Patent US4876731A successfully refreshed from Google Patents.
US4876731A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Neural network model in pattern recognition using probabilistic contextual information
Assignee: Google LLC
•
Filing: 1988-02-19
•
Granted: 1989-10-24
•
Expires: 2008-02-19
Abstract
Abstract A pattern recognition system for recognizing an unknown pattern comprised of symbols which are part of a pattern system which is devoid of inherent context such as numbers. Artificial contextual information based on other than symbol features and the pattern system and in the form of probability weighted expected interpretations are stored and used in the processing phase of recognition. In the system disclosed, the system comprises a neural network whose forward and feedback paths are controlled by the output cells of the network based, in part, on the contextual information.
Inventors
- 👤 Keith Loris
- 👤 James Euchner
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-26 13:01
CPC Classifications
G; G06; G06V; G06V30/00; G06V30/10; G06V30/26; G06V30/262; G06V30/274; G06N; G06N3/00; G06N3/02; G06N3/04; G06N3/042
Claims Analysis
Total Claims: 34
Claim 1
Independent Claim
123 words
A system for recognizing an unknown pattern, said unknown pattern including one or more symbols which are part of a pattern system which is devoid of inherent context, comprising: storing means for storing artificial contextual information associated with said unknown pattern and based upon other than information as to symbol features and the pattern system; and recognizer means responsive to said known pattern and to said artificial contextual information, said recognizer means having a processing phase wherein said recognizer means processes at least one symbol of said unknown pattern and outputs as a recognition decision that said one symbol is one of a plurality of category names and wherein in reaching said recognition decision during said processing phase utilizes said artificial contextual information.
Claim 2
Independent Claim
123 words
A system for recognizing an unknown pattern, said unknown pattern including one or more symbols which are part of a pattern system which is devoid of inherent context, comprising: storing means for storing artificial contextual information associated with said unknown pattern and based upon other than information as to symbol features and the pattern system; and recognizer means responsive to said known pattern and to said artificial contextual information, said recognizer means having a processing phase wherein said recognizer means processes at least one symbol of said unknown pattern and outputs as a recognition decision that said one symbol is one of a plurality of category names and wherein in reaching said recognition decision during said processing phase utilizes said artificial contextual information.
Claim 3
Dependent Claim
70 words
A system in accordance with claim 1 wherein: said unknown pattern includes a plurality of symbols; and said recognizer means during said processing phase processes each of said symbols of said unknown pattern and outputs as a recognition decision for each of said symbols that the symbol is one of a plurality of category names and in reaching said recognition decision during said processing phase utilizes said derived contextual information.
Claim 4
Dependent Claim
27 words
A system in accordance with claim 2 wherein: said derived contextual information comprises expected interpretations of said unknown pattern and probability values associated with such expected interpretations.
Claim 5
Dependent Claim
17 words
A system in accordance with claim 3 wherein: said unknown pattern is a sequence of individual numbers.
Claim 6
Dependent Claim
18 words
A system in accordance with claim 4 wherein: said numbers correspond to the dollar amount of a check.
Claim 7
Dependent Claim
17 words
A system in accordance with claim 5 wherein: said check is utilized in payment of a bill.
Claim 8
Dependent Claim
25 words
A system in accordance with claim 6 wherein: one of said expected interpretations of said unknown pattern corresponds to the dollar amount of said bill.
Claim 9
Dependent Claim
42 words
A system in accordance with claim 7 wherein: others of said expected interpretations of said unknown pattern correspond to the dollar amount of a bill previous to said bill and the sum of the dollar amounts of said bill and previous bill.
Claim 10
Dependent Claim
13 words
A system in accordance with claim 8 wherein: said bills are telephone bills.
Claim 11
Dependent Claim
40 words
A system in accordance with claim 2 wherein: said recognizer means comprises: neural network means responsive to said unknown pattern for carrying out said processing phase; and selector means responsive to said derived contextual information for controlling said neural network.
Claim 12
Dependent Claim
97 words
A system in accordance with claim 10 wherein: said neural network means comprises: a plurality of stages of cells including an input stage for receiving said unknown pattern and an output stage having output cells each corresponding to a different one of said category names and to a different set of essential symbol features; an afferent path connecting said stages; and an efferent path connecting said stages for facilitating and inhibiting said afferent path; and said selector means is responsive to said output cells and based thereon and on said derived contextual information controls said efferent path.
Claim 13
Dependent Claim
46 words
A system in accordance with claim 11 wherein: said selector means selects one of said output cells to condition said efferent path so said efferent path facilitates transmission in said afferent path associated with the set of essential symbol features corresponding to said selected output cell.
Claim 14
Dependent Claim
24 words
A system in accordance with claim 12 wherein: said selector means comprises a rule based selector operating on the basis of certainty factor calculus.
Claim 15
Dependent Claim
38 words
A system in accordance with claim 10 wherein: said recognizer means comprises means for automatically changing the attention of said neural network when said recognizer means outputs a recognition decision with respect to a symbol in said pattern.
