This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 1
Neural Networks
• Outline– Introduction
– From biological to artificial neurons
– Self organizing maps
– Backpropagation network
– Radial basis functions
– Associative memories Hopfield networks
– Other applications of neural networks
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 2
Introduction
• Why neural networks?– Algorithms developed over centuries do not fit the complexity of
real world problem
– The human brain: most sophisticated computer suitable for solving extremely complex problems
• Historical knowledge on human brain– Greeks thought that the brain was where the blood is cooled off!
– Even till late 19th century not much was known about the brain and it was assumed to be a continuum of non-structured cells
– Phineas Gage’s Story• In a rail accident, a metal bar was shot
through the head of Mr. Phineas P.
Gage at Cavendish, Vermont, Sept 14, 1848 – Iron bar was 3 feet 7 inches long and weighed 13 1/2 pounds. It was 1 1/4
inches in diameter at one end
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 3
Introduction (cont’d)
• He survived the accident!– Originally he seemed to have fully recovered with no clear effect(s)
• After a few weeks, Phineas exhibited profound personality changes
– This is the first time, researchers have a clear evidence that the brain is not a continuum of cell mass and rather each region has relatively independent task
• information is saved mainly in the connections among neurons
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 5
• learning and generalization through examples
• simple building block: neuron– Dendrites: collecting
signals from other neurons
– Soma (cell body): spatial
summation and processing
– Axon: transmitting signals to
dendrites of other cells
Introduction (Continued)
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 6
Introduction (Continued)
• biological neural networks:
formation of neurons with different connection strengths
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 7
• Biological vs. artificial neurons– From biological neuron to schematic structure of artificial
neuron• biological:
– Inputs
– Summation
of inputs
– Processing
unit
– Output
• artificial:
From biological to artificial neural nets
1x
Nx
1w
Nw
)...( 11 nn xwxwfy
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 8
– Artificial neural nets:• Formation of artificial neurons
From biological to artificial neural net (continued)
neuron 1
1x
Nx
NMw
neuron 2
neuron M
iw1
11w
Niw
neuron i
neuron M-1
1y
2y
iy
1My
My
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 9
– Multi-layer neural nets:• Serial connection of single layers:
– Training: finding the best values of weights wij
• Training happens iteratively and through exposing the network to examples:
From biological to artificial neural nets (continued)
1x
Nx
NMw
iw1
11w
Niw
1y
2y
iy
1My
My
ijijij www )old()new(
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 10
– Activation functions:
• Hard limiter (binary step):
– Role of threshold
– Biologically supported
– Non-differentiable
From biological to artificial neural nets (continued)
x
x
x
xf
if1
if0
if1
)(
1
1
x
)(xf
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 11
• Binary sigmoid (exponential sigmoid)
– Differentiable
– Biologically supported
– Saturation curve is controlled
by
– In limit when , hard limiter is achieved
• Bipolar sigmoid (atan)
– As popular as binary sigmoid
From biological to artificial neural nets (continued)
)(tan)( 1 xxf
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 12
• Supervised Vs Unsupervised Learning– Supervised learning (classification)
• Training data are labeled, i.e the output class of all training data are given
• Example: recognition of birds and insects– Training set:
– Classification:
– Unsupervised learning (clustering)• Training data are not labeled• Output classes must be generated during training• Similarity between features of training example creates different
classes
birdowlinsectantinsectbeebirdeagle ,,,
?sparrow
From biological to artificial neural nets (continued)
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 13
• Example : types of companies– Features: Number
of employees &
rate of growth
– Training data
create natural
clusters
– From graph:
Class #1: small size companies with small rate of growth
Class #2: small size companies with large rate of growth
Class #3: medium size companies with medium rate of growth
Class #4: large size companies with small rate of growth– Classification: a company with NOE=600 & ROG=12%
is mapped to Class #3
Number of Employees
Rat
e of
Gro
wt h
100 500 1000
5%20
%10
%
From biological to artificial neural nets (continued)
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 14
From biological to artificial neural nets (continued)
– Artificial neural networks as classification tools:• If training data are labeled, supervised neural nets are used• Supervised leaning normally results to better performance• Most successful types of supervised ANNs:
– Multi-layer perceptrons– Radial basis function networks
