Data processing: artificial intelligence – Knowledge processing system
Reexamination Certificate
2001-10-04
2004-07-13
Davis, George B. (Department: 2121)
Data processing: artificial intelligence
Knowledge processing system
C706S047000, C706S050000
Reexamination Certificate
active
06763341
ABSTRACT:
TECHNICAL FIELD
This invention relates to object-oriented knowledge base systems.
BACKGROUND ART
Smart explanation and commentary on the technical term ‘knowledge base’ is given in {circle around (∘)}“Joho Sisutemu Handobukku” (In Japanese). It is summarized as follows:
Roughly speaking, a knowledge base is an integrated items of knowledge (expert knowledge and/or empirical knowledge, etc.) of real world (usually of a restricted domain of the real world) which are usually stored in computer systems and are described in a form readily used for the purpose of, say, inference.
Items of knowledge in a knowledge base can be roughly classified into ‘facts’ and ‘rules’. For example, the knowledge that “A crow is a bird” is an example of a ‘fact’, and the knowledge that “Any bird flies in the air” (i.e. the knowledge that “No matter what a creature may be, the creature flies in the air if the creature is a bird.”) is an example of a ‘rule’. If an ‘inference’ is carried out on the bases of these two items of knowledge, the answer to the question, for example “Dose a crow fly?” can be answered, if a perfect knowledge base system whose ability matches human ability of thought would exist and are used. A portion carrying out such inferences like this is usually called ‘inference mechanism’.
Usually, the main purpose to construct a knowledge base is to give solutions to varieties of problems by using the knowledge base in combination with an inference mechanism. If a very universal inference mechanism exists, then, procedures for solving problems have not necessarily be explicitly described as computer programs; That is, if a knowledge base system has a very universal inference mechanism, then a user of the knowledge base system can solve problems only by representing the necessary items of knowledge according to the style of representation of the knowledge base. Therefore, if so, the user may not necessarily be skillful at coding computer programs.
An important issue to be discussed when a knowledge base is constructed, is the way of ‘knowledge representation’. As ideal merits that a good ‘knowledge representations’ should have, for examples,
1) Having high power of expression and being able to describe varieties of things and matters in a systematic manner.
2) Having high readability and being easily understood by human.
3) Having high degree of modularity and being highly flexible to the renewal of items of knowledge stored in the knowledge base.
4) With which, being able to implement (=embody) high performance processing systems.
Four types of methods for knowledge representation are broadly known:
1) Predicate logic.
2) Production rule.
3) Semantic network.
4) Frame.
The ‘semantic network’ model is originally introduced to describe human memory and ability of association in the field of cognitive psychology (see for example {circle around (∘)}“Gurafikku Ninchi shinrigaku” p.86), but is frequently used also in the field of artificial intelligence: (see for example, {circle around (∘)}“The CLASSIC Knowledge Representation System or, KL-ONE”).
The KL-ONE is one of the most famous knowledge representation systems. The KL-ONE's root is in ‘semantic network’, but KL-ONE is influenced in part by ‘Frame’. The CLASSIC Knowledge Representation system is a new generation of KL-ONE-like systems ({circle around (∘)}“The CLASSIC Knowledge Representation System or, KL-ONE”). The KL-ONE and its descendants are one of the most long-lived knowledge representation systems, and its research has lasted over two decades at least up to today. For more details, see for example, recent US Patents
{circle around (∘)}“U.S. Pat. No. 5,974,405 Oct. 26, 1999 Knowledge base management system with enhanced explanation of derived information and error objects”, and,
{circle around (∘)}“U.S. Pat. No. 5,659,724 Aug. 19, 1997 Interactive data analysis apparatus employing a knowledge base”.
As the fifth type of method for knowledge representation,
5)‘Object-oriented knowledge base’
is known.
The original idea ‘Object’ was first introduced when a computer language ‘Simula67’ was designed. The Simula67 got extremely high reputation as a computer simulation system (Chapter 2 of {circle around (∘)}“MODELLING the WORLD with OBJECTS”). Thus, the original idea, ‘Object’ was born primary as a concept of programming language (i.e. Simula67) rather than as a method only for knowledge representation ({circle around (∘)}“Chishiki no hyo-gen to kousoku suiron”, p.11).
The term ‘object-oriented programming language’ is a general and somewhat vaguely used term, and in many cases represents a programming language supporting the way to implement (=embody) an ‘object’, which is usually represented in a source code of a computer program as an integrated unit composed of algorithms and data ({circle around (∘)}“Joho Sisutemu Handobukku” p.4-202). The SIMULA67 is of course an object-oriented programming language. Smalltalk and C++ are also object-oriented programming languages. Usually, each of “graphical objects” in a GUI (Graphical User Interfaces) of computers, such as icons, windows, dialog boxes, Mouse cursors etc. are usually implemented by using object oriented languages. In such cases, above mentioned type of an integrated unit composed of algorithms and data defined in the source code of the object oriented software supporting the GUI, is usually used to embody a “graphical object”.
Some persons believe in that object-oriented representation is just suitable to be used not only as a method to implement “graphical objects” in a GUI, but also as a method of knowledge representations in a knowledge base system ({circle around (∘)}“MODELLING the WORLD with OBJECTS” Chapter 12). It is highly desirable today to propose varieties of types of object-oriented knowledge base systems of practical use until the ultimate and perfect solution is got.
As a matter of fact, some pioneering and challenging work to construct object-oriented knowledge base systems have been done, including, O-logic, ({circle around (∘)}“A Logic for Object-Oriented Logic Programming (Maier's O-Logic: Revisited)”, and, Transaction Logic ({circle around (∘)}“Transaction Logic Programming (or, a Logic of Procedural and Declarative Knowledge”), and, Quixote({circle around (∘)}“Specific Features of a Deductive Object-Oriented Database Language Quixote”). Knowledge representations used in these knowledge base systems are very precise and are extremely mathematical.
However, these object-oriented knowledge base system is not so popular when compared with other software systems, say, ‘©MS-DOS’, ‘©Windows 95’, ‘©Office’ (presented by Microsoft). It is highly desirable to give object-oriented knowledge base systems with facts and rules having high readability and being easily understood by human to present a knowledge base system of very wide and popular use. One of the final purposes to be attained in developing an ultimate object-oriented knowledge base system is to give
a knowledge base system in which
a flexible, systematic, exact, exhaustive, and user-friendly way of ‘knowledge representation’ which covers not only what can be described mathematically but also covers what can be described linguistically, and,
a mighty ‘inference mechanism’ for inference widely applicable to wide varieties of sentences in the ‘knowledge representation’, coexist.
DISCLOSURE OF INVENTION
1. SUMMARY
A useful object-oriented knowledge base system is provided, which comprises an ‘object-oriented knowledge base’, an inference mechanism, and an ‘object-oriented knowledge base management system’. In the knowledge base, simple English sentences are used as a rule and/or as a fact. An ‘ideal dictionary’ exists in the ‘object-oriented knowledge base management system’, wherein an object-oriented-lexical-definition of nouns is given, and dichotomy in combination with c-language-like way of description of English sentences is used to give a lexical definition of a verb. The ‘
LandOfFree
Object-oriented knowledge base system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Object-oriented knowledge base system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Object-oriented knowledge base system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3212155