SlideShare une entreprise Scribd logo
1  sur  6
Télécharger pour lire hors ligne
66//1010//20132013
Training in Neural Network
• Learn values of weights from I/O pairs
• Start with random weights
• Load training example’s input
• Observe computed input
• Modify weights to reduce difference
• Iterate over all training examples
• Terminate when weights stop changing
OR when error is very small
Single Layer Perceptrons
• A network with all the inputs connected directly to the
output (figure 1)
• In simple cases, feature space is divided by drawing a
hyperplane across it, which is Known as a decision
boundary
• Problems which can be thus classified are linearly
separable (figure 2)
++
++
++
++ --
--
--
--
xx11
xx22
FigureFigure 11 FigureFigure 22
66//1010//20132013
Single Layer Perceptrons
• Classical Measure of Error
– Squared error
– Where Err is the difference between the target value
and the output by the network
• Weight Modifying
– Use gradient descent to reduce the squared error by
calculating the partial derivative of E with respect to
each weight
j
jj
xinfErr
Err
Err
E
*** )('−=
∂
∂
=
∂
∂
ww
: Learning Rate: Learning Rate
Error Back-Propagation
The gradient descent
• The gradient of the error E gives the direction
where the error function at the current setting of
the w will increase. In order to decrease E, we
take a small step in the opposite direction, -G
66//1010//20132013
Error Back-Propagation
The gradient descentThe gradient descent
InIn 22D (one weight)D (one weight)
By repeating this over and over, we moveBy repeating this over and over, we move
"downhill" in"downhill" in EE until we reach a minimumuntil we reach a minimum
Error Back-Propagation
The gradient descentThe gradient descent
66//1010//20132013
Single Layer Perceptrons –
a basic application
• Suppose we have data about the height and
the age of a population of 100 people.
• So we can plot a 2D sketch (x is the age, y
the height)
• How can we predict the height of a 101th
person, given his age?
UsingUsing aa modelmodel of the data. The simplestof the data. The simplest
model can be :model can be : y = wy = w11 x + wx + w00
This may exactly be the equation of theThis may exactly be the equation of the
output of a neuron networkoutput of a neuron network
Different Non-Linearly
Separable Problems
StructureStructure
ExclusiveExclusive--OROR
ProblemProblem
Classes withClasses with
Meshed regionsMeshed regions
Most GeneralMost General
Region ShapesRegion Shapes
SingleSingle--LayerLayer
TwoTwo--LayerLayer
ThreeThree--LayerLayer
AA
AABB
BB
AA
AABB
BB
AA
AABB
BB
BB
AA
BB
AA
BB
AA
66//1010//20132013
Multilayer Perceptrons(MLP)
HiddenHidden LayerLayer
Output LayerOutput Layer
AdjustableAdjustable
wweightseights
IntputIntput LayerLayer
AdjustableAdjustable
wweightseights
InputInputUnitUnits(ExternalStimuli)s(ExternalStimuli)
OutputValuesOutputValues
Types of Layers
• Input layer (units)
– Introduces input values into the network
– No activation function or other processing
• Hidden layer(s)
– Perform classification of features
– Two hidden layers are sufficient to solve any problem
– Features imply more layers may be better
• Output layer
– Functionally just like the hidden layers
– Outputs are passed on to the world outside the neural
network
66//1010//20132013
MLP Characteristics
• Input propagates in a forward direction,
layer-by-layer basis
– also called Multilayer Feedforward Network, MLP
• Non-linear activation function
– differentiable
– nonlinearity prevent reduction to single-layer perceptron
• One or more layers of hidden neurons
– progressively extracting more meaningful features from
input patterns
• High degree of connectivity
Problems of MLP
• Nonlinearity and high degree of connectivity
makes theoretical analysis difficult
• Output vector rather than a single output value
• Error at output layer is clear, but error at the
hidden layers seems mysterious
• Learning process is hard to visualize
• So, Error Back-Propagation Algorithm (BPA) is
a computationally efficient training for MLP

Contenu connexe

Tendances

Coupling_of_IGA_plates_and_3D_FEM_domain
Coupling_of_IGA_plates_and_3D_FEM_domainCoupling_of_IGA_plates_and_3D_FEM_domain
Coupling_of_IGA_plates_and_3D_FEM_domain
Nguyen Vinh Phu
 
Oleksandr Obiednikov “Affine transforms and how CNN lives with them”
Oleksandr Obiednikov “Affine transforms and how CNN lives with them”Oleksandr Obiednikov “Affine transforms and how CNN lives with them”
Oleksandr Obiednikov “Affine transforms and how CNN lives with them”
Lviv Startup Club
 