– Artificial neural networks as clustering tools:• If training data are not labeled, unsupervised neural nets are used• Most successful types of supervised ANNs:
– Kohonen network
– Some networks can be trained both in supervised and unsupervised modes
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 15
Perceptron
• A more advanced version of
simple neuron
• Structure (architecture):– Very similar to simple neuron
– The only difference:
activation function is
bipolar hard limiter
iny
iny
iny
y
_if1
_if0
_if1
1
1
iny
y
1x
ix
nx
n
iiij wxbiny
1
_
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 16
Competitive Self-Organizing Networks
• Biological procedure:– Each neuron (group of neurons)
stores a pattern– Neurons in a neighborhood store
similar patterns– During classification the similarity
of new pattern with all the patterns
in all neurons is calculated– The neurons (or neighborhoods)
with highest similarity are the winners– The new pattern is attributed to the
class of the winner neighborhood
Rocking Chair
Study Chair Desk
Dining Table
DishesFood
Forest
Books
Trees
Rocking Chair
Study Chair Desk
Dining Table
DishesFood
Forest
Books
Trees
Winner: Desk (and its neighborhood)
New pattern: Conference Table
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 17
Kohonen Self-Organizing Maps
• Main idea: placing similar objects close to each other– Example: object classification using one-dimensional array
– New pattern: hovercraft • is mapped to the nationhood of truck or airplane
weight (0 to 1) texture (wood=0, metal =1)
size (0 to 1) flies (1) or not (0)
pencil
pencil
airplane
airplane
chair
chair
book
book
wooden house
wooden house
stapler
stapler
metal desk
metal desktruck
truck
kite
kite
Inputs
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 18
Kohonen Self-Organizing Maps (continued)
• Two-dimensional arrays– Proximity exists on two dimensions– Better chance of positioning similar
objects in the same vicinity– The most popular
type of Kohonen network– How to define neighborhoods
RectangularHexagonal
• Radius of neighborhood
(R)
• R = 0 means that each
neighborhood has only
one member
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 19
Kohonen Self-Organizing Maps (continued)
• Example: two-dimensional array– Classification of objects
• Objective: clustering of
a number of objects • 10 input features: # of lines,
thickness of lines, angles, ...• Competitive layer is
a 2D-grid of neurons• Trained network clusters
objects rather successfully• Proximity of some objects
is not optimal
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 20
Backpropagation neural networks
• Idea: not as biologically-supported as Kohonen • Architecture:
– Number
of layers &
number of
neurons in
each layer– Most popular structure
• Activation functions: – sigmoid
1x
Nx
NLw
iw1
11w
Niw
1y
2y
iy
1Ly
My
jkvjz
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 21
Backpropagation neural networks (continued)
• Updating weights:– Based on Delta Rule – Function to be minimized:
– Best updating of
wJK is toward
the gradient– Error of output layer is propagated back towards the input layer– Error is calculated at output layer and propagated back towards
the input layer– As layers receive the backpropagated error, they adjust their
weights according to the negative direction of gradient
k
kk ytE 25.0
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 22
Backpropagation neural networks (continued)
– Define:
– Then:
and:
– Now:
)_( KKKK inyfyt
JKJK
zw
E
IJk
JkkIJ
xinzfwv
E_
jkjk
jk zw
Ew
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 23
Backpropagation neural networks (continued)
• Applications:– Time-series analysis and prediction
• Problem statement (simplest case): – The future value of a signal depends on the previous values of
the same signal, i.e.• Objective: to use a neural net to estimate function “g”• Procedure:
– Form a number of training points as:
– Train a backpropagation net to learn the input-output relation• Advanced cases:
– Procedure is similar to the simple case
)(,...),2(),1()( ptytytygty
1....,,1,
)()(,...),2(),1(
pnppi
iypiyiyiy
)(,...),2(),1(),(,...),2(),1()( rtxtxtxptytytygty
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 24
Backpropagation neural networks (continued)
• Feedforward neural time series models are used in many fields including:
– Stock market prediction– Weather prediction– Control– System identification– Signal and image processing– Signal and image compression
• Classification:– When a hard limiter is added to the output neurons
(only in classification and not during the training phase), backpropagation network is used to classify complex data sets
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 25
Radial Basis Function networks (RBFN’s)
• Architecture:– Weights of the input
layer are all “1”, i.e.:
– Based on the exact definition
of radial basis functions ,
many different families of
RBFN’s are defined– Main properties of all “radial” basis functions: “radial”
• Example: is a radial function because the value of
is the same for
all points with equal
distance from point C.
