Back to results
US 2023/0142543 A1Published application
Augmented knowledge base and reasoning with uncertainties and/or incompleteness
OwnerEugene S. SANTOS, Eunice E. SANTOS +2inventors
InventorsEugene S. SANTOS, Eunice E. SANTOS, Evelyn W. SANTOS, Eugene SANTOS, Jr.
Life of this patent
Estimating the term…
Abstract
A knowledge-based system under uncertainties and/or incompleteness, referred to as augmented knowledge base (AKB) is provided, including constructing, reasoning, analyzing and applying AKBs by creating objects in the form E→A, where A is a rule in a knowledgebase and E is a set of evidences that supports the rule A. A reasoning scheme under uncertainties and/or incompleteness is provided as augmented reasoning (AR).
Claims20
A server configured to implement a service to output a result in response to a query, input through a user interface, of a target rule L as a target knowledge to be associated with a set of evidences E which is a subset of set U of evidences in an electronically stored knowledge base (KB) of the set U based on a collection of knowledge in form of rules A in form of a relational database and/or a deductive database in support of the target rule L, the server comprising:
- a processor coupled to a memory storing instructions which when executed by the processor causes the processor to execute a process to,
- obtain mapping information which associates evidences in the set U of evidences with respective weight values representing a corresponding strength characteristic of an evidence among the evidences;
- transform data from the KB to a Free-Form Database (FFDB);
- create, based on the FFDB, data objects in form E→A for data in the KB, where A is as a rule, among rules in the KB, and E is a subset of evidences from the set U of evidences to support the rule A, and an evidence among the evidences in E indicatable as uncertain and/or incomplete based on the corresponding strength characteristic of the evidence, according to a mapped value among the mapped values, so that the rule A is supportable by the set of evidences E according to the mapped values;
- compute relationship constraints κ based on the respective weight values indicating validity and/or plausibility values, the relationship constraints κ indicating relations among a plurality of sets of evidences E, including at least one or any combination of a subset relation or a disjoint relation among the plurality of sets of evidences E;
- compute for the queried target rule L, a first composite object ω1=(G→L0), the first composite object wi constructed from a combination of the data objects E→A to create a composite set of evidences G subject to the relationship constraints κ which support a composite rule L0, the composite rule L0 being a combination of the rules implying the target rule L,
- the first composite object ω1 being indicative, based upon the computed relationship constraints κ, of deductive reasoning of a first validity and/or a first plausibility value for the target rule L;
- and
- output the result representing an extracted knowledge from the KB in form of the queried target rule L associated with the E according to the first composite object ω1.
6 more elements
A server, comprising:
- a non-transitory computer readable storage medium to store:
- a Free-Form Database as a knowledge base (KB) resulting from a conversion of data which is indicated as uncertain according to a certainty factor and/or incomplete from sources inclusive of relational databases, deductive databases, and/or inductive databases,
- mapping information, which associates information as evidences in a set U of evidences in the KB with respective weight values representing a corresponding strength characteristic of an evidence among the evidences in the KB, the strength characteristic indicative of a certainty factor and/or incompleteness, and
- objects in the form E→A, where A is a rule among rules in the KB, and E is a subset of the sets of evidences U from the knowledge base that supports the rule A, so that the rule A is supportable by the subset of evidences E;
- and
- a hardware processor coupled to at least one memory to execute instructions stored in the at least one memory, which instructions when executed by the at least one hardware processor, control the server to,
- augment the data of the KB resulting from a composite object w of data which is indicated as uncertain and/or incomplete in form of an AKB, the composite object ω resulting from,
- obtaining relationship constraints κ in form of set relations among a plurality of subsets of evidences E;
- and
- computing for a target rule L, a composite object with respect to validity (v), ω1p=G1→L0, and/or with respect to plausibility (p), ω1p=G1→L0, the composite object ω1p or ω1p constructed from a combination of the objects E→A to create a composite set of evidences G1p or G1p subject to the relationship constraints κ which support a composite rule L0vor L0p, the composite rule L0vor L0p being a combination of the rules implying the target rule L according to deductive reasoning indicated by the mapping of the values to the sets of evidences U,
- the composite object ω1p or ω1p representing derived new knowledge based upon the relationship constraints κ as indicative according to the deductive reasoning based upon a validity value (v) or a plausibility value (p) for the target rule L.
8 more elements
At least one server, comprising:
- at least one processor;
- and
- at least one non-transitory computer readable storage medium configured to store,
- a Free-Form Database as a knowledge base (KB) resulting from a conversion of data which is indicated as uncertain according to a certainty factor and/or incomplete from sources inclusive of relational databases, deductive databases, and/or inductive databases;
- mapping information, which associates information as evidences in the KB with respective weight values representing a corresponding strength characteristic of an evidence among the evidences in the KB, the strength characteristic indicative of a certainty factor and/or incompleteness;
- objects in the form E→A, where A is a rule among rules in the KB, and E is a subset of a set U of evidences from the KB that supports the rule A, so that the rule A is supportable by the subset of evidences E;
- and
- instructions which when executed by the at least one processor cause the at least one server to execute a process to,
- augment the data of the KB in form of an augmented knowledge base (AKB) resulting from at least one composite object w to provide the data which is indicated as uncertain and/or incomplete in form of the AKB, the at least one composite object w resulting from,
- obtaining relationship constraints κ in form of set relations among a plurality of subsets of evidences E;
- and
- computing for a target rule L, a composite object with respect to validity (v), ω1v=G1→L0 , and/or with respect to plausibility (p), ω1p=G1→L0, the composite object ω1p or ω1p constructed from a combination of the objects E→A to create a composite set of evidences G1p or G1p subject to the relationship constraints κ which support a composite rule L0vor L0p, the composite rule L0vor L0p being a combination of the rules implying the target rule L according to deductive reasoning indicated by the mapping of the values to the sets of evidences U,
- the composite object ω1p or ω1p indicative according to the deductive reasoning of a first validity value or a first plausibility value for the target rule L;
- and/or
- computing for at least one target rule L, a composite object with respect to i-validity (v), ω1iv=G1→L0, and/or with respect to i-plausibility (p), ω1ip=G1→L0, the composite object ω1iv or ω1ip constructed from a combination of the objects E→A to create a composite sets of evidences G1ivor G1ip subject to the relationship constraints κ which support a composite rule L0iv or L0ip, the composite rule L0iv or L0ip being a combination of the rules implied by the target rule L according to inductive reasoning indicated by the mapping of the values to the sets of evidences U,
- the composite object ω1iv or ω1ip indicative according to the inductive reasoning of a first i-validity value or a first i-plausibility value for the target rule L.
13 more elements
Dependencies are inferred from the claim text.