US4953099A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Information discrimination cell
Assignee: Massachusetts Institute of Technology
•
Filing: 1988-06-07
•
Granted: 1990-08-28
•
Expires: 2008-06-07
Abstract
Abstract Methods of information discrimination based on extremization of local, energy related functionals defined on macroscopic states of random code information processors are disclosed. In one embodiment, an assembly of processors exchanging information through the exchange of random codes serves an an information discriminator cell between the digitally encoded, analog, quasi-periodic input function and the digitally encoded, analog, quasi-periodic memory function stored within the cell.
Inventors
- 👤 Alexander N. Jourjine
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-05 11:14
CPC Classifications
G; G06; G06N; G06N3/00; G06N3/02; G06N3/04; G06F; G06F18/00; G06F18/20; G06F18/21; G06F18/213; G06F18/2132; G06N3/044; G06N3/0442
Claims Analysis
Total Claims: 10
Claim 1
Independent Claim
118 words
A random code neural network with learning capability for discrimination of the difference between digitally encoded input and digitally encoded memory information comprising: a plurality of input cells, a plurality of memory cells and a plurality of discrimination cells interconnected for communication therebetween to exchange signals in the form of random binary codes, each input cell, memory cell, and discrimination cell including means to emit said signals periodically within time spans of length Tp within larger time spans Tc and including means for learning external input by extremizing a local functional of the total input of said cell's binary code by adjustment of the position of the time span of duration Tp within the larger time span Tc.
Claim 2
Dependent Claim
118 words
A random code neural network with learning capability for discrimination of the difference between digitally encoded input and digitally encoded memory information comprising: a plurality of input cells, a plurality of memory cells and a plurality of discrimination cells interconnected for communication therebetween to exchange signals in the form of random binary codes, each input cell, memory cell, and discrimination cell including means to emit said signals periodically within time spans of length Tp within larger time spans Tc and including means for learning external input by extremizing a local functional of the total input of said cell's binary code by adjustment of the position of the time span of duration Tp within the larger time span Tc.
Claim 3
Dependent Claim
30 words
A neural network as claimed in claim 1 wherein the means for learning extremizes a local functional which is a piece-wise linear function of the position of the time span.
Claim 4
Dependent Claim
30 words
A neural network as claimed in claim 1 wherein the means for learning extremizes a local functional which is a piece-wise quadratic polynomial in the position of the time span.
Claim 5
Independent Claim
131 words
A method for using a random code neural network with learning capability for discrimination of the difference between digitally encoded input and digitally encoded memory information comprising: arranging a plurality of input processors, memory processors and discrimination processors in a random code neural network to exchange information between the processors in the form of random binary codes periodically by emitting random binary codes within time spans of duration Tp within larger time spans of duration Tc; and in each of the memory processors and discrimination processors, extremizing a local functional of the total input to the processor and said processor's binary code by adjusting the position of the time span of duration Tp within the larger time span Tc to learn in the memory processor external input to the input processor.
Claim 6
Independent Claim
131 words
A method for using a random code neural network with learning capability for discrimination of the difference between digitally encoded input and digitally encoded memory information comprising: arranging a plurality of input processors, memory processors and discrimination processors in a random code neural network to exchange information between the processors in the form of random binary codes periodically by emitting random binary codes within time spans of duration Tp within larger time spans of duration Tc; and in each of the memory processors and discrimination processors, extremizing a local functional of the total input to the processor and said processor's binary code by adjusting the position of the time span of duration Tp within the larger time span Tc to learn in the memory processor external input to the input processor.
Claim 7
Dependent Claim
25 words
A method as claimed in claim 4 wherein the functional in each processor is a piece-wise linear function of the position of the time span.
Claim 8
Dependent Claim
23 words
A method of claim 4 wherein the functional in each processor is a piece-wise quadratic polynomial in said position of the time span.
Claim 9
Dependent Claim
19 words
A method as claimed in claim 4 wherein the binary codes are emitted in the form of radiant energy.
Claim 10
Dependent Claim
30 words
A method as claimed in claim 4 wherein the binary codes are in the form of binary strings and the input processors, memory processors and discrimination processors are digital processors.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1993-08-27 | 1994-02-26 | 1994-08-28 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1997-08-27 | 1998-02-26 | 1998-08-28 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2001-08-27 | 2002-02-26 | 2002-08-28 | $7700.00 | Window Closed / Lapsed |