International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)
IJERD Editor
 
Mitchell's Face Recognition
Mitchell's Face RecognitionMitchell's Face Recognition
Mitchell's Face Recognition
butest
 

Tendances (19)

Neural network
Neural networkNeural network
Neural network
 
Back propagation network
Back propagation networkBack propagation network
Back propagation network
 
Coupling_of_IGA_plates_and_3D_FEM_domain
Coupling_of_IGA_plates_and_3D_FEM_domainCoupling_of_IGA_plates_and_3D_FEM_domain
Coupling_of_IGA_plates_and_3D_FEM_domain
 
cheb_conf_aksenov.pdf
cheb_conf_aksenov.pdfcheb_conf_aksenov.pdf
cheb_conf_aksenov.pdf
 
Interconnection Network
Interconnection NetworkInterconnection Network
Interconnection Network
 
SET
SETSET
SET
 
The Birefringent Property of an Optical Resinfor the Study of a Stress Field ...
The Birefringent Property of an Optical Resinfor the Study of a Stress Field ...The Birefringent Property of an Optical Resinfor the Study of a Stress Field ...
The Birefringent Property of an Optical Resinfor the Study of a Stress Field ...
 
neural networksNnf
neural networksNnfneural networksNnf
neural networksNnf
 
Acem neuralnetworks
Acem neuralnetworksAcem neuralnetworks
Acem neuralnetworks
 
Oleksandr Obiednikov “Affine transforms and how CNN lives with them”
Oleksandr Obiednikov “Affine transforms and how CNN lives with them”Oleksandr Obiednikov “Affine transforms and how CNN lives with them”
Oleksandr Obiednikov “Affine transforms and how CNN lives with them”
 
KAIST 2012 Fall 전자공학개론 6조 발표 PPT
KAIST 2012 Fall 전자공학개론 6조 발표 PPTKAIST 2012 Fall 전자공학개론 6조 발표 PPT
KAIST 2012 Fall 전자공학개론 6조 발표 PPT
 
Finite element analysis of space truss by abaqus
Finite element analysis of space truss by abaqus Finite element analysis of space truss by abaqus
Finite element analysis of space truss by abaqus
 
International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)
 
International journal of applied sciences and innovation vol 2015 - no 1 - ...
International journal of applied sciences and innovation   vol 2015 - no 1 - ...International journal of applied sciences and innovation   vol 2015 - no 1 - ...
International journal of applied sciences and innovation vol 2015 - no 1 - ...
 
Kohonen self organizing maps
Kohonen self organizing mapsKohonen self organizing maps
Kohonen self organizing maps
 
Mc culloch pitts neuron
Mc culloch pitts neuronMc culloch pitts neuron
Mc culloch pitts neuron
 
Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...Computation of electromagnetic fields scattered from dielectric objects of un...
Computation of electromagnetic fields scattered from dielectric objects of un...
 
Computational fracture mechanics
Computational fracture mechanicsComputational fracture mechanics
Computational fracture mechanics
 
Mitchell's Face Recognition
Mitchell's Face RecognitionMitchell's Face Recognition
Mitchell's Face Recognition
 

En vedette

Foundations of business strategy SoA
Foundations of business strategy SoAFoundations of business strategy SoA
Foundations of business strategy SoA
Mart Waterval
 

En vedette (7)

Prã¡ctica 2
Prã¡ctica 2Prã¡ctica 2
Prã¡ctica 2
 
Foundations of business strategy SoA
Foundations of business strategy SoAFoundations of business strategy SoA
Foundations of business strategy SoA
 
Разработка веб-проектов в компании без проект-менеджеров — Артур Богданова
Разработка веб-проектов в компании без проект-менеджеров — Артур Богданова Разработка веб-проектов в компании без проект-менеджеров — Артур Богданова
Разработка веб-проектов в компании без проект-менеджеров — Артур Богданова
 
Coolant for lathe
Coolant for latheCoolant for lathe
Coolant for lathe
 
Tgv sócretino
Tgv sócretinoTgv sócretino
Tgv sócretino
 
Esofago
EsofagoEsofago
Esofago
 
32 Ways a Digital Marketing Consultant Can Help Grow Your Business
32 Ways a Digital Marketing Consultant Can Help Grow Your Business32 Ways a Digital Marketing Consultant Can Help Grow Your Business
32 Ways a Digital Marketing Consultant Can Help Grow Your Business
 

