SlideShare ist ein Scribd-Unternehmen logo
1 von 22
Downloaden Sie, um offline zu lesen
A Brief Introduction to
Artificial Neural Networks
http://www.raymazza.com/uploads/1/0/6/2/10622510/4006097_orig.jpg
• A neuron is electrically excitable
cell that processes and transmits
information through electrical and
chemical signals.
• It’s the basic unit of computation in
our nervous system.
• A typical neuron possesses :
• Dendrites : Input
• Cell body (Soma) : Processor
• Axon : Output
• Neurons communicate with one
another via synapses which can be
excitatory or inhibitory.
What is a neuron ?
http://www.urbanchildinstitute.org/sites/all/files/databooks/2011/ch1-fg2-communication-between-neurons.jpg
Neural Network
• A neural network is an interconnected web
of neurons transmitting elaborate patterns
of electrical signals.
• Dendrites receive input signals and, based
on those inputs, fire an output signal via an
axon.
• A neural network is a adaptive system,
meaning it can change its internal structure
based on the information flowing through it.
• Typically, this is achieved through the
adjusting of weights.
http://www.fridayfonts.com/wp-content/uploads/2009/05/neurons.jpg
Artificial neuron
• An artificial neuron is a mathematical model of biological neurons which receives one or
more inputs and sums them to produce an output.
• Usually the sums of each node are weighted, and the sum is passed through a non-
linear function known as an activation function.
http://upload.wikimedia.org/wikipedia/commons/thumb/6/60/ArtificialNeuronModel_english.png/600px-ArtificialNeuronModel_english.png
Artificial Neural Network (ANN)
• An ANN is typically defined by three types of parameters:
• The interconnection pattern between the different layers of neurons
• The learning process for updating the weights of the interconnections
• The activation function that converts a neuron's weighted input to its output
http://wp.fobiss.com/wp-content/uploads/2013/06/Brain2.png
Artificial Neuron Models
Perceptron
• Invented in 1957 by Frank Rosenblatt at the Cornell Aeronautical Laboratory.
• It is a computational model of a single neuron.
• It consists of one or more binary inputs, a processor, and a single binary output.
• Binary inputs are multiplied by weights [real numbers expressing the importance of the
respective inputs to the output] & then fed into processor.
Activation function
http://neuralnetworksanddeeplearning.com/chap1.html
Stochastic Model
𝑝 𝑥 =
1
1 + 𝑒− 𝑣
𝑇
𝑊𝑖𝑡ℎ 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 1 − 𝑝(𝑥)
𝑊𝑖𝑡ℎ 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑝(𝑥)
Where T is pseudo temperature.
• For T=0 :
• Sigmoid neuron inputs and output can take any value between 0 and 1.
• Sigmoid neuron activation function is given as :
𝜎 𝑤. 𝑥 + 𝑏 =
1
1+𝑒(−𝑤.𝑥−𝑏)
Sigmoidal Neuron
• The smoothness of σ allows us to write a small change Δoutput in terms of small
changes Δwj in the weights and Δb in the bias as :
http://neuralnetworksanddeeplearning.com/chap1.html
Neural network architecture
• The leftmost layer in this network is called the input layer, and the neurons within the
layer are called input neurons. The rightmost or output layer contains the output neurons.
The middle layer is called a hidden layer, since the neurons in this layer are neither inputs
nor outputs.
Feed-forward network
http://neuralnetworksanddeeplearning.com/chap1.html
Design of Input and Output Layers
• The design of the input and output layers in a network is often straightforward.
• For example, suppose we're trying to determine whether a handwritten image depicts a
"9" or not.
• A natural way to design the network is to encode the intensities of the image pixels into
the input neurons.
