Patent US4875183A successfully refreshed from Google Patents.
US4875183A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Neural networks
Assignee: Nokia Bell Labs
•
Filing: 1987-11-19
•
Granted: 1989-10-17
•
Expires: 2007-11-19
Abstract
Abstract The operation of neural networks begins with the initialization of the system with the information to be processed. Presently, this initialization is performed by pinning the system with rather large analog or digital signals representing this information. The problems associated with the high power required for such initialization are eliminated and accuracy is maintained by utilizing a specific set of input points and appropriately positioned switches. In particular, a switch corresponding to each amplifier is introduced, and the initializing data is introduced between the amplifier and this switch.
Inventors
- 👤 Hans P. Graf
- 👤 Richard E. Howard
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-26 13:11
CPC Classifications
G; G06; G06N; G06N3/00; G06N3/02; G06N3/06; G06N3/063; G06N3/065
Claims Analysis
Total Claims: 11
Claim 1
Independent Claim
114 words
A process for the parallel processing of signals by a neural network, said neural network comprising (a) a multiplicity of inputs for a matrix with rows and columns connected by interconnected elements and (b) active elements interacting with said matrix, said processes comprising the steps of (1) applying a signal to said multiplicity of inputs wherein switches present in said matrix are located such that each of said inputs is made between a switch in said matrix and the input of one of said active elements and such that, with the switch open, essentially no processing of information occurs, and (2) closing said switches wherein the parallel processing of said applied inputs is initiated.
Claim 2
Dependent Claim
114 words
A process for the parallel processing of signals by a neural network, said neural network comprising (a) a multiplicity of inputs for a matrix with rows and columns connected by interconnected elements and (b) active elements interacting with said matrix, said processes comprising the steps of (1) applying a signal to said multiplicity of inputs wherein switches present in said matrix are located such that each of said inputs is made between a switch in said matrix and the input of one of said active elements and such that, with the switch open, essentially no processing of information occurs, and (2) closing said switches wherein the parallel processing of said applied inputs is initiated.
Claim 3
Dependent Claim
11 words
The process of claim 1 wherein said interconnective elements comprise resistors.
Claim 4
Dependent Claim
12 words
The process of claim 2 wherein said active element comprises an amplifier.
Claim 5
Dependent Claim
12 words
The process of claim 1 wherein said active element comprises an amplifier.
Claim 6
Dependent Claim
95 words
A neural network comprising (1) a multiplicity of inputs for a matrix, said matrix with rows and columns connected by interconnective elements and (2) active elements interacting with said matrix CHARACTERIZED IN THAT a plurality of switches are positioned in said matrix such that each of said inputs is made between one of said switches and the input of one of said active elements and such that when said switches are open said matrix does not process signals applied to said inputs and when said switches are closed parallel processing of said signals is induced.
Claim 7
Dependent Claim
95 words
A neural network comprising (1) a multiplicity of inputs for a matrix, said matrix with rows and columns connected by interconnective elements and (2) active elements interacting with said matrix CHARACTERIZED IN THAT a plurality of switches are positioned in said matrix such that each of said inputs is made between one of said switches and the input of one of said active elements and such that when said switches are open said matrix does not process signals applied to said inputs and when said switches are closed parallel processing of said signals is induced.
Claim 8
Dependent Claim
11 words
The neural-network of claim 5 wherein said interconnective elements comprises resistors.
Claim 9
Dependent Claim
12 words
The neural-network of claim 6 wherein said active element comprises an amplifier.
Claim 10
Dependent Claim
14 words
The neural-network of claim 6 wherein a portion of said resistors has infinite resistance.
Claim 11
Dependent Claim
12 words
The neural-network of claim 5 wherein said active element comprises an amplifier.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1992-10-16 | 1993-04-17 | 1993-10-17 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1996-10-16 | 1997-04-17 | 1997-10-17 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2000-10-16 | 2001-04-17 | 2001-10-17 | $7700.00 | Window Closed / Lapsed |