Similaire à Jst part4

10 Backpropagation Algorithm for Neural Networks (1).pptx
10 Backpropagation Algorithm for Neural Networks (1).pptx10 Backpropagation Algorithm for Neural Networks (1).pptx
10 Backpropagation Algorithm for Neural Networks (1).pptx
SaifKhan703888
 
nural network ER. Abhishek k. upadhyay
nural network ER. Abhishek  k. upadhyaynural network ER. Abhishek  k. upadhyay
nural network ER. Abhishek k. upadhyay
abhishek upadhyay
 
Chapter No. 6: Backpropagation Networks
Chapter No. 6:  Backpropagation NetworksChapter No. 6:  Backpropagation Networks
Chapter No. 6: Backpropagation Networks
RamkrishnaPatil17
 
ES_SAA_OG_PF_ECCTD_Pos
ES_SAA_OG_PF_ECCTD_PosES_SAA_OG_PF_ECCTD_Pos
ES_SAA_OG_PF_ECCTD_Pos
Syed Asad Alam
 
Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...
Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...
Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...
Simplilearn
 

Similaire à Jst part4 (20)

03 Single layer Perception Classifier
03 Single layer Perception Classifier03 Single layer Perception Classifier
03 Single layer Perception Classifier
 
10 Backpropagation Algorithm for Neural Networks (1).pptx
10 Backpropagation Algorithm for Neural Networks (1).pptx10 Backpropagation Algorithm for Neural Networks (1).pptx
10 Backpropagation Algorithm for Neural Networks (1).pptx
 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural Network
 
Unit 1
Unit 1Unit 1
Unit 1
 
2-Perceptrons.pdf
2-Perceptrons.pdf2-Perceptrons.pdf
2-Perceptrons.pdf
 
Artificial Neural Networks
Artificial Neural NetworksArtificial Neural Networks
Artificial Neural Networks
 
Artificial neural network by arpit_sharma
Artificial neural network by arpit_sharmaArtificial neural network by arpit_sharma
Artificial neural network by arpit_sharma
 
Multi Layer Network
Multi Layer NetworkMulti Layer Network
Multi Layer Network
 
ann-ics320Part4.ppt
ann-ics320Part4.pptann-ics320Part4.ppt
ann-ics320Part4.ppt
 
ann-ics320Part4.ppt
ann-ics320Part4.pptann-ics320Part4.ppt
ann-ics320Part4.ppt
 
nural network ER. Abhishek k. upadhyay
nural network ER. Abhishek  k. upadhyaynural network ER. Abhishek  k. upadhyay
nural network ER. Abhishek k. upadhyay
 
Introduction to Neural networks (under graduate course) Lecture 4 of 9
Introduction to Neural networks (under graduate course) Lecture 4 of 9Introduction to Neural networks (under graduate course) Lecture 4 of 9
Introduction to Neural networks (under graduate course) Lecture 4 of 9
 
Chapter No. 6: Backpropagation Networks
Chapter No. 6:  Backpropagation NetworksChapter No. 6:  Backpropagation Networks
Chapter No. 6: Backpropagation Networks
 
NN-Ch6.PDF
NN-Ch6.PDFNN-Ch6.PDF
NN-Ch6.PDF
 
InternImage: Exploring Large-Scale Vision Foundation Models with Deformable C...
InternImage: Exploring Large-Scale Vision Foundation Models with Deformable C...InternImage: Exploring Large-Scale Vision Foundation Models with Deformable C...
InternImage: Exploring Large-Scale Vision Foundation Models with Deformable C...
 
Artificial neural networks
Artificial neural networksArtificial neural networks
Artificial neural networks
 
6
66
6
 
ES_SAA_OG_PF_ECCTD_Pos
ES_SAA_OG_PF_ECCTD_PosES_SAA_OG_PF_ECCTD_Pos
ES_SAA_OG_PF_ECCTD_Pos
 
Neural
NeuralNeural
Neural
 
Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...
Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...
Deep Learning Interview Questions And Answers | AI & Deep Learning Interview ...
 