1x
2
1)(
Cxxx
…..
2x
nx
(x)1
(x)2
(x)m
…... y
nxxx ...,,,x 21
(.)i
C
m
iiiy
1
x
1
2
m
1all
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 26
RBFN’s (continued)
• Basis functions– Gaussian basis functions:
• Coordinates of center:• Width parameter:
– Reciprocal Multi-Quadratic (RMQ) functions:
• Width parameter:
• Training: Least Mean Square Techniques
2
2xx
exp(x)i
C
ii
iCx
i
2
Cxx1
1(x)
ii
i
b
ib
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 27
RBFN’s (continued)• Example:
– Trying to approximate a no-linear function using RMQ-RBFN’s– Function to be estimated:
– 135 Training points– 19 Basis functions are
uniformly centered
between -10 and 10– All basis functions have: b = 0.5– Training method: batch– Estimation is rather successful
-10 -8 -6 -4 -2 0 2 4 6 8 10-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
Input
Actu
al and E
stim
ate
d O
uto
ut
RBFNs (batch method)
Red: Actual
Blue: Estimated
xxxg 1ln).sin()(
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 28
Associative Memories
• Concept:– Object or pattern A (input) reminds the network of object or pattern
B (output)
• Heteroassociative Vs. autoassociative memories– If A and B are different, the system is called heteroassociative net
• Example: you see a large lake (A) and that reminds you of the Pacific (B) ocean you visited last year
– If A and B are the same, the system is called autoassociative net• Example: you see the Pacific ocean for the second time (A) and that
reminds you of the Pacific (B) ocean you visited last year
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 29
Associative Memories (continued)
• Recognizing new or incomplete patterns– Recognizing patterns that are similar to one of the patterns stored in
memory (generalization)• Example: recognizing a football player you haven’t seen before from
his clothes
– Recognizing incomplete or noisy patterns whose complete (correct) forms were previously stored in memory
• Example: recognizing somebody’s face from a picture that is partially torn
• Unidirectional Vs. bidirectional memories– Unidirectional: A reminds you of B
– Bidirectional: A reminds you of B and B reminds you of A
• Many biological neural nets are associative memories
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 30
Hopfield Network
• Concept:– A more advance type of autoassociative memory
– Is almost fully connected
• Architecture– Symmetric weights
– No feedback from a
cell to itself
– Notice the “feedback” in
the network structure
jiij ww
0iiw
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 31
• Concept:– Bidirectional memory: pattern A reminds you of pattern B and pattern B
reminds you of pattern A– Is almost fully connected
• Architecture– Symmetric weights
– No feedback from a
cell to itself
– Notice the “feedback” in
the network structure
Bidirectional Associative Memories (BAM)
jiij ww
0iiw
YmYjY1
XnXiX1
… …
……
nmw11wijw
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 32
Other Applications of NNs
• Control– Structure:
– Example: Robotic manipulation
SystemNeuro-controller
Actualbehavior
ControldecisionDesired
behavior
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 33
• Finance and Marketing– Stock market prediction
– Fraud detection
– Loan approval
– Product bundling
– Strategic planning
• Signal and image processing– Signal prediction (e.g. weather prediction)
– Adaptive noise cancellation
– Satellite image analysis
– Multimedia processing
Applications of NNs (continued)
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 34
• Bioinformatics– Functional classification of protein
– Functional classification of genes
– Clustering of genes based on their expression (using DNA microarray data)
• Astronomy– Classification of objects (into stars and galaxies ad so on)
– Compression of astronomical data
• Function estimation
Applications of NNs (continued)
Computer Science Dept, UNC Charlotte
Copyright 2002 Kayvan Najarian 35
• Biomedical engineering– Modeling and control of complex biological system (e.g. modeling
of human respiratory system)
– Automated drug-delivery
– Biomedical image processing and diagnostics
– Treatment planning
• Clustering, classification, and recognition– Handwriting recognition