inputs Thresholding unit Thresholding output Output
  • 7. Simple Model of Artificial Neuron Let ๐ผ be the total input received by the soma of the artificial neuron ๐ผ = ๐‘ค1๐‘ฅ1 + ๐‘ค2๐‘ฅ2+. . . +๐‘ค๐‘›๐‘ฅ๐‘› ๐ผ = ๐‘–=1 ๐‘› ๐‘ค๐‘–๐‘ฅ๐‘– To generate the output ๐‘ฆ, the sum ๐ผ is passed on to a non-linear filter ๐œ™ called the Activation function or Transfer function or Squash Function ๐‘ฆ = ๐œ™ ๐ผ UITE221 SOFT COMPUTING 7
  • 8. Activation Functions: Heaviside function Very commonly used activation function: Thresholding function The sum is compared with a threshold value ๐œƒ. If ๐ผ > ๐œƒ, then the output is 1 else it is 0 ๐‘ฆ = ๐œ™ ๐‘–=1 ๐‘› ๐‘ค๐‘–๐‘ฅ๐‘– โˆ’ ๐œƒ where, ๐œ™ is the step function known as Heaviside function and is such that ๐œ™ ๐ผ = 1, ๐ผ > 0 0, ๐ผ โ‰ค 0 UITE221 SOFT COMPUTING 8 Input I 1 0 ๐œƒ Threshold ๐œ™(๐ผ) Output
  • 9. Activation Functions: Signum function Also known as Quantizer function ๐œ™ ๐ผ = +1, ๐ผ > 0 โˆ’1, ๐ผ โ‰ค 0 UITE221 SOFT COMPUTING 9 Input I +1 0 Threshold -1 ๐œƒ ๐œ™(๐ผ) Output
  • 10. Activation Functions: Sigmoidal function UITE221 SOFT COMPUTING 10 This Photo by Unknown Author is licensed under CC BY-SA Varies gradually between the asymptotic values 0 and 1 or -1 and +1 ๐œ™ ๐ผ = 1 1 + ๐‘’โˆ’๐›ผ๐ผ where, ๐›ผ is the slope parameter The function is differentiable Prone to vanishing gradient problem When gradient reaches 0, the network do not learn
  • 11. Activation Functions: Hyperbolic tangent function Also known as tanh function ๐œ™ ๐ผ = tanh ๐ผ Scaled version of sigmoid function Leads to vanishing gradient problem in very deep neural networks UITE221 SOFT COMPUTING 11 This Photo by Unknown Author is licensed under CC BY-SA
  • 12. Other popular activation functions: ReLU and Softmax UITE221 SOFT COMPUTING 12 This Photo by Unknown Author is licensed under CC BY-SA โ€ข Most widely used โ€ข Does not activate all neurons at the same time โ€ข If input is negative the neuron will not get activated โ€ข Overcomes the vanishing gradient problem โ€ข Suited for hidden layers Softmax Function Softmax is a type of sigmoid function Used in handling Ideally used in output layer of the classification ๐ผ๐‘› = ๐‘’๐‘ง๐‘› ๐‘˜=1 ๐‘š ๐‘’๐‘ง๐‘˜
  • 13. Broader classification UITE221 SOFT COMPUTING 13 Neural Networks Single layer feedforward networks Multilayer feedforward networks Recurrent networks
  • 14. Neural Network Architectures UITE221 SOFT COMPUTING 14 This Photo by Unknown Author is licensed under CC BY-SA This Photo by Unknown Author is licensed under CC BY-SA This Photo by Unknown Author is licensed under CC BY-SA Single layer feedforward Network Multilayer feedforward Network Recurrent Networks
  • 15. Characteristics of Neural Networks โ€ข Exhibit mapping capabilities โ€ข Learn by examples โ€ข Possess the capability to generalize โ€ข Robust and Fault-tolerant โ€ข Can process information in parallel, at a high speed and in a distributed manner. UITE221 SOFT COMPUTING 15
  • 16. Learning Methods UITE221 SOFT COMPUTING 16 Learning Methods Supervised Unsupervised Reinforced Gradient Descent Stochastic Hebbian Competitive Least Mean Square Backpropagation
  • 17. Supervised Learning โ€ข Every input pattern that is used to train the network is associated with an output pattern, which is the target or the desired pattern โ€ข A teacher is assumed to be present to make comparison between computed output and expected output, to determine error โ€ข The error can be used to change network parameters which can be used to improve the performance of the network UITE221 SOFT COMPUTING 17
