 |
|
|
|
|
|
 Title: |
US5613040:
Neural network shell for application programs
[ Derwent Title ]
|
| Country: |
US United States of America
|
| |
| Inventor: |
Bigus, Joseph P.; Rochester, MN
|
| Assignee: |
International Business Machines Corporation, Armonk, NY
other patents from INTERNATIONAL BUSINESS MACHINES CORPORATION (280070) (approx. 44,393)
News, Profiles, Stocks and More about this company
|
| Published / Filed: |
1997-03-18
/ 1995-06-02
|
| Application Number: |
US1995000460351
|
| IPC Code: |
Advanced:
G06N 3/04;
Core:
G06N 3/00;
IPC-7:
G06F 15/18;
|
| U.S. Class: |
Current:
706/025;
706/044;
706/060;
Original:
395/023;
395/026;
|
| Field of Search: |
395/023,26,76
|
| Priority Number: |
|
| Abstract: |
A neural network shell has a defined interface to an application program. By interfacing with the neural network shell, any application program becomes a neural network application program. The neural network shell contains a set of utility programs that transfers data into and out of a neural network data structure. This set of utility programs allows an application program to define a new neural network model, create a neural network data structure, train a neural network, and run a neural network. Once trained, the neural network data structure can be transported to other computer systems or to application programs written in different computing languages running on similar or different computer systems.
|
| Attorney, Agent or Firm: |
Gamon, Owen J. ;
Rose, Curtis G. ;
|
| Primary / Asst. Examiners: |
Downs, Robert W.;
|
| Maintenance Status: |
E3 Expired Check current status
|
| INPADOC Legal Status: |
Show legal status actions
Family Legal Status Report
|
|
|
|
|
|
|
| Parent Case: |
This is a divisional of application Ser. No. 08/075,370 filed Jun. 11, 1993, now issued as U.S. Pat. No. 5,450,529, which is a divisional of application Ser. No. 07/849,116 filed on Mar. 10, 1992, now issued as U.S. Pat. No. 5,222,196, which was a divisional of application Ser. No. 07/482,450, filed on Feb. 20, 1990, now issued as U.S. Pat. No. 5,142,665.
|
| Designated Country: |
CH DE ES FR GB IT LI NL SE
|
| Family: |
Show 14 known family members
|
First Claim:
Show all 14 claims |
What is claimed is:
1. A method of teaching a neural network in a computer system, the neural network having an input, an output for a result, comprising the steps of:
- creating a neural network data structure in memory of the computer system, wherein the neural network data structure defines a structure and organization of a plurality of data arrays in the memory, wherein the plurality of data arrays have data types and an order within the neural network data structure, wherein the neural network data structure is common and generic to all of a plurality of defined neural network models, and wherein the plurality of data arrays have data values specific to a selected one of the plurality of defined neural network models;
- presenting training data at said input of said neural network;
- repeatedly adjusting the values of said plurality of data arrays until said result at said output is within tolerance of a correct result; and
- locking the values of said plurality of data arrays responsive to said adjusting step.
|
| Background / Summary: |
Show background / summary
|
| Drawing Descriptions: |
Show drawing descriptions
|
| Description: |
Show description
|
|
|
|
|
|
|
| U.S. References: |
Go to Result Set:
No U.S. references
| No patents reference this one
|
Backward references (0)
|
|
Nominate this for the Gallery...