• If the image is a 64 by 64 greyscale image, then we'd have 4,096=64×64 input neurons,
with the intensities scaled appropriately between 0 and 1.
• The output layer will contain just a single neuron, with output values of less than 0.5
indicating "input image is not a 9", and values greater than 0.5 indicating "input image is
a 9 ".
Learning strategy
• Now we want an algorithm which lets us find weights and biases so that the output from
the network approximates the desired output for all training inputs x. For that purpose we
define a cost function :
𝐶 𝑤, 𝑏 =
1
2𝑛 𝑥 𝑑 − 𝑦(𝑥) 2
• The aim of our training algorithm is to find a set of weights and biases which minimizes
the cost function.
• We'll do that using an algorithm known as gradient descent.
Where,
w : the collection of all weights in the network
b : all the biases
n : total number of training inputs,
d : desired output from network
y(x) : Actual output from network
Gradient Descent Algorithm
• Let's suppose we're trying to minimize some function :
𝐶 𝒗 𝑤ℎ𝑒𝑟𝑒, 𝒗 = (𝑣1, 𝑣2, … . , 𝑣 𝑚) 𝑇 𝑣1, 𝑣2, … . , 𝑣 𝑚 ∈ ℝ
• To visualize C(v) it helps to imagine C as a function of just two variables, which we'll call
v1 and v2
What we'd like is to find where C achieves
its global minimum.
http://neuralnetworksanddeeplearning.com/chap1.html
• A small change in 𝐶(𝒗) can be written as :
∆𝐶 =
𝜕𝐶
𝜕𝑣1
∆𝑣1 +
𝜕𝐶
𝜕𝑣2
∆𝑣2 + ⋯ +
𝜕𝐶
𝜕𝑣 𝑚
∆𝑣 𝑚
• Above expression can also be written as :
∆𝐶 = 𝜵𝑪. ∆𝒗
• Where ,
𝜵𝑪 =
𝜕𝐶
𝜕𝑣1
,
𝜕𝐶
𝜕𝑣2
, … ,
𝜕𝐶
𝜕𝑣 𝑚
𝑇
∆𝒗 = ∆𝑣1, ∆𝑣2, … , ∆𝑣 𝑚
• Now ,we want to choose ∆𝒗 such that ∆𝐶 ≤ 0. If we choose :
∆𝒗 = −𝜂𝜵𝑪 ⇒ ∆𝐶 = −𝜂 𝜵𝑪 2 ≤ 0
• Now , the next point on the surface can be given as :
𝒗′ = 𝒗 + ∆𝑣
𝒗′ = 𝒗 − 𝜂𝜵𝑪
• We'll use above updating rule to compute a value for new value of 𝑣1, 𝑣2,…, 𝑣 𝑚 :
𝑣1
′
= 𝑣1 − 𝜂
𝜕𝐶
𝜕𝑣1
𝑣2
′
= 𝑣2 − 𝜂
𝜕𝐶
𝜕𝑣2
.
.
𝑣 𝑚
′
= 𝑣 𝑚 − 𝜂
𝜕𝐶
𝜕𝑣 𝑚
• Then we'll use this update rule again, to make another move. If we keep doing this, over
and over, we'll keep decreasing C until , we hope , we reach a global minimum.
𝜂 = 𝐿𝑒𝑎𝑟𝑛𝑖𝑛𝑔 𝑟𝑎𝑡𝑒
Update rule for weights & bias
• In case of a neural network our variables are : 𝑤𝑖 𝑎𝑛𝑑 𝑏𝑖
• Using gradient descent algorithm, the update rule for weights and biases can be given as :
𝑤𝑖
′ = 𝑤𝑖 − 𝜂
𝜕𝐶
𝜕𝑤𝑖
𝑏𝑖
′
= 𝑏𝑖 − 𝜂
𝜕𝐶
𝜕𝑏𝑖
• By repeatedly applying this update rule we can "roll down the hill", and hopefully find a
minimum of the cost function. In other words, this is a rule which can be used to learn in
a neural network.
Practical applications
• To recognize individual digits we will use a three-layer neural network:
http://neuralnetworksanddeeplearning.com/chap1.html
Fitting Straight to the data
• A linear model of neuron is used for fitting straight line to our data.
• In case of N dimension , there are N free parameters : (N-1) slope components and an
intercept.
𝑥1
𝑥 𝑛−1
𝑦𝑛
𝑤 𝑛0
𝑤 𝑛1
𝑤 𝑛(𝑛−1)
+1
𝑦𝑛 = 𝑤 𝑛0 +
𝑗=1
𝑛−1
𝑥𝑗 𝑤 𝑛𝑗
Output Input
𝑦𝑝
𝑡 𝑝
𝑝
𝐶 =
𝑝
𝐶 𝑝
𝐶 𝑝 =
1
2
𝑡 𝑝 − 𝑦𝑝
2
Update rule for weights
• Using gradient descent algorithm, the update rule for weights and biases can be given as :
𝑤 𝑛𝑖
′
= 𝑤 𝑛𝑖 − 𝜂
𝜕𝐶
𝜕𝑤 𝑛𝑖
• Now , the derivative in above equation can be calculated using chain rule as follows :
𝜕𝐶
𝜕𝑤 𝑛𝑖
=
𝜕
𝜕𝑤 𝑛𝑖
𝑝
𝐶 𝑝 =
𝑝
𝜕𝐶 𝑝
𝜕𝑤 𝑛𝑖
𝜕𝐶 𝑝
𝜕𝑤 𝑛𝑖
=
𝜕𝐶 𝑝
𝜕𝑦𝑝
𝜕𝑦𝑝
𝜕𝑤 𝑛𝑖
𝜕𝐶 𝑝
𝜕𝑦𝑝
=
𝜕
𝜕𝑦𝑝
1
2
𝑡 𝑝 − 𝑦𝑝
2
= − 𝑡 𝑝 − 𝑦𝑝
𝜕𝑦𝑝
𝜕𝑤 𝑛𝑖
=
𝜕
𝜕𝑤 𝑛𝑖
𝑗=0
𝑛
𝑥𝑗 𝑤 𝑛𝑗 = 𝑥𝑖
• Using gradient descent algorithm, the update rule for weights and biases can be given as :
𝑤 𝑛𝑖
′
= 𝑤 𝑛𝑖 − 𝜂
𝜕𝐶
𝜕𝑤 𝑛𝑖
𝜕𝐶
𝜕𝑤 𝑛𝑖
=
𝑝
−𝑥𝑖 𝑡 𝑝 − 𝑦𝑝
𝑤 𝑛𝑖
′
= 𝑤 𝑛𝑖 + 𝜂
𝑝
𝑥𝑖 𝑡 𝑝 − 𝑦𝑝
References
• http://neuralnetworksanddeeplearning.com/chap1.html
• http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-034-artificial-intelligence-fall-2010/
• https://en.wikipedia.org/
• https://images.google.com/
Thank You

