Scribd
Upload
38 views

Lecture 14 - Associative Neural Networks Using Matlab

The document provides examples of using Matlab code to simulate auto-associative neural networks, hetero-associative neural networks, Hopfield networks, and continuous auto-associative memories. It includes code to calculate weight matrices, test network responses, and store/recall patterns from these different network types. Key functions used include newhop to create a Hopfield network and sim to simulate a neural network over multiple time steps.

Uploaded by

Ammar Alkindy
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
38 views

Lecture 14 - Associative Neural Networks Using Matlab

The document provides examples of using Matlab code to simulate auto-associative neural networks, hetero-associative neural networks, Hopfield networks, and continuous auto-associative memories. It includes code to calculate weight matrices, test network responses, and store/recall patterns from these different network types. Key functions used include newhop to create a Hopfield network and sim to simulate a neural network over multiple time steps.

Uploaded by

Ammar Alkindy
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Dr.

Qadri Hamarsheh

Associative Neural Networks using Matlab

Example 1: Write a matlab program to find the weight matrix of an auto


associative net to store the vector (1 1 -1 -1). Test the response of the network
by presenting the same pattern and recognize whether it is a known vector
or unknown vector.
The auto associative net has the same inputs and targets. The MATLAB program
for the auto associative net is as follows:
Program
%Auotassociative net to store the vector
clc;
clear;
x = [1 1 –1 –1];
w=zeros (4, 4);
w=x'*x;
yin=x*w;
for i=1:4
if yin(i)>0
y(i)=1;
else
y(i) = –1;
end
end
disp ('Weight matrix');
disp (w);
if x == y
disp ('The vector is a Known Vector');
else
disp ('The vector is a Unknown Vector');
end
Output
Weight matrix
1 1 –1 –1
1 1–1 –1
–1 –1 1 1
–1 –1 1 1
The vector is a known vector.

1
Dr. Qadri Hamarsheh

Example 2: Write an M–file to store the vectors (–1 –1 –1 –1) and (–1 –1 1 1) in
an auto associative net. Find the weight matrix. Test the net with (1 1 1 1) as
input.
The MATLAB program for the auto association problem is as follows:
Program
clc;
clear;
x=[–1 –1 –1 –1;–1 –1 1 1];
t=[1 1 1 1];
w=zeros (4, 4);
for i=1:2
w=w + x(i,1:4)'*x(i,1:4);
end
yin = t*w;
for i=1:4
if yin(i)>0
y(i)=1;
else
y(i)=–1;
end
end
disp ('The calculated weight matrix');
disp (w);
if x(1,1:4)==y(1:4) | x(2,1:4)==y(1:4)
disp ('The vector is a Known Vector');
else
disp ('The vector is a unknown vector');
end
Output
The calculated weight matrix
2 2 0 0
2 2 0 0
0 0 2 2
0 0 2 2
The vector is an unknown vector.
Example 3: Write an M–file to calculate the weights for the following
patterns using hetero associative neural net for mapping four input vectors
to two output vectors.
S1 S2 S3 S4 t1 t2
1 1 0 0 1 0
1 0 1 0 1 0
1 1 1 0 0 1
0 1 1 0 0 1

2
Dr. Qadri Hamarsheh

Solution
%Hetero associative NN for mapping input vectors to output vectors
clc;
clear;
x= [1 1 0 0; 1 0 1 0; 1 1 1 0; 0 1 1 0];
t= [1 0; 1 0; 0 1; 0 1];
w=zeros (4, 2);
for i=1:4
w=w+x (i, 1:4)'*t(i,1:2);
end
disp('Weight matrix');
disp(w);
Output
Weight matrix
2 1
1 2
1 2
0 0
Example 4: Write a MATLAB program to store the vector (1 1 1 0) in bipolar
binary form and calculate the weight matrix for Hopfield net.
%The MATLAB program for calculating the weight matrix is as follows
%Discrete Hopfield net
clc;
clear;
x=[1 1 1 0];
w=(2*x'–1)*(2*x–1);
for i=1:4
w (i, i)=0;
end
disp('Weight matrix');
disp(w);
Output
Weight matrix
0 1 1 -1
1 0 1 -1
1 1 0 -1
-1 -1 -1 0
Example 5: Auto-associative Memories (continuous):
%Auto-associative Memories (continuous)
x1=[-0.3; 0.9; -0.2];
x2=[0.44; -0.7; 0.9];
x3=[0.9; 0.6; 0.8];
Total_M = x1*x1' + x2*x2' + x3*x3';
estimate_x1= Total_M *x1;
estimate_x2= Total_M *x2;
estimate_x3= Total_M *x3;
%Estimates are not perfect because of non-orthogonality of the vectors

