Bibliographic Details
| Title: |
DEVICE FOR AND COMPUTER IMPLEMENTED METHOD OF DIGITAL SIGNAL PROCESSING |
| Document Number: |
20210366160 |
| Publication Date: |
November 25, 2021 |
| Appl. No: |
17/242710 |
| Application Filed: |
April 28, 2021 |
| Abstract: |
A device for and a computer implemented method of digital signal processing. The method includes providing a first set of data, mapping the first set of data with to a second set of data, and determining an output of the digital signal processing depending on the second set of data. The second set of data is determined depending on a sum of a finite series of terms. At least one term of the series is determined depending on a result of a convolution of the first set of data with a kernel and at least one term of the series is determined depending on the first set of data and independent of the kernel. |
| Claim: |
1. A computer implemented method for digital signal processing, the method comprising the following steps: providing a first set of data; mapping the first set of data to a second set of data; and determining an output of the digital signal processing depending on the second set of data; wherein the second set of data is determined depending on a sum of a finite series of terms, wherein at least one term of the finite series is determined depending on a result of a convolution of the first set of data with a kernel and at least one term of the finite series is determined depending on the first set of data and independent of the kernel, at least one term of the finite series is determined depending on a first result, and wherein the first result is determined depending on a first convolution of the first set of data with the kernel and wherein the at least one term is determined depending on a second convolution of the first result with the kernel. |
| Claim: |
2. The method according to claim 1, wherein at least one term of the finite series that is determined depending on the kernel is determined depending on a convolution of the kernel with an output of an autoregressive function defined by the first set of data. |
| Claim: |
3. The method according to claim 1, wherein the first set of data represents an input at a hidden layer of an artificial neural network between an input layer and an output layer, wherein the second set of data represents an output of the artificial neural network, and wherein the input and the output have the same dimension. |
| Claim: |
4. The method according to claim 1, wherein training input data representing a digital signal is mapped to training output data, wherein the training output data is determined depending on a sum of a series of terms, wherein at least one term of the series of terms is determined depending on a result of a convolution of the training input data with the kernel and at least one term of the series of terms is determined depending on the training input data and independent of the kernel, and wherein at least one parameter of an artificial neural network is determined depending on the training output data. |
| Claim: |
5. The method according to claim 1, characterized by a multiplication of elements of the kernel by a negative value. |
| Claim: |
6. The method according to claim 1, wherein the negative value is −1. |
| Claim: |
7. The method according to claim 4, wherein, in training, the sum of the series of terms is determined at a hidden layer of an artificial neural network arranged between an input layer of the artificial neural network for the training input data and an output layer of the artificial neural network for the training output data. |
| Claim: |
8. The method according to claim 1, wherein an encoder is defined depending on the kernel, wherein a decoder is defined depending on the kernel, wherein a training of the encoder and/or the decoder includes mapping a digital signal with the encoder to a representation thereof, and mapping the representation with the decoder to a synthetic signal. |
| Claim: |
9. The method according to claim 1, characterized in that the at least one term of the series, that is determined depending on a number of convolutions, is determined depending on a result of a division by a factorial, wherein the factorial is defined by the product of positive integers less than or equal to the number. |
| Claim: |
10. The method according to claim 1, further comprising: determining the first set of data depending on a digital signal, wherein a representation of the digital signal is determined depending on the second set of data, or by determining the first set of data depending on a representation of a digital signal; wherein a synthetic sample of the digital signal is determined depending on the second set of data. |
| Claim: |
11. The method according to claim 10, further comprising: determining a density depending on the second set of data, wherein the output of the digital signal processing indicates whether the density meets a condition or not. |
| Claim: |
12. The method according to claim 1, further comprising: determining a plurality of sums of terms depending on a plurality of kernels. |
| Claim: |
13. The method according to claim 12, further comprising: determining the plurality of sums of terms depending on kernels that differ from each other. |
| Claim: |
14. The method according to claim 1, wherein the first set of data is sampled from a random distribution having a density or the first set of data is determined depending on an autoregressive function from input data. |
| Claim: |
15. The method according to claim 1, wherein the kernel defines a square matrix for a matrix multiplication with an input that is equivalent to the convolution of the input with the kernel, wherein a spectral norm of the matrix is less than 1. |
| Claim: |
16. A device for digital signal processing, the device configured to: providing a first set of data; mapping the first set of data to a second set of data; and determining an output of the digital signal processing depending on the second set of data; wherein the second set of data is determined depending on a sum of a finite series of terms, wherein at least one term of the finite series is determined depending on a result of a convolution of the first set of data with a kernel and at least one term of the finite series is determined depending on the first set of data and independent of the kernel, at least one term of the finite series is determined depending on a first result, and wherein the first result is determined depending on a first convolution of the first set of data with the kernel and wherein the at least one term is determined depending on a second convolution of the first result with the kernel. |
| Claim: |
17. A non-transitory computer-readable storage medium on which is stored a computer program for digital signal processing, the computer program, when executed by a computer, cause the computer to perform the following steps: providing a first set of data; mapping the first set of data to a second set of data; and determining an output of the digital signal processing depending on the second set of data; wherein the second set of data is determined depending on a sum of a finite series of terms, wherein at least one term of the finite series is determined depending on a result of a convolution of the first set of data with a kernel and at least one term of the finite series is determined depending on the first set of data and independent of the kernel, at least one term of the finite series is determined depending on a first result, and wherein the first result is determined depending on a first convolution of the first set of data with the kernel and wherein the at least one term is determined depending on a second convolution of the first result with the kernel. |
| Current International Class: |
06; 06; 06 |
| Accession Number: |
edspap.20210366160 |
| Database: |
USPTO Patent Applications |