US5047951A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Inference processor using meta knowledge
Assignee: Toshiba Corp
•
Filing: 1989-05-10
•
Granted: 1991-09-10
•
Expires: 2009-05-10
Abstract
Abstract An inference processor includes a production rule memory, a meta knowledge generator, a meta knowledge memory, a production system interpreter, and a working memory. The meta knowledge generator obtains production rules satisfying transition conditions between all the combinations of the states of the respective variables in the production rules as a directed-graph matrix. The meta knowledge generator then decomposes the directed-graph matrix into strongly connected components, and obtains, as meta knowledge, 1) variable-conditions of the respective strongly-connected components and 2) production rules satisfying transition conditions between the respective strongly-connected components. The interpreter generally determines an inference path from a given state to another state as a target on the basis of the meta knowledge, and efficiently executes production rules, thereby performing inference processing.
Inventors
- 👤 Kenji Ono
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-31 02:28
CPC Classifications
G; G06; G06N; G06N5/00; G06N5/04; G06N5/046
Claims Analysis
Total Claims: 6
Claim 1
Independent Claim
129 words
An inference processor comprising: rule storage means for storing production rules formed of pairs of condition part and conclusion part; meta knowledge generating means for obtaining a directed-graph matrix formed of production rules satisfying transition conditions between all combinations of states of variables included in the production rules stored in said rule storage means, decomposing the directed-graph matrix into strongly-connected components, and obtaining conditions of variables of respective strongly-connected components and production rules satisfying transition conditions between the respective strongly-connected components as meta knowledge; and interpreter means for obtaining a general direction of inference on the basis of the meta knowledge obtained in said meta knowledge generating means, and executing inference processing by executing a production rule stored in said rule storage means on the basis of the direction.
Claim 2
Dependent Claim
129 words
An inference processor comprising: rule storage means for storing production rules formed of pairs of condition part and conclusion part; meta knowledge generating means for obtaining a directed-graph matrix formed of production rules satisfying transition conditions between all combinations of states of variables included in the production rules stored in said rule storage means, decomposing the directed-graph matrix into strongly-connected components, and obtaining conditions of variables of respective strongly-connected components and production rules satisfying transition conditions between the respective strongly-connected components as meta knowledge; and interpreter means for obtaining a general direction of inference on the basis of the meta knowledge obtained in said meta knowledge generating means, and executing inference processing by executing a production rule stored in said rule storage means on the basis of the direction.
Claim 3
Dependent Claim
72 words
The processor according to claim 1 in which said interpreter means receives an initial state and a certain state, determines whether or not there is a path leading from a strongly-connected component including the initial state to a strongly-connected component including the certain state by referring to the meta knowledge, and obtains a production rule or an execution order of the production rules when it is determined that there is the path.
Claim 4
Dependent Claim
54 words
The processor according to claim 2, in which if the obtained rule is applicable, said interpreter means immediately executes the rule, and if the rule is not applicable, other applicable rules are applied until the rule becomes applicable so that transition to other strongly-connected components from the strongly-connected component of interest does not occur.
Claim 5
Dependent Claim
165 words
The processor according to claim 1 in which said metal knowledge generating means comprises: means for storing a first matrix of n×n (n is a total number of the combinations); means for registering a production rule including a transition condition from a node i to node j in an (i,j) component of the first matrix, and registering zero in remaining components; means for simultaneously replacing rows and columns of the first matrix such that zeros are set in a lower left portion of the first matrix and square block matrices are set on diagonal lines thereof; means for obtaining a logical OR of states of variables of each node included in each strongly-connected component; means for storing a second matrix of m×m (m is a total number of strongly-connected components); and means for registering a production rule including a transition condition from a strongly-connected component k to a strongly-connected component l in a (k,l) component of the second matrix, and registering zeros in remaining components.
Claim 6
Dependent Claim
26 words
The processor according to claim 1 in which said meta knowledge generating means comprises means for obtaining partial orders of the strongly-connected components as metal knowledge.
References Cited (Backward Citations)
JPS6072031A
Cited
US4704695A
Cited
JPS6219940A
Cited
US4761746A
Cited
US4783752A
Cited
US4853873A
Cited
US4847784A
Cited
US4849905A
Cited
US4890240A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1994-09-09 | 1995-03-11 | 1995-09-10 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1998-09-09 | 1999-03-11 | 1999-09-10 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2002-09-09 | 2003-03-11 | 2003-09-10 | $7700.00 | Window Closed / Lapsed |