3
Dr. Qadri Hamarsheh

%Euclidean distance between x1 and other key vectors


d11 = norm(x1-estimate_x1);
d21 = norm(x2-estimate_x1);
d31 = norm(x3-estimate_x1);
% As expected the response vector estimate_x1 is closest to x1
%Euclidean distance between x2 and other key vectors
d12 = norm(x1-estimate_x2);
d22 = norm(x2-estimate_x2);
d32 = norm(x3-estimate_x2);
%As expected the response vector estimate_x2 is closest to x2
%Euclidean distance between x3 and other key vectors
d13 = norm(x1-estimate_x3);
d23 = norm(x2-estimate_x3);
d33 = norm(x3-estimate_x3);
%As expected the response vector estimate_x3 is closest to x3
 Matlab code for converting continuous data (x1 vector) to binary bipolar
data.
for i=1:length(x1)
if x1(i)>0
x1(i) = 1;
else
x1(i) = -1;
end
end
 In Matlab Hopfield networks can be implemented as vector matrix
manipulations. To make the pattern vectors as easy as possible to read and
write we define them as row vectors.
 We prefer to make the calculations within the interval [−1, 1] (bipolar) as
this makes the calculations simpler. It is, however, easier to type in and to
visually recognize values in the range [0, 1] (binary). Therefore, it may be
better to use this for input and output. Translate a vector from the binary
format into the bipolar.
Example 6:
%Enter three test patterns.
x1b= [0 0 1 0 1 0 0 1];
x2b= [0 0 0 0 0 1 0 0];
x3b= [0 1 1 0 0 1 0 1];
%Translate a vector from the binary format into the bipolar.
%x1= [-1 -1 1 -1 1 -1 -1 1];
%x2= [-1 -1 -1 -1 -1 1 -1 -1];
%x3=[-1 1 1 -1 -1 1 -1 1];
x1 = 2* x1b -1;
x2 = 2* x2b -1;
x3 = 2* x3b -1;
%Calculate a weight matrix.
w=x1'*x1+x2'*x2+x3'*x3-3*eye (8, 8);
%Check if the network was able to store all three patterns.

4
Dr. Qadri Hamarsheh

x1test=sign (w*x1');
x2test=sign (w*x2');
x3test=sign (w*x3');
%Convergence and attractors.
%Can the memory recall the stored patterns from distorted inputs
% patterns? Define a few new patters which are distorted versions of
%the original ones:
x1d= [1 0 1 0 1 0 0 1];
x2d= [1 1 0 0 0 1 0 0];
x3d= [1 1 1 0 1 1 0 1];
%x1d has a one bit error, x2d and x3d have two bit errors.
x1d=[1 -1 1 -1 1 -1 -1 1];
x2d=[1 1 -1 -1 -1 1 -1 -1];
x3d=[1 1 1 -1 1 1 -1 1];

Hopfield neural networks using Matlab Neural Network Tool Box


 Hopfield neural networks can be simulated by using the Neural Network
Tool Box. The architecture is shown below.

 net = newhop (T) takes one input argument:


o T - R x Q matrix of Q target vectors. (Values must be +1 or -1) and
returns a new Hopfield recurrent neural network with stable points at
the vectors in T.
o Hopfield networks consist of a single layer with the dotprod weight
function, netsum net input function, and the satlins transfer function.
Example 7:
 Consider the following design example. Suppose that we want to design a
network with two stable points in a three-dimensional space T.
T = [-1 -1 1; 1 -1 1]';
 We can execute the design with:
net = newhop (T);
 To check that the network is stable at these points use them as initial layer
delay conditions. If the network is stable we would expect that the outputs
5
Dr. Qadri Hamarsheh

Y will be the same. (Since Hopfield networks have no inputs, the second
argument to sim is Q = 2 when using matrix notation).
Ai = T;
[Y, Pf, Af] = sim (net, 2, [ ], Ai);
%Y
 To see if the network can correct a corrupted vector, run the following
code, which simulates the Hopfield network for five time steps. (Since
Hopfield networks have no inputs, the second argument to sim is {Q TS} =
[1 5] when using cell array notation.)
Ai = {[-0.9; -0.8; 0.7]};
[Y, Pf, Af] = sim (net, {1 5}, { },Ai);
%Y{1}
 If you run the above code, Y{1} will equal T(:,1) if the network has
managed to convert the corrupted vector Ai to the nearest target vector.
Description of sim function
Purpose: simulate a neural network.
Syntax:
[Y, Pf, Af, E, perf] = sim (net, P, Pi, Ai, T)
[Y, Pf, Af, E, perf] = sim (net, {Q TS}, Pi, Ai, T)
[Y,Pf,Af,E,perf] = sim(net,Q,Pi,Ai,T)
Description sim simulates neural networks.
[Y, Pf, Af, E, perf] = sim (net, P, Pi, Ai, T) takes,
net - Network.
P - Network inputs.
Pi - Initial input delay conditions, default = zeros.
Ai - Initial layer delay conditions, default = zeros.
T - Network targets, default = zeros.
and returns,
Y - Network outputs.
Pf - Final input delay conditions.
Af - Final layer delay conditions.
E - Network errors.
perf- Network performance.
Note that arguments Pi, Ai, Pf, and Af are optional and need only be used for
networks that have input or layer delays.
sim’s signal arguments can have two formats: cell array or matrix.

You might also like