Neural cortex

• US 7,305,370 B2
• Filed: 11/30/2000
• Issued: 12/04/2007
• Est. Priority Date: 11/30/2000
• Status: Expired due to Term

• Current Assignee: Alexei Mikhailov, Yang Ming Pok
• Original Assignee: Alexei Mikhailov, Yang Ming Pok
• Inventors: Mikhailov, Alexei, Pok, Yang Ming
• Primary Examiner(s): Knight; Anthony
• Assistant Examiner(s): Holmes; Michael B.
• Application Number:
  US10/398,279
• Publication Number:
  US 20040034609A1
• Time in Patent Office: 2,560 Days
• Field of Search: 706/20
• US Class Current: 706/20
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06N 3/063 using electronic means

Abstract

A neural network system includes a random access memory (RAM); and an index-based weightless neural network with a columnar topography; wherein patterns of binary connections and values of output nodes'"'"' activities are stored in the RAM. Information is processed by pattern recognition using the neural network by storing a plurality of output patterns to be recognized in a pattern index; accepting an input pattern and dividing the input pattern into a plurality of components; and processing each component according to the pattern index to identify a recognized output pattern corresponding to the input pattern.

Claims

1. A method of processing an input pattern by pattern recognition using a neural network storing an index to identify at least one output pattern as a recognized output pattern corresponding to the input pattern, the method including the acts of:

storing a plurality of output pattern indications, each corresponding to an output pattern to be recognized;

accepting an input pattern;

dividing the input pattern into a plurality of input components;

processing each input component of the plurality according to the index, the processing comprising;

looking up, in a RAM, a first link associated with a first output pattern;

using the first link to look up, in the RAM, a second link associated with a second output pattern; and

for each of the first link and the second link looked up, increasing a corresponding stored score corresponding to the associated first output pattern and the associated second output pattern, respectively; and

using each input component of the plurality of input components as the address of a corresponding output pattern column wherein the stored score is allocated to one or more recognized output patterns that are found in the corresponding output pattern columns;

outputting as the recognized output pattern the output pattern with a highest score.

2. The method according to claim 1, wherein each of the at least one output pattern is divided into a plurality of recognized components with each recognized component being stored in the index for recognition.

3. The method according to claim 2, wherein the index consists of columns with each column corresponding to one or more recognized components, each column being represented with a plurality of linked index values.

4. The method according to claim 3, wherein the index is divided into a number of columns which is equal to or less than the number of recognized components.

5. The method according to claim 3, wherein the index is divided into a number of columns which is equal to the number of recognized components.

6. The method according to any one of claims 2 to 4, wherein each input component of the plurality of input components is used as the address of a corresponding output pattern column and a score is allocated to one or more recognized output patterns that are found in the corresponding output pattern columns.

7. The method according to claim 6, wherein the output pattern with the highest score is identified as the recognized output pattern.