Plus de Rendy Ardiwinata (20)

Hoopers a. kamasutra. sex positions-dk
Hoopers a. kamasutra. sex positions-dkHoopers a. kamasutra. sex positions-dk
Hoopers a. kamasutra. sex positions-dk
 
4 plc-dasar-dasar-pemrograman-new-6
4 plc-dasar-dasar-pemrograman-new-64 plc-dasar-dasar-pemrograman-new-6
4 plc-dasar-dasar-pemrograman-new-6
 
Jst part6
Jst part6Jst part6
Jst part6
 
Jst part5
Jst part5Jst part5
Jst part5
 
Jst part3
Jst part3Jst part3
Jst part3
 
Jst part2
Jst part2Jst part2
Jst part2
 
Jst part1
Jst part1Jst part1
Jst part1
 
Fuzzy logic part7
Fuzzy logic part7Fuzzy logic part7
Fuzzy logic part7
 
Fuzzy logic part6
Fuzzy logic part6Fuzzy logic part6
Fuzzy logic part6
 
Fuzzy logic part4
Fuzzy logic part4Fuzzy logic part4
Fuzzy logic part4
 
Fuzzy logic part3
Fuzzy logic part3Fuzzy logic part3
Fuzzy logic part3
 
Fuzzy logic part2
Fuzzy logic part2Fuzzy logic part2
Fuzzy logic part2
 
Fuzzy logic part1
Fuzzy logic part1Fuzzy logic part1
Fuzzy logic part1
 
Fuzzy logic part5
Fuzzy logic part5Fuzzy logic part5
Fuzzy logic part5
 
Customer ptcpi for lifting process
Customer ptcpi for lifting processCustomer ptcpi for lifting process
Customer ptcpi for lifting process
 
1 n4148 1n4448
1 n4148 1n44481 n4148 1n4448
1 n4148 1n4448
 
Hukum tajwid
Hukum tajwidHukum tajwid
Hukum tajwid
 
Bridge circuits
Bridge circuitsBridge circuits
Bridge circuits
 
Filters DAC and ADC
Filters DAC and ADCFilters DAC and ADC
Filters DAC and ADC
 
Bab 2 kontrol sekuensial PLC
Bab 2 kontrol sekuensial PLCBab 2 kontrol sekuensial PLC
Bab 2 kontrol sekuensial PLC
 

Dernier

Spellings Wk 3 English CAPS CARES Please Practise
Spellings Wk 3 English CAPS CARES Please PractiseSpellings Wk 3 English CAPS CARES Please Practise
Spellings Wk 3 English CAPS CARES Please Practise
AnaAcapella
 
The basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptxThe basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptx
heathfieldcps1
 

Dernier (20)

Spellings Wk 3 English CAPS CARES Please Practise
Spellings Wk 3 English CAPS CARES Please PractiseSpellings Wk 3 English CAPS CARES Please Practise
Spellings Wk 3 English CAPS CARES Please Practise
 
General Principles of Intellectual Property: Concepts of Intellectual Proper...
General Principles of Intellectual Property: Concepts of Intellectual  Proper...General Principles of Intellectual Property: Concepts of Intellectual  Proper...
General Principles of Intellectual Property: Concepts of Intellectual Proper...
 
Unit-IV- Pharma. Marketing Channels.pptx
Unit-IV- Pharma. Marketing Channels.pptxUnit-IV- Pharma. Marketing Channels.pptx
Unit-IV- Pharma. Marketing Channels.pptx
 
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfKey note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdf
 
The basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptxThe basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptx
 
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
 
Google Gemini An AI Revolution in Education.pptx
Google Gemini An AI Revolution in Education.pptxGoogle Gemini An AI Revolution in Education.pptx
Google Gemini An AI Revolution in Education.pptx
 
ICT role in 21st century education and it's challenges.
ICT role in 21st century education and it's challenges.ICT role in 21st century education and it's challenges.
ICT role in 21st century education and it's challenges.
 
Unit-IV; Professional Sales Representative (PSR).pptx
Unit-IV; Professional Sales Representative (PSR).pptxUnit-IV; Professional Sales Representative (PSR).pptx
Unit-IV; Professional Sales Representative (PSR).pptx
 
Dyslexia AI Workshop for Slideshare.pptx
Dyslexia AI Workshop for Slideshare.pptxDyslexia AI Workshop for Slideshare.pptx
Dyslexia AI Workshop for Slideshare.pptx
 
Understanding Accommodations and Modifications
Understanding  Accommodations and ModificationsUnderstanding  Accommodations and Modifications
Understanding Accommodations and Modifications
 
Sociology 101 Demonstration of Learning Exhibit
Sociology 101 Demonstration of Learning ExhibitSociology 101 Demonstration of Learning Exhibit
Sociology 101 Demonstration of Learning Exhibit
 
How to Create and Manage Wizard in Odoo 17
How to Create and Manage Wizard in Odoo 17How to Create and Manage Wizard in Odoo 17
How to Create and Manage Wizard in Odoo 17
 