Weitere ähnliche Inhalte

Was ist angesagt?

lecture07.ppt
lecture07.pptlecture07.ppt
lecture07.pptbutest
 
Introduction to Neural networks (under graduate course) Lecture 8 of 9
Introduction to Neural networks (under graduate course) Lecture 8 of 9Introduction to Neural networks (under graduate course) Lecture 8 of 9
Introduction to Neural networks (under graduate course) Lecture 8 of 9Randa Elanwar
 
Artificial Neural Networks Lect2: Neurobiology & Architectures of ANNS
Artificial Neural Networks Lect2: Neurobiology & Architectures of ANNSArtificial Neural Networks Lect2: Neurobiology & Architectures of ANNS
Artificial Neural Networks Lect2: Neurobiology & Architectures of ANNSMohammed Bennamoun
 
Neural networks
Neural networksNeural networks
Neural networksSlideshare
 
Perceptron (neural network)
Perceptron (neural network)Perceptron (neural network)
Perceptron (neural network)EdutechLearners
 
Activation function
Activation functionActivation function
Activation functionAstha Jain
 
Multilayer perceptron
Multilayer perceptronMultilayer perceptron
Multilayer perceptronsmitamm
 
2.5 backpropagation
2.5 backpropagation2.5 backpropagation
2.5 backpropagationKrish_ver2
 
Artificial Neural Network
Artificial Neural Network Artificial Neural Network
Artificial Neural Network Iman Ardekani
 
Neural networks1
Neural networks1Neural networks1
Neural networks1Mohan Raj
 
Introduction to Neural networks (under graduate course) Lecture 7 of 9
Introduction to Neural networks (under graduate course) Lecture 7 of 9Introduction to Neural networks (under graduate course) Lecture 7 of 9
Introduction to Neural networks (under graduate course) Lecture 7 of 9Randa Elanwar
 
Comparative study of ANNs and BNNs and mathematical modeling of a neuron
Comparative study of ANNs and BNNs and mathematical modeling of a neuronComparative study of ANNs and BNNs and mathematical modeling of a neuron
Comparative study of ANNs and BNNs and mathematical modeling of a neuronSaransh Choudhary
 
Artificial neural network for concrete mix design
Artificial neural network for concrete mix designArtificial neural network for concrete mix design
Artificial neural network for concrete mix designMonjurul Shuvo
 
Counter propagation Network
Counter propagation NetworkCounter propagation Network
Counter propagation NetworkAkshay Dhole
 
Neural networks
Neural networksNeural networks
Neural networksSlideshare
 
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
 

Was ist angesagt? (20)

04 Multi-layer Feedforward Networks
04 Multi-layer Feedforward Networks04 Multi-layer Feedforward Networks
04 Multi-layer Feedforward Networks
 
lecture07.ppt
lecture07.pptlecture07.ppt
lecture07.ppt
 
Introduction to Neural networks (under graduate course) Lecture 8 of 9
Introduction to Neural networks (under graduate course) Lecture 8 of 9Introduction to Neural networks (under graduate course) Lecture 8 of 9
Introduction to Neural networks (under graduate course) Lecture 8 of 9
 
