Natural language processing via a two-dimensional symbol having multiple ideograms contained therein

Patent metadata

Field	Value
Publication number	US-10102453-B1
Application number	US-201715694711-A
Country	US
Kind code	B1
Filing date	Sep 1, 2017
Priority date	Aug 3, 2017
Publication date	Oct 16, 2018
Grant date	Oct 16, 2018

Abstract

A string of natural language texts is received and formed a multi-layer 2-D symbol in a first computing system. The 2-D symbol comprises a matrix of N×N pixels of data representing a “super-character”. The matrix is divided into M×M sub-matrices with each sub-matrix containing (N/M)×(N/M) pixels. N and M are positive integers, and N is preferably a multiple of M. Each sub-matrix represents one ideogram defined in an ideogram collection set. “Super-character” represents a meaning formed from a specific combination of a plurality of ideograms. The meaning of the “super-character” is learned in a second computing system by using an image processing technique to classify the 2-D symbol, which is formed in the first computing system and transmitted to the second computing system. Image process technique includes predefining a set of categories and determining a probability for associating each of the predefined categories with the meaning of the “super-character”.

First claim

What is claimed is:

1. A method of machine learning of written natural languages comprising:
receiving a string of natural language texts in a first computing system having at least one application module installed thereon;
forming, with the at least one application module in the first computing system, a multi-layer two-dimensional (2-D) symbol from the received string of natural language texts based on a set of rules, the 2-D symbol being a matrix of N×N pixels of data that contains a super-character, the matrix being divided into M×M sub-matrices with each of the sub-matrices containing (N/M)×(N/M) pixels, said each of the sub-matrices representing one ideogram defined in an ideogram collection set, and the super-character representing a meaning formed from a specific combination of a plurality of ideograms, where N and M are positive integers, and N is a multiple of M; and
learning the meaning of the super-character in a second computing system by using an image processing technique to classify the 2-D symbol, which is formed with the at least one application module in the first computing system and transmitted to the second computing system.

2. The method of claim 1 , wherein the super-character contains a maximum of M×M ideograms.

3. The method of claim 1 , wherein the set of rules comprises:
determining a size of the received string of natural language texts;
if the size of the receiving string is greater than M×M, modifying the received string via at least one language text reduction scheme such that the size of the string is equal to M×M, and then converting the string to the multi-layer 2-D symbol;
if the size of the receiving string is equal to M×M, converting the received string to the multi-layer 2-D symbol;
if the size of the receiving string is less than M×M and a padding operation is desired, adding at least one text to pad the string such that the size of the string is equal to M×M in accordance with at least one language text increase scheme, and then converting the padded string to the multi-layer 2-D symbol;
otherwise, converting the received string to the multi-layer 2-D symbol, which contains at least one empty space.

4. The method of claim 3 , wherein said at least one language text reduction scheme comprises deleting at least one unimportant text from the received string according to at least one relevant grammar based rule.

5. The method of claim 4 , wherein the at least one relevant grammar based rule is associated with the received string of natural language texts.

6. The method of claim 4 , wherein said at least one language text reduction scheme comprises a randomized text reduction scheme.

7. The method of claim 6 , wherein the randomized text reduction scheme comprises truncating the string such that the size of the string is reduced to M×M.

8. The method of claim 6 , wherein the randomized text reduction scheme comprises arbitrarily selecting certain texts in the string such that the size of the string is equal to M×M.

9. The method of claim 1 , wherein at least one language text increase scheme comprises repeatedly appending one or more texts from the received string to the string.

10. The method of claim 1 , wherein the multi-layer 2-D symbol contains three layers for representing red, green and blue hues, respectively.

11. The method of claim 1 , said using the image processing technique further comprises predefining a set of categories.

12. The method of claim 11 , said using the image processing technique further comprises determining a probability for associating each of the predefined categories with the meaning of the super-character.

13. The method of claim 12 , wherein the set of categories comprises commands for smart electronic devices.

14. The method of claim 1 , wherein the image processing technique comprises a convolutional neural networks algorithm.

15. The method of claim 1 , wherein the image processing technique comprises a support vector machine.

16. The method of claim 1 , wherein the second computing system comprises a Cellular Neural Networks or Cellular Nonlinear Networks (CNN) based computing system.

17. The method of claim 16 , wherein CNN based computing system comprises a semi-conductor chip containing digital circuits dedicated for performing the convolutional neural networks algorithm.

18. The method of claim 1 , wherein N is 224, M is 4, M×M is 16 and N/M is 56.

19. The method of claim 1 , wherein N is 224, M is 8, M×M is 64 and N/M is 28.

Assignees

• Gyrfalcon Tech Inc

Inventors

• Yang Lin
• Dong Patrick Z
• Sun Baohua

Classifications

• G06V30/19173

  Classification techniques · CPC title

• G06V10/82 Primary

  using neural networks · CPC title

• G06F18/2415

  based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate · CPC title

• G06F18/24133

  Distances to prototypes · CPC title

• G06F18/2137

  based on criteria of topology preservation, e.g. multidimensional scaling or self-organising maps · CPC title