How to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POSHow to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POS
 
ICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptxICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptx
 
Mehran University Newsletter Vol-X, Issue-I, 2024
Mehran University Newsletter Vol-X, Issue-I, 2024Mehran University Newsletter Vol-X, Issue-I, 2024
Mehran University Newsletter Vol-X, Issue-I, 2024
 
On National Teacher Day, meet the 2024-25 Kenan Fellows
On National Teacher Day, meet the 2024-25 Kenan FellowsOn National Teacher Day, meet the 2024-25 Kenan Fellows
On National Teacher Day, meet the 2024-25 Kenan Fellows
 
Graduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - EnglishGraduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - English
 
Introduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The BasicsIntroduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The Basics
 
Kodo Millet PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...
Kodo Millet  PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...Kodo Millet  PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...
Kodo Millet PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...
 

Jst part4

  • 1. 66//1010//20132013 Training in Neural Network • Learn values of weights from I/O pairs • Start with random weights • Load training example’s input • Observe computed input • Modify weights to reduce difference • Iterate over all training examples • Terminate when weights stop changing OR when error is very small Single Layer Perceptrons • A network with all the inputs connected directly to the output (figure 1) • In simple cases, feature space is divided by drawing a hyperplane across it, which is Known as a decision boundary • Problems which can be thus classified are linearly separable (figure 2) ++ ++ ++ ++ -- -- -- -- xx11 xx22 FigureFigure 11 FigureFigure 22
  • 2. 66//1010//20132013 Single Layer Perceptrons • Classical Measure of Error – Squared error – Where Err is the difference between the target value and the output by the network • Weight Modifying – Use gradient descent to reduce the squared error by calculating the partial derivative of E with respect to each weight j jj xinfErr Err Err E *** )('−= ∂ ∂ = ∂ ∂ ww : Learning Rate: Learning Rate Error Back-Propagation The gradient descent • The gradient of the error E gives the direction where the error function at the current setting of the w will increase. In order to decrease E, we take a small step in the opposite direction, -G
  • 3. 66//1010//20132013 Error Back-Propagation The gradient descentThe gradient descent InIn 22D (one weight)D (one weight) By repeating this over and over, we moveBy repeating this over and over, we move "downhill" in"downhill" in EE until we reach a minimumuntil we reach a minimum Error Back-Propagation The gradient descentThe gradient descent
  • 4. 66//1010//20132013 Single Layer Perceptrons – a basic application • Suppose we have data about the height and the age of a population of 100 people. • So we can plot a 2D sketch (x is the age, y the height) • How can we predict the height of a 101th person, given his age? UsingUsing aa modelmodel of the data. The simplestof the data. The simplest model can be :model can be : y = wy = w11 x + wx + w00 This may exactly be the equation of theThis may exactly be the equation of the output of a neuron networkoutput of a neuron network Different Non-Linearly Separable Problems StructureStructure ExclusiveExclusive--OROR ProblemProblem Classes withClasses with Meshed regionsMeshed regions Most GeneralMost General Region ShapesRegion Shapes SingleSingle--LayerLayer TwoTwo--LayerLayer ThreeThree--LayerLayer AA AABB BB AA AABB BB AA AABB BB BB AA BB AA BB AA
  • 5. 66//1010//20132013 Multilayer Perceptrons(MLP) HiddenHidden LayerLayer Output LayerOutput Layer AdjustableAdjustable wweightseights IntputIntput LayerLayer AdjustableAdjustable wweightseights InputInputUnitUnits(ExternalStimuli)s(ExternalStimuli) OutputValuesOutputValues Types of Layers • Input layer (units) – Introduces input values into the network – No activation function or other processing • Hidden layer(s) – Perform classification of features – Two hidden layers are sufficient to solve any problem – Features imply more layers may be better • Output layer – Functionally just like the hidden layers – Outputs are passed on to the world outside the neural network
  • 6. 66//1010//20132013 MLP Characteristics • Input propagates in a forward direction, layer-by-layer basis – also called Multilayer Feedforward Network, MLP • Non-linear activation function – differentiable – nonlinearity prevent reduction to single-layer perceptron • One or more layers of hidden neurons – progressively extracting more meaningful features from input patterns • High degree of connectivity Problems of MLP • Nonlinearity and high degree of connectivity makes theoretical analysis difficult • Output vector rather than a single output value • Error at output layer is clear, but error at the hidden layers seems mysterious • Learning process is hard to visualize • So, Error Back-Propagation Algorithm (BPA) is a computationally efficient training for MLP