Artificial Neural Networks Lect2: Neurobiology & Architectures of ANNS
Artificial Neural Networks Lect2: Neurobiology & Architectures of ANNSArtificial Neural Networks Lect2: Neurobiology & Architectures of ANNS
Artificial Neural Networks Lect2: Neurobiology & Architectures of ANNS
 
Neural networks
Neural networksNeural networks
Neural networks
 
Perceptron (neural network)
Perceptron (neural network)Perceptron (neural network)
Perceptron (neural network)
 
Activation function
Activation functionActivation function
Activation function
 
Multilayer perceptron
Multilayer perceptronMultilayer perceptron
Multilayer perceptron
 
Neural network and mlp
Neural network and mlpNeural network and mlp
Neural network and mlp
 
2.5 backpropagation
2.5 backpropagation2.5 backpropagation
2.5 backpropagation
 
Artificial Neural Network
Artificial Neural Network Artificial Neural Network
Artificial Neural Network
 
Neural networks1
Neural networks1Neural networks1
Neural networks1
 
Introduction to Neural networks (under graduate course) Lecture 7 of 9
Introduction to Neural networks (under graduate course) Lecture 7 of 9Introduction to Neural networks (under graduate course) Lecture 7 of 9
Introduction to Neural networks (under graduate course) Lecture 7 of 9
 
Comparative study of ANNs and BNNs and mathematical modeling of a neuron
Comparative study of ANNs and BNNs and mathematical modeling of a neuronComparative study of ANNs and BNNs and mathematical modeling of a neuron
Comparative study of ANNs and BNNs and mathematical modeling of a neuron
 
Artificial neural network for concrete mix design
Artificial neural network for concrete mix designArtificial neural network for concrete mix design
Artificial neural network for concrete mix design
 
03 Single layer Perception Classifier
03 Single layer Perception Classifier03 Single layer Perception Classifier
03 Single layer Perception Classifier
 
Counter propagation Network
Counter propagation NetworkCounter propagation Network
Counter propagation Network
 
Unit 1
Unit 1Unit 1
Unit 1
 
Neural networks
Neural networksNeural networks
Neural networks
 
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
 

Ähnlich wie Artificial neural networks

08 neural networks
08 neural networks08 neural networks
08 neural networksankit_ppt
 
nural network ER. Abhishek k. upadhyay
nural network ER. Abhishek  k. upadhyaynural network ER. Abhishek  k. upadhyay
nural network ER. Abhishek k. upadhyayabhishek upadhyay
 
Neural-Networks.ppt
Neural-Networks.pptNeural-Networks.ppt
Neural-Networks.pptRINUSATHYAN
 
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
 
Neural network basic and introduction of Deep learning
Neural network basic and introduction of Deep learningNeural network basic and introduction of Deep learning
Neural network basic and introduction of Deep learningTapas Majumdar
 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural NetworkAtul Krishna
 
Machine learning Module-2, 6th Semester Elective
Machine learning Module-2, 6th Semester ElectiveMachine learning Module-2, 6th Semester Elective
Machine learning Module-2, 6th Semester ElectiveMayuraD1
 
Artificial Neural Network_VCW (1).pptx
Artificial Neural Network_VCW (1).pptxArtificial Neural Network_VCW (1).pptx
Artificial Neural Network_VCW (1).pptxpratik610182
 
Introduction to Neural networks (under graduate course) Lecture 9 of 9
Introduction to Neural networks (under graduate course) Lecture 9 of 9Introduction to Neural networks (under graduate course) Lecture 9 of 9
Introduction to Neural networks (under graduate course) Lecture 9 of 9Randa Elanwar
 
artificialneuralnetwork-130409001108-phpapp02 (2).pptx
artificialneuralnetwork-130409001108-phpapp02 (2).pptxartificialneuralnetwork-130409001108-phpapp02 (2).pptx
artificialneuralnetwork-130409001108-phpapp02 (2).pptxREG83NITHYANANTHANN
 
Multilayer Perceptron Neural Network MLP
Multilayer Perceptron Neural Network MLPMultilayer Perceptron Neural Network MLP
Multilayer Perceptron Neural Network MLPAbdullah al Mamun
 
Hands on machine learning with scikit-learn and tensor flow by ahmed yousry
Hands on machine learning with scikit-learn and tensor flow by ahmed yousryHands on machine learning with scikit-learn and tensor flow by ahmed yousry
Hands on machine learning with scikit-learn and tensor flow by ahmed yousryAhmed Yousry
 

Ähnlich wie Artificial neural networks (20)

08 neural networks
08 neural networks08 neural networks
08 neural networks
 
nural network ER. Abhishek k. upadhyay
nural network ER. Abhishek  k. upadhyaynural network ER. Abhishek  k. upadhyay
nural network ER. Abhishek k. upadhyay
 