  • 18. Unsupervised learning โ€ข Target output is not presented to the network โ€ข System learns of its own by discovering and adapting to structural features in the input pattern UITE221 SOFT COMPUTING 18
  • 19. Reinforced Learning โ€ข Though teacher will be present, does not present the expected answer but will indicate whether the computed output is correct or not โ€ข Reward is given for correct answer and penalty for a wrong answer โ€ข Not a widely used method of learning UITE221 SOFT COMPUTING 19
  • 20. Hebbian Learning โ€ข Proposed by Hebb (1949) โ€ข Based on correlative weight adjustment ๐‘Š = ๐‘–=1 ๐‘› ๐‘‹๐‘–๐‘Œ๐‘– ๐‘‡ where (๐‘‹๐‘–, ๐‘Œ๐‘–) is the input-output pattern pair UITE221 SOFT COMPUTING 20
  • 21. Gradient descent learning โ€ข Based on minimization of error defined in terms of weights and activation function of the network โ€ข Activation function deployed whould be differentiable since the weight update depends on the gradient of the error โ€ข If ฮ”๐‘Š๐‘–๐‘— is the weight update of the link connecting the ith and jth neuron of two neighboring layers, then ฮ”๐‘Š๐‘–๐‘— = ๐œ‚ ๐œ•๐ธ ๐œ•๐‘Š๐‘–๐‘— where ๐œ‚ is the learning rate parameter and ๐œ•๐ธ ๐œ•๐‘Š๐‘–๐‘— is the error gradient with reference to the weight ๐‘Š๐‘–๐‘— UITE221 SOFT COMPUTING 21
  • 22. Competitive Learning โ€ข Those neurons which respond strongly to input stimuli have their weights updated โ€ข When an input pattern is presented, all neurons in the layer compete and the winning neuron undergoes weight adjustment โ€ข โ€œWinner-takes-allโ€ strategy UITE221 SOFT COMPUTING 22
  • 23. Stochastic learning โ€ข Weights are adjusted in a probabilistic fashion โ€ข e.g.: simulated annealing deployed by Boltzmann and Cauchy machines UITE221 SOFT COMPUTING 23
  • 24. Taxonomy of Neural Network Architectures Adaptive Linear Neural Element (ADALINE) Adaptive Resonance Theory (ART) Associative Memory (AM) Bidirectional Associative Memory (BAM) Boltzmann Machine Brain-State-in-a-Box (BSB) Cascade Correlation (CCN) Cauchy Machine Counter Propagation Network (CPN) Hamming Network UITE221 SOFT COMPUTING 24 Hopfield Network Learning Vector Quantization (LVQ) Many ADALINE (MADALINE) Multilayer Feedforward Network (MLFF) Neocognitron Perceptron Radial Basis Function (RBF) Recurrent Neural Network (RNN) Self-organizing Feature Map (SOFM)
  • 25. Evolution of Neural Networks Year Neural Network Designer 1943 McCulloch and Pitts Neuron McCulloch and Pitts 1949 Hebb Network Hebb 1958, 1959, 1962, 1988 Perceptron Frank Rosenblatt, Block, Minsky and Papert 1960 ADALINE Widrow and Hoff 1972 Kohonen self organizing feature map Kohonen 1982, 1984, 1985, 1986, 1987 Hopfield Network John Hopfield and Tank 1986 Back Propagation Network Rumelhart, Hinton and Williams 1988 Counter-Propagation network Grossberg 1987-1990 Adaptive Resonance Theory (ART) Carpenter and Grossberg 1988 Radial basis function network Broomhead and Lowe 1988 Neo cognitron Fukushima UITE221 SOFT COMPUTING 25
  • 26. Basic Models Models of ANN are specified by three basic entities namely 1. Synaptic interconnections of the models 2. Training or learning rules adopted for updating and adjusting the connection weights 3. Activation functions UITE221 SOFT COMPUTING 26
  • 27. Connections Five basic types of neuron connection architectures 1. Single-layer feed-forward network 2. Multilayer feed-forward network 3. Single node with its own feedback 4. Single-layer recurrent network 5. Multilayer recurrent network UITE221 SOFT COMPUTING 27
