Visualization and self-organization of multidimensional data through equalized orthogonal mapping
First Claim
1. A system for organizing multi-dimensional pattern data into a reduced-dimension representation comprising:
- a neural network comprised of a plurality of layers of nodes, the plurality of layers including;
an input layer comprised of a plurality of input nodes, a hidden layer, and an output layer comprised of a plurality of non-linear output nodes, wherein the number of non-linear output nodes is less than the number of input nodes;
receiving means for receiving multi-dimensional pattern data into the input layer of the neural network;
output means for generating an output signal for each of the output nodes of the output layer of the neural network corresponding to received multi-dimensional pattern data; and
training means for completing a training of the neural network, wherein the training means includes means for equalizing and orthogonalizing the output signals of the output nodes by reducing a covariance matrix of the output signals to the form of a diagonal matrix.
3 Assignments
0 Petitions
Accused Products
Abstract
The subject system provides reduced-dimension mapping of pattern data. Mapping is applied through conventional single-hidden-layer feed-forward neural network with non-linear neurons. According to one aspect of the present invention, the system functions to equalize and orthogonalize lower dimensional output signals by reducing the covariance matrix of the output signals to the form of a diagonal matrix or constant times the identity matrix. The present invention allows for visualization of large bodies of complex multidimensional data in a relatively “topologically correct” low-dimension approximation, to reduce randomness associated with other methods of similar purposes, and to keep the mapping computationally efficient at the same time.
44 Citations
16 Claims
-
1. A system for organizing multi-dimensional pattern data into a reduced-dimension representation comprising:
-
a neural network comprised of a plurality of layers of nodes, the plurality of layers including;
an input layer comprised of a plurality of input nodes, a hidden layer, and an output layer comprised of a plurality of non-linear output nodes, wherein the number of non-linear output nodes is less than the number of input nodes;
receiving means for receiving multi-dimensional pattern data into the input layer of the neural network;
output means for generating an output signal for each of the output nodes of the output layer of the neural network corresponding to received multi-dimensional pattern data; and
training means for completing a training of the neural network, wherein the training means includes means for equalizing and orthogonalizing the output signals of the output nodes by reducing a covariance matrix of the output signals to the form of a diagonal matrix. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
and the elements of the covariance matrix of the output signals of the output nodes are defined by;
where p=1, 2, . . . , P;
Ok 1 p is the output signal of the k1th node of the output layer for the pth input data pattern vector;
Ok 2 p is the output signal of the k2th node of the output layer for the pth input data pattern vector;
<
Ok1 >
is the average of Ok1 p evaluated over the set of input data pattern vectors<
Ok2 >
is the average of Ok2 p evaluated over the set of input data pattern vectorsk1=1 to K;
k2=1 to K;
K is the number of dimensions in the reduced-dimension representation; and
<
>
denotes the mean evaluated over the set of input data pattern vectors for each indicated component.
-
-
5. A system according to claim 4, wherein weights Δ
- wkj between the hidden layer and the output layer are iteratively updated according to the expression;
where η
is a constant of suitable value chosen to provide efficient convergence but to avoid oscillation;Ojp is the output signal from the jth node in the layer preceeding the output layer due to the pth input data pattern vector;
E is the error given by;
where k1=k2=k;
k=1, . . . , K; and
rkk is a positive constant which has an effect of increasing the speed of training,where k2>
k1;
k1=1, . . . , K; and
rk1 k2 is a positive constant which has an effect of increasing the speed of training; andδ
kp=δ
kp,1+δ
kp,2+δ
kp,3, where δ
kp is a value proportional to the contribution to the error E by the outputs of the kth node of the output layer, for the pth input data pattern vector, and δ
kp,1, δ
kp,2, and δ
kp,3 are components of δ
kp.