Neural-Networks.ppt
Neural-Networks.pptNeural-Networks.ppt
Neural-Networks.ppt
 
Neural
NeuralNeural
Neural
 
Neural Networks
Neural NetworksNeural Networks
Neural Networks
 
Deep learning
Deep learningDeep learning
Deep learning
 
Unit 2 ml.pptx
Unit 2 ml.pptxUnit 2 ml.pptx
Unit 2 ml.pptx
 
Artificial neural network by arpit_sharma
Artificial neural network by arpit_sharmaArtificial neural network by arpit_sharma
Artificial neural network by arpit_sharma
 
ANN.pptx
ANN.pptxANN.pptx
ANN.pptx
 
Neural network basic and introduction of Deep learning
Neural network basic and introduction of Deep learningNeural network basic and introduction of Deep learning
Neural network basic and introduction of Deep learning
 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural Network
 
Machine learning Module-2, 6th Semester Elective
Machine learning Module-2, 6th Semester ElectiveMachine learning Module-2, 6th Semester Elective
Machine learning Module-2, 6th Semester Elective
 
Artificial Neural Network_VCW (1).pptx
Artificial Neural Network_VCW (1).pptxArtificial Neural Network_VCW (1).pptx
Artificial Neural Network_VCW (1).pptx
 
UNIT 5-ANN.ppt
UNIT 5-ANN.pptUNIT 5-ANN.ppt
UNIT 5-ANN.ppt
 
Introduction to Neural networks (under graduate course) Lecture 9 of 9
Introduction to Neural networks (under graduate course) Lecture 9 of 9Introduction to Neural networks (under graduate course) Lecture 9 of 9
Introduction to Neural networks (under graduate course) Lecture 9 of 9
 
artificialneuralnetwork-130409001108-phpapp02 (2).pptx
artificialneuralnetwork-130409001108-phpapp02 (2).pptxartificialneuralnetwork-130409001108-phpapp02 (2).pptx
artificialneuralnetwork-130409001108-phpapp02 (2).pptx
 
Multilayer Perceptron Neural Network MLP
Multilayer Perceptron Neural Network MLPMultilayer Perceptron Neural Network MLP
Multilayer Perceptron Neural Network MLP
 
Unit iii update
Unit iii updateUnit iii update
Unit iii update
 
19_Learning.ppt
19_Learning.ppt19_Learning.ppt
19_Learning.ppt
 
Hands on machine learning with scikit-learn and tensor flow by ahmed yousry
Hands on machine learning with scikit-learn and tensor flow by ahmed yousryHands on machine learning with scikit-learn and tensor flow by ahmed yousry
Hands on machine learning with scikit-learn and tensor flow by ahmed yousry
 

Kürzlich hochgeladen

Caco-2 cell permeability assay for drug absorption
Caco-2 cell permeability assay for drug absorptionCaco-2 cell permeability assay for drug absorption
Caco-2 cell permeability assay for drug absorptionPriyansha Singh
 
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxSOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxkessiyaTpeter
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisDiwakar Mishra
 
Types of different blotting techniques.pptx
Types of different blotting techniques.pptxTypes of different blotting techniques.pptx
Types of different blotting techniques.pptxkhadijarafiq2012
 
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptxUnlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptxanandsmhk
 
Isotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on IoIsotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on IoSérgio Sacani
 
Stunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCR
Stunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCRStunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCR
Stunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCRDelhi Call girls
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Patrick Diehl
 
Recombination DNA Technology (Nucleic Acid Hybridization )
Recombination DNA Technology (Nucleic Acid Hybridization )Recombination DNA Technology (Nucleic Acid Hybridization )
Recombination DNA Technology (Nucleic Acid Hybridization )aarthirajkumar25
 
Work, Energy and Power for class 10 ICSE Physics
Work, Energy and Power for class 10 ICSE PhysicsWork, Energy and Power for class 10 ICSE Physics
Work, Energy and Power for class 10 ICSE Physicsvishikhakeshava1
 
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...Sérgio Sacani
 
Botany 4th semester file By Sumit Kumar yadav.pdf
Botany 4th semester file By Sumit Kumar yadav.pdfBotany 4th semester file By Sumit Kumar yadav.pdf
Botany 4th semester file By Sumit Kumar yadav.pdfSumit Kumar yadav
 
Boyles law module in the grade 10 science
Boyles law module in the grade 10 scienceBoyles law module in the grade 10 science
Boyles law module in the grade 10 sciencefloriejanemacaya1
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxgindu3009
 