  • 28. Basic Models: Connection: Single-layer feed-forward network UITE221 SOFT COMPUTING 28 x1 x2 xn y1 y2 ym w11 w21 wn1 w12 w22 wn2 w1m w2m wnm | | | | | | | | | | | | | | Output Neurons Input Neurons Output layer Input layer
  • 29. Basic Models: Connection: Multilayer feed-forward network UITE221 SOFT COMPUTING 29 x1 x2 xn | | | | | | | z1 z2 zk | | | | | | | R1 R2 Rq | | | | | | | y1 y2 ym | | | | | | | Output Neurons Input Neurons Input layer Output layer
  • 30. Basic Models: Connection: Single node with its own feedback UITE221 SOFT COMPUTING 30 A1 Ai Am Ak -๐œ€ -๐œ€ -๐œ€ -๐œ€ -๐œ€ -๐œ€ Input Output Feedback Competitive Nets
  • 31. Basic Models: Connection: Single-layer recurrent network UITE221 SOFT COMPUTING 31 x1 x2 xn | | | | | | | Input layer y1 y2 ym | | | | | | | Output layer w11 w22 wnm
  • 32. Basic Models: Connection: Multilayer recurrent network UITE221 SOFT COMPUTING 32 x1 x2 xn | | | | | | | z1 z2 zk | | | | | | | y1 y2 ym | | | | | | | Input layer Output layer w11 w22 wn2 v11 v22 vkm
  • 33. Basic Models: Connection: Lateral inhibition structure UITE221 SOFT COMPUTING 33
  • 34. Basic Models: Learning Two kinds of learning 1. Parameter Learning: updates the connecting weights in neural network 2. Structure learning: focuses on the change in network structure (no. of processing elements, connection types) Three categories of learning 1. Supervised Learning 2. Unsupervised Learning 3. Reinforcement Learning UITE221 SOFT COMPUTING 34
  • 35. Basic Models: Learning: Supervised Learning UITE221 SOFT COMPUTING 35 Neural Network W Error Signal Generator Y (Actual Output) D (Desired Output) X (Input) Error (D-Y) signals
  • 36. Basic Models: Learning: Unsupervised Learning UITE221 SOFT COMPUTING 36 Artificial Neural Network W Y (Actual Output) X (Input)
  • 37. Basic Models: Learning: Reinforcement Learning UITE221 SOFT COMPUTING 37 Neural Network W Error Signal Generator Y (Actual Output) R (Reinforcement Signal) X (Input) Error signals
  • 38. Basic Models: Activation Functions 1. Identity function 2. Binary step function 3. Bipolar step function 4. Sigmoidal function i. Binary sigmoid function ii. Bipolar sigmoid function 5. Ramp function UITE221 SOFT COMPUTING 38
  • 39. Basic Models: Activation Functions: Identity function ๐‘“ ๐‘ฅ = ๐‘ฅ for all ๐‘ฅ UITE221 SOFT COMPUTING 39 This Photo by Unknown Author is licensed under CC BY
  • 40. Basic Models: Activation Functions: Binary step function ๐‘“ ๐‘ฅ = 1 ๐‘–๐‘“ ๐‘ฅ โ‰ฅ ๐œƒ 0 ๐‘–๐‘“ ๐‘ฅ < ๐œƒ ๐œƒ represents the threshold values โ€ข Widely used in single-layer nets to convert the net input to an output that is a binary (1 or 0) โ€ข Also known as Heaviside function (Refer slide 8) UITE221 SOFT COMPUTING 40 Input I 1 0 ๐œƒ Threshold ๐œ™(๐ผ) Output
  • 41. Basic Models: Activation Functions: Bipolar step function ๐‘“ ๐‘ฅ = 1 ๐‘–๐‘“ ๐‘ฅ โ‰ฅ ๐œƒ โˆ’1 ๐‘–๐‘“ ๐‘ฅ < ๐œƒ ๐œƒ represents the threshold values โ€ข Used in single-layer nets to convert the net input to an output that is a bipolar (+1 or -1) โ€ข Also known as signum function (Refer slide 9) UITE221 SOFT COMPUTING 41 Input I +1 0 Threshold -1 ๐œƒ ๐œ™(๐ผ) Output
  • 42. Basic Models: Activation Functions: Sigmoidal function Binary sigmoid function โ€ข Also known as logistic sigmoid function or unipolar sigmoid function โ€ข Range is from 0 to 1 ๐‘“ ๐‘ฅ = 1 1 + ๐‘’โˆ’๐œ†๐‘ฅ where ๐œ† is the steepness parameter Derivative of this function is ๐‘“โ€ฒ ๐‘ฅ = ๐œ†๐‘“ ๐‘ฅ [1 โˆ’ ๐‘“(๐‘ฅ)] UITE221 SOFT COMPUTING 42 This Photo by Unknown Author is licensed under CC BY-SA