- wkj between the hidden layer and the output layer are iteratively updated according to the expression;
-
6. A system according to claim 5, wherein:
-
where Δ
wkj,1 is the contribution from the diagonal terms of the covariance matrix of the outputs,Δ
wkj,2 is the contribution from the off-diagonal terms in kth row,Δ
wkj,3 is the contribution from the off-diagonal terms in kth column, andOjp is the output signal from the jth node in the layer preceeding the output layer for the pth input data pattern vector.
-
-
7. A system according to claim 6, wherein:
-
where Okp is the output signal from the kth node in the output layer for the pth input data pattern vector, and <
Okp>
is the average of Okp evaluated over the set of input data pattern vectors.
-
-
8. A system according to claim 5, wherein backpropogation of error to the weights Δ
- wji between the jth node in a layer of nodes and the ith node in its'"'"' preceeding layer;
where, δ
jp is given by;
- wji between the jth node in a layer of nodes and the ith node in its'"'"' preceeding layer;
-
9. A method for effecting the organization of multi-dimensional pattern data into a reduced dimensional representation using a neural network having an input layer comprised of a plurality of input nodes, a hidden layer, and an output layer comprised of a plurality of non-linear output nodes, wherein the number of non-linear output nodes is less than the number of input nodes, said method comprising:
-
receiving multi-dimensional pattern data into the input layer of the neural network;
generating an output signal for each of the output nodes of the neural network corresponding to received multi-dimensional pattern data; and
training the neural network by equalizing and orthogonalizing the output signals of the output nodes by reducing a covariance matrix of the output signals to the form of a diagonal matrix. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16)
and the elements of the covariance matrix of the output signals of the output nodes is;
where p=1, 2, . . . , P;
Ok 1 p is the output signal of the k1th node of the output layer for the pth input data pattern vector;
Ok 2 p is the output signal of the k2th node of the output layer for the pth input data pattern vector;
<
Ok1 p>
is the average of Ok1 p evaluated over the set of input data pattern vectors<
Ok2 p>
is the average of Ok2 p evaluated over the set of input data pattern vectorsk1=1 to K;
k2=1 to K;
K is the number of dimensions in the reduced-dimension representation; and
<
>
denotes the mean evaluated over the set of input data pattern vectors for each indicated component.
-
-
13. A method according to claim 12, wherein weights Δ
- wkj between the hidden layer and the output layer are iteratively updated according to the expression;
where η
is a constant of suitable value chosen to provide efficient convergence but to avoid oscillation;Ojp is the output signal from the jth node in the layer preceeding the output layer, due to the pth input data pattern vector;
E is the error given by;
where k1=k2=k;
k=1, . . . , K; and
rkk is a positive constant which has an effect of increasing the speed of training,where k2>
k1;
k1=1, . . . , K−
1;
k2=k1+1, . . . , K; and
rk1 k2 is a positive constant which has an effect of increasing the speed of training; andδ
kp=δ
kp,1+δ
kp,2+δ
kp,3, where δ
kp is a value proportional to the contribution to the error E by the outputs of the kth node of the output layer, for the pth input data pattern vector, and δ
kp,1, δ
kp,2, and δ
kp,3 are components of δ
kp.
- wkj between the hidden layer and the output layer are iteratively updated according to the expression;
-
14. A method according to claim 13, wherein:
-
where Δ
wkj,1 is the contribution from the diagonal term,Δ
wkj,2 is the contribution from the off-diagonal terms in kth row, andΔ
wkj,3 is the contribution from the off-diagonal terms in kth column.
-
-
15. A method according to claim 14, wherein δ
-
kp,1, δ
kp,2 and δ
kp,3 are given by;
where Okp is the output signal from the kth node in the layer preceeding the output layer for the pth input data pattern vector, and <
Okp>
is the average of Okp evaluated over the set of input data pattern vectors.
-
kp,1, δ
-
16. A method according to claim 13, wherein backpropogation of error to the weights Δ
- wji between the jth node in a layer of nodes and the ith node in its'"'"' preceeding layer are;
where, δ
jp is given by;
- wji between the jth node in a layer of nodes and the ith node in its'"'"' preceeding layer are;
Specification