Analytical Profile of Coleus Forskohlii | Forskolin .pdf
Analytical Profile of Coleus Forskohlii | Forskolin .pdfAnalytical Profile of Coleus Forskohlii | Forskolin .pdf
Analytical Profile of Coleus Forskohlii | Forskolin .pdfSwapnil Therkar
 
Disentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOSTDisentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOSTSérgio Sacani
 
GFP in rDNA Technology (Biotechnology).pptx
GFP in rDNA Technology (Biotechnology).pptxGFP in rDNA Technology (Biotechnology).pptx
GFP in rDNA Technology (Biotechnology).pptxAleenaTreesaSaji
 
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...jana861314
 
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxAnalytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxSwapnil Therkar
 

Kürzlich hochgeladen (20)

Caco-2 cell permeability assay for drug absorption
Caco-2 cell permeability assay for drug absorptionCaco-2 cell permeability assay for drug absorption
Caco-2 cell permeability assay for drug absorption
 
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptxSOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
SOLUBLE PATTERN RECOGNITION RECEPTORS.pptx
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
 
The Philosophy of Science
The Philosophy of ScienceThe Philosophy of Science
The Philosophy of Science
 
Types of different blotting techniques.pptx
Types of different blotting techniques.pptxTypes of different blotting techniques.pptx
Types of different blotting techniques.pptx
 
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptxUnlocking  the Potential: Deep dive into ocean of Ceramic Magnets.pptx
Unlocking the Potential: Deep dive into ocean of Ceramic Magnets.pptx
 
Isotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on IoIsotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on Io
 
Stunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCR
Stunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCRStunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCR
Stunning ➥8448380779▻ Call Girls In Panchshil Enclave Delhi NCR
 
Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?Is RISC-V ready for HPC workload? Maybe?
Is RISC-V ready for HPC workload? Maybe?
 
Recombination DNA Technology (Nucleic Acid Hybridization )
Recombination DNA Technology (Nucleic Acid Hybridization )Recombination DNA Technology (Nucleic Acid Hybridization )
Recombination DNA Technology (Nucleic Acid Hybridization )
 
Work, Energy and Power for class 10 ICSE Physics
Work, Energy and Power for class 10 ICSE PhysicsWork, Energy and Power for class 10 ICSE Physics
Work, Energy and Power for class 10 ICSE Physics
 
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
 
Botany 4th semester file By Sumit Kumar yadav.pdf
Botany 4th semester file By Sumit Kumar yadav.pdfBotany 4th semester file By Sumit Kumar yadav.pdf
Botany 4th semester file By Sumit Kumar yadav.pdf
 
Boyles law module in the grade 10 science
Boyles law module in the grade 10 scienceBoyles law module in the grade 10 science
Boyles law module in the grade 10 science
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptx
 
Analytical Profile of Coleus Forskohlii | Forskolin .pdf
Analytical Profile of Coleus Forskohlii | Forskolin .pdfAnalytical Profile of Coleus Forskohlii | Forskolin .pdf
Analytical Profile of Coleus Forskohlii | Forskolin .pdf
 
Disentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOSTDisentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOST
 
GFP in rDNA Technology (Biotechnology).pptx
GFP in rDNA Technology (Biotechnology).pptxGFP in rDNA Technology (Biotechnology).pptx
GFP in rDNA Technology (Biotechnology).pptx
 
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
Traditional Agroforestry System in India- Shifting Cultivation, Taungya, Home...
 
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptxAnalytical Profile of Coleus Forskohlii | Forskolin .pptx
Analytical Profile of Coleus Forskohlii | Forskolin .pptx
 