  • 43. Basic Models: Activation Functions: Sigmoid function (Contdโ€ฆ) Bipolar sigmoid function ๐‘“ ๐‘ฅ = 2 1 + ๐‘’โˆ’๐œ†๐‘ฅ โˆ’ 1 = 1 โˆ’ ๐‘’โˆ’๐œ†๐‘ฅ 1 + ๐‘’โˆ’๐œ†๐‘ฅ where ๐œ† is the steepness parameter The derivative of this function is ๐‘“โ€ฒ ๐‘ฅ = ๐œ† 2 1 + ๐‘“ ๐‘ฅ 1 โˆ’ ๐‘“ ๐‘ฅ This function is closely related to the hyperbolic tangent function UITE221 SOFT COMPUTING 43 This Photo by Unknown Author is licensed under CC BY-SA
  • 44. Basic Models: Activation Functions: Ramp Function ๐‘“ ๐‘ฅ = 1 ๐‘–๐‘“ ๐‘ฅ > 1 ๐‘ฅ ๐‘–๐‘“ 0 โ‰ค ๐‘ฅ โ‰ค 1 0 ๐‘–๐‘“ ๐‘ฅ < 0 UITE221 SOFT COMPUTING 44 Input 1 1 0 ๐œƒ Threshold ๐œ™(๐ผ) Output
  • 45. Important terminologies โ€ข Weights โ€ข Bias โ€ข Threshold โ€ข Learning Rate โ€ข Momentum Factor โ€ข Vigilance Parameter UITE221 SOFT COMPUTING 45
  • 46. Weights โ€ข Contains information about the input signal which is used to solve a problem โ€ข Can be represented in terms of matrix โ€ข Also known as connection matrix โ€ข Weights encode long-term memory [LTM] and the activation states short-term memory [STM] โ€ข Assume โ€œnโ€ processing elements and each element has โ€œmโ€ adaptive weights, weight matrix W is defined by ๐‘Š = ๐‘ค1 ๐‘‡ ๐‘ค2 ๐‘‡ โ‹ฎ ๐‘ค๐‘› ๐‘‡ = ๐‘ค11 ๐‘ค12 โ€ฆ ๐‘ค21 ๐‘ค22 โ€ฆ โ‹ฎ โ‹ฎ โ€ฆ ๐‘ค๐‘›1 ๐‘ค๐‘›2 โ€ฆ ๐‘ค1๐‘š ๐‘ค2๐‘š โ‹ฎ ๐‘ค๐‘›๐‘š UITE221 SOFT COMPUTING 46
  • 47. Bias โ€ข Almost like another weight, say ๐‘ค0๐‘— = ๐‘๐‘— ๐‘ฆ๐‘–๐‘›๐‘— = ๐‘๐‘— + ๐‘–=1 ๐‘› ๐‘ฅ๐‘–๐‘ค๐‘–๐‘— โ€ข Consider the line equation ๐‘ฆ = ๐‘š๐‘ฅ + ๐‘, ๐‘ may be considered as a bias โ€ข Two types: positive bias and negative bias UITE221 SOFT COMPUTING 47
  • 48. Threshold โ€ข A value upon which the final output of a network may be calculates โ€ข Used in activation function โ€ข Based on the threshold value, the activation functions are defined and the output is calculated ๐‘“ ๐‘›๐‘’๐‘ก = 1 ๐‘–๐‘“ ๐‘›๐‘’๐‘ก โ‰ฅ ๐œƒ โˆ’1 ๐‘–๐‘“ ๐‘›๐‘’๐‘ก < ๐œƒ where ๐œƒ is the threshold UITE221 SOFT COMPUTING 48
  • 49. Learning Rate โ€ข Denoted by ๐›ผ โ€ข Used to control the amount of weight adjustment at each step of training โ€ข Ranges from 0 to 1 and determines the rate of learning at each time step UITE221 SOFT COMPUTING 49
  • 50. Momentum Factor โ€ข Faster convergence if momentum factor is added to the weight update process โ€ข Generally done in back propagation network โ€ข Weights from one or more previous training patterns must be saved to use momentum UITE221 SOFT COMPUTING 50
  • 51. Vigilance Parameter โ€ข Denoted by ๐œŒ โ€ข Generally used in adaptive resonance theory (ART) network โ€ข Used to control the degree of similarity required for patterns to be assigned to same cluster โ€ข Ranges approximately from 0.7 to 1 to perform useful work in controlling the number of clusters UITE221 SOFT COMPUTING 51
  • 52. Applications โ€ข Pattern recognition / Image processing โ€ข Optimization / constraint satisfaction โ€ข Forecasting and risk assessment โ€ข Control Systems UITE221 SOFT COMPUTING 52
  • 53. References Rajasekaran, S., & Pai, G. V. (2017). Neural Networks, Fuzzy Systems and Evolutionary Algorithms: Synthesis and Applications. PHI Learning Pvt. Ltd.. Haykin, S. (2010). Neural Networks and Learning Machines, 3/E. Pearson Education India. Sivanandam, S. N., & Deepa, S. N. (2007). Principles of soft computing. John Wiley & Sons. UITE221 SOFT COMPUTING 53

Editor's Notes

  1. Unsupervised learning