Claim 16
Dependent Claim
36 words
A system in accordance with claim 2 wherein; said recognizer means comprises means for deleting from said unknown pattern the symbols of said unknown pattern for which said recognizer means has already made a recognition decision.
Claim 17
Dependent Claim
38 words
A system in accordance with claim 2 wherein: said recognizer means includes means for adjusting said derived contextual information in dependence on said recognizer means reaching a recognition decision with respect to the individual symbols in said pattern.
Claim 18
Independent Claim
90 words
A method of recognizing an unknown pattern, said unknown pattern including one or more symbols which are part of a pattern system which is devoid of inherent context, comprising: storing artificial contextual information associated with said unknown pattern and based upon other than information as to symbol features and the pattern system; and recognizing said unknown pattern by processing at least one symbol of said unknown pattern and outputting as a recognition decision that said one symbol is one of a plurality of category names utilizing said artificial contextual information.
Claim 19
Independent Claim
90 words
A method of recognizing an unknown pattern, said unknown pattern including one or more symbols which are part of a pattern system which is devoid of inherent context, comprising: storing artificial contextual information associated with said unknown pattern and based upon other than information as to symbol features and the pattern system; and recognizing said unknown pattern by processing at least one symbol of said unknown pattern and outputting as a recognition decision that said one symbol is one of a plurality of category names utilizing said artificial contextual information.
Claim 20
Dependent Claim
55 words
A method in accordance with claim 17 wherein: said unknown pattern includes a plurality of symbols; and said recognizing includes processing each of said symbols of said unknown pattern and outputting a recognition decision for each of said symbols that the symbol is one of a plurality of category names utilizing said derived contextual information.
Claim 21
Dependent Claim
27 words
A method in accordance with claim 18 wherein: said derived contextual information comprises expected interpretations of said unknown pattern and probability values associated with such expected interpretations.
Claim 22
Dependent Claim
17 words
A method in accordance with claim 19 wherein: said unknown pattern is a sequence of individual numbers.
Claim 23
Dependent Claim
18 words
A method in accordance with claim 20 wherein: said numbers correspond to the dollar amount of a check.
Claim 24
Dependent Claim
17 words
A method in accordance with claim 21 wherein: said check is utilized in payment of a bill.
Claim 25
Dependent Claim
25 words
A method in accordance with claim 22 wherein: one of said expected interpretations of said unknown pattern corresponds to the dollar amount of said bill.
Claim 26
Dependent Claim
42 words
A method in accordance with claim 23 wherein: others of said expected interpretations of said unknown pattern correspond to the dollar amount of a bill previous to said bill and the sum of the dollar amounts of said bill and previous bill.
Claim 27
Dependent Claim
13 words
A method in accordance with claim 24 wherein: said bills are telephone bills.
Claim 28
Dependent Claim
38 words
A method in accordance with claim 18 wherein: said processing step is carried out by neural network means responsive to said unknown pattern and by selector means responsive to said derived contextual information for controlling said neural network.
Claim 29
Dependent Claim
98 words
A method in accordance with claim 26 wherein: said neural network means comprises: a plurality of stages of cells including an input stage for receiving said unknown pattern and an output stage having output cells each corresponding to a different one of said category names and to a different set of essential symbol features; an afferent path connecting said stages; and an efferent path connecting said stages for facilitating and inhibiting said afferent path; and said selector means is responsive to said output cells and based thereon and on said derived contextual information and controls said efferent path.
Claim 30
Dependent Claim
46 words
A method in accordance with claim 27 wherein: said selector means selects one of said output cells to condition said efferent path so said efferent path facilitates transmission in said afferent path associated with the set of essential symbol features corresponding to said selected output cell.
Claim 31
Dependent Claim
24 words
A method in accordance with claim 28 wherein: said selector means comprises a rule based selector operating on the basis of certainty factor calculus.
Claim 32
Dependent Claim
33 words
A method in accordance with claim 26 wherein: said recognizing comprises automatically changing the attention of said neural network when a recognition decision with respect to a symbol in said pattern is reached.
Claim 33
Dependent Claim
32 words
A method in accordance with claim 18 wherein: said recognizing step includes deleting from said unknown pattern the symbols of said unknown pattern for which a recognition decision has already been reached.
Claim 34
Dependent Claim
33 words
A method in accordance with claim 18 wherein: said recognizing step includes adjusting said derived contextual information in dependence on reaching a recognition decision with respect to the individual symbols in said pattern.
References Cited (Backward Citations)
US3440617A
Cited
US3601811A
Cited
US3638196A
Cited
US3810162A
Cited
US4081791A
Cited
US4385285A
Cited
US4479241A
Cited
US4741045A
Cited
US4599692A
Cited
US4754489A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1992-10-23 | 1993-04-24 | 1993-10-24 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1996-10-23 | 1997-04-24 | 1997-10-24 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2000-10-23 | 2001-04-24 | 2001-10-24 | $7700.00 | Window Closed / Lapsed |