Artificial neural networks

  • 1. A Brief Introduction to Artificial Neural Networks http://www.raymazza.com/uploads/1/0/6/2/10622510/4006097_orig.jpg
  • 2. • A neuron is electrically excitable cell that processes and transmits information through electrical and chemical signals. • It’s the basic unit of computation in our nervous system. • A typical neuron possesses : • Dendrites : Input • Cell body (Soma) : Processor • Axon : Output • Neurons communicate with one another via synapses which can be excitatory or inhibitory. What is a neuron ? http://www.urbanchildinstitute.org/sites/all/files/databooks/2011/ch1-fg2-communication-between-neurons.jpg
  • 3. Neural Network • A neural network is an interconnected web of neurons transmitting elaborate patterns of electrical signals. • Dendrites receive input signals and, based on those inputs, fire an output signal via an axon. • A neural network is a adaptive system, meaning it can change its internal structure based on the information flowing through it. • Typically, this is achieved through the adjusting of weights. http://www.fridayfonts.com/wp-content/uploads/2009/05/neurons.jpg
  • 4. Artificial neuron • An artificial neuron is a mathematical model of biological neurons which receives one or more inputs and sums them to produce an output. • Usually the sums of each node are weighted, and the sum is passed through a non- linear function known as an activation function. http://upload.wikimedia.org/wikipedia/commons/thumb/6/60/ArtificialNeuronModel_english.png/600px-ArtificialNeuronModel_english.png
  • 5. Artificial Neural Network (ANN) • An ANN is typically defined by three types of parameters: • The interconnection pattern between the different layers of neurons • The learning process for updating the weights of the interconnections • The activation function that converts a neuron's weighted input to its output http://wp.fobiss.com/wp-content/uploads/2013/06/Brain2.png
  • 7. Perceptron • Invented in 1957 by Frank Rosenblatt at the Cornell Aeronautical Laboratory. • It is a computational model of a single neuron. • It consists of one or more binary inputs, a processor, and a single binary output. • Binary inputs are multiplied by weights [real numbers expressing the importance of the respective inputs to the output] & then fed into processor. Activation function http://neuralnetworksanddeeplearning.com/chap1.html
  • 8. Stochastic Model 𝑝 𝑥 = 1 1 + 𝑒− 𝑣 𝑇 𝑊𝑖𝑡ℎ 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 1 − 𝑝(𝑥) 𝑊𝑖𝑡ℎ 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑝(𝑥) Where T is pseudo temperature. • For T=0 :
  • 9. • Sigmoid neuron inputs and output can take any value between 0 and 1. • Sigmoid neuron activation function is given as : 𝜎 𝑤. 𝑥 + 𝑏 = 1 1+𝑒(−𝑤.𝑥−𝑏) Sigmoidal Neuron • The smoothness of σ allows us to write a small change Δoutput in terms of small changes Δwj in the weights and Δb in the bias as : http://neuralnetworksanddeeplearning.com/chap1.html
  • 10. Neural network architecture • The leftmost layer in this network is called the input layer, and the neurons within the layer are called input neurons. The rightmost or output layer contains the output neurons. The middle layer is called a hidden layer, since the neurons in this layer are neither inputs nor outputs. Feed-forward network http://neuralnetworksanddeeplearning.com/chap1.html
  • 11. Design of Input and Output Layers • The design of the input and output layers in a network is often straightforward. • For example, suppose we're trying to determine whether a handwritten image depicts a "9" or not. • A natural way to design the network is to encode the intensities of the image pixels into the input neurons. • If the image is a 64 by 64 greyscale image, then we'd have 4,096=64×64 input neurons, with the intensities scaled appropriately between 0 and 1. • The output layer will contain just a single neuron, with output values of less than 0.5 indicating "input image is not a 9", and values greater than 0.5 indicating "input image is a 9 ".
  • 12. Learning strategy • Now we want an algorithm which lets us find weights and biases so that the output from the network approximates the desired output for all training inputs x. For that purpose we define a cost function : 𝐶 𝑤, 𝑏 = 1 2𝑛 𝑥 𝑑 − 𝑦(𝑥) 2 • The aim of our training algorithm is to find a set of weights and biases which minimizes the cost function. • We'll do that using an algorithm known as gradient descent. Where, w : the collection of all weights in the network b : all the biases n : total number of training inputs, d : desired output from network y(x) : Actual output from network
  • 13. Gradient Descent Algorithm • Let's suppose we're trying to minimize some function : 𝐶 𝒗 𝑤ℎ𝑒𝑟𝑒, 𝒗 = (𝑣1, 𝑣2, … . , 𝑣 𝑚) 𝑇 𝑣1, 𝑣2, … . , 𝑣 𝑚 ∈ ℝ • To visualize C(v) it helps to imagine C as a function of just two variables, which we'll call v1 and v2 What we'd like is to find where C achieves its global minimum. http://neuralnetworksanddeeplearning.com/chap1.html
  • 14. • A small change in 𝐶(𝒗) can be written as : ∆𝐶 = 𝜕𝐶 𝜕𝑣1 ∆𝑣1 + 𝜕𝐶 𝜕𝑣2 ∆𝑣2 + ⋯ + 𝜕𝐶 𝜕𝑣 𝑚 ∆𝑣 𝑚 • Above expression can also be written as : ∆𝐶 = 𝜵𝑪. ∆𝒗 • Where , 𝜵𝑪 = 𝜕𝐶 𝜕𝑣1 , 𝜕𝐶 𝜕𝑣2 , … , 𝜕𝐶 𝜕𝑣 𝑚 𝑇 ∆𝒗 = ∆𝑣1, ∆𝑣2, … , ∆𝑣 𝑚 • Now ,we want to choose ∆𝒗 such that ∆𝐶 ≤ 0. If we choose : ∆𝒗 = −𝜂𝜵𝑪 ⇒ ∆𝐶 = −𝜂 𝜵𝑪 2 ≤ 0
  • 15. • Now , the next point on the surface can be given as : 𝒗′ = 𝒗 + ∆𝑣 𝒗′ = 𝒗 − 𝜂𝜵𝑪 • We'll use above updating rule to compute a value for new value of 𝑣1, 𝑣2,…, 𝑣 𝑚 : 𝑣1 ′ = 𝑣1 − 𝜂 𝜕𝐶 𝜕𝑣1 𝑣2 ′ = 𝑣2 − 𝜂 𝜕𝐶 𝜕𝑣2 . . 𝑣 𝑚 ′ = 𝑣 𝑚 − 𝜂 𝜕𝐶 𝜕𝑣 𝑚 • Then we'll use this update rule again, to make another move. If we keep doing this, over and over, we'll keep decreasing C until , we hope , we reach a global minimum. 𝜂 = 𝐿𝑒𝑎𝑟𝑛𝑖𝑛𝑔 𝑟𝑎𝑡𝑒
  • 16. Update rule for weights & bias • In case of a neural network our variables are : 𝑤𝑖 𝑎𝑛𝑑 𝑏𝑖 • Using gradient descent algorithm, the update rule for weights and biases can be given as : 𝑤𝑖 ′ = 𝑤𝑖 − 𝜂 𝜕𝐶 𝜕𝑤𝑖 𝑏𝑖 ′ = 𝑏𝑖 − 𝜂 𝜕𝐶 𝜕𝑏𝑖 • By repeatedly applying this update rule we can "roll down the hill", and hopefully find a minimum of the cost function. In other words, this is a rule which can be used to learn in a neural network.
  • 17. Practical applications • To recognize individual digits we will use a three-layer neural network: http://neuralnetworksanddeeplearning.com/chap1.html
  • 18. Fitting Straight to the data • A linear model of neuron is used for fitting straight line to our data. • In case of N dimension , there are N free parameters : (N-1) slope components and an intercept. 𝑥1 𝑥 𝑛−1 𝑦𝑛 𝑤 𝑛0 𝑤 𝑛1 𝑤 𝑛(𝑛−1) +1 𝑦𝑛 = 𝑤 𝑛0 + 𝑗=1 𝑛−1 𝑥𝑗 𝑤 𝑛𝑗 Output Input 𝑦𝑝 𝑡 𝑝 𝑝 𝐶 = 𝑝 𝐶 𝑝 𝐶 𝑝 = 1 2 𝑡 𝑝 − 𝑦𝑝 2
  • 19. Update rule for weights • Using gradient descent algorithm, the update rule for weights and biases can be given as : 𝑤 𝑛𝑖 ′ = 𝑤 𝑛𝑖 − 𝜂 𝜕𝐶 𝜕𝑤 𝑛𝑖 • Now , the derivative in above equation can be calculated using chain rule as follows : 𝜕𝐶 𝜕𝑤 𝑛𝑖 = 𝜕 𝜕𝑤 𝑛𝑖 𝑝 𝐶 𝑝 = 𝑝 𝜕𝐶 𝑝 𝜕𝑤 𝑛𝑖 𝜕𝐶 𝑝 𝜕𝑤 𝑛𝑖 = 𝜕𝐶 𝑝 𝜕𝑦𝑝 𝜕𝑦𝑝 𝜕𝑤 𝑛𝑖 𝜕𝐶 𝑝 𝜕𝑦𝑝 = 𝜕 𝜕𝑦𝑝 1 2 𝑡 𝑝 − 𝑦𝑝 2 = − 𝑡 𝑝 − 𝑦𝑝 𝜕𝑦𝑝 𝜕𝑤 𝑛𝑖 = 𝜕 𝜕𝑤 𝑛𝑖 𝑗=0 𝑛 𝑥𝑗 𝑤 𝑛𝑗 = 𝑥𝑖
  • 20. • Using gradient descent algorithm, the update rule for weights and biases can be given as : 𝑤 𝑛𝑖 ′ = 𝑤 𝑛𝑖 − 𝜂 𝜕𝐶 𝜕𝑤 𝑛𝑖 𝜕𝐶 𝜕𝑤 𝑛𝑖 = 𝑝 −𝑥𝑖 𝑡 𝑝 − 𝑦𝑝 𝑤 𝑛𝑖 ′ = 𝑤 𝑛𝑖 + 𝜂 𝑝 𝑥𝑖 𝑡 𝑝 − 𝑦𝑝