Data processing: artificial intelligence – Knowledge processing system – Knowledge representation and reasoning technique
Reexamination Certificate
1998-08-19
2003-03-04
Starks, Jr., Wilbert L. (Department: 2122)
Data processing: artificial intelligence
Knowledge processing system
Knowledge representation and reasoning technique
C702S002000
Reexamination Certificate
active
06529890
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to an expert system for obtaining information about synoptic climatology. It further relates to a method for representing synoptic climatology information in a frame based hierarchy.
2. Description of the Related Art
2A. Artificial Intelligence and Expert Systems
This invention presumes that the practitioner is familiar with knowledge-based systems terminology, including object-oriented programming techniques as well as terminology used for knowledge processing applications that is, applications conventionally associated with the field of artificial intelligence (AI). This invention also presumes that the practitioner is familiar with terminology in the field of synoptic climatology. This invention relates to the field of artificial intelligence, for example to the field of expert systems or knowledge-based systems. It is to be noted that practitioners in this field use the terms expert systems and Knowledge-based Systems interchangeably. Within the scope of this application and invention, the terms Knowledge-based Systems and expert systems to mean the same thing. Principles of AI and Expert Systems are described in detail in U.S. Pat. Nos. 5,313,636, 4,930,071, 4,918,621 and 4,675,829 all of which are incorporated herein by reference.
Artificial intelligence (AI) technology is a branch of computer science with an ultimate goal of providing a machine that is capable of reasoning, making inferences and following rules in a manner believed to model the human mind. There have been substantial advances in the theoretical aspects of AI though much remains to be done. Principles developed in Artificial intelligence theory are increasingly finding applications. It is being accepted now that AI principles can be effectively applied to develop better computer software. AI also provides users sophisticated ways to use computer power to solve day to day practical problems. These include assisting in the analysis of massive amounts of relatively unprocessed data to aid in decision-making processes.
It is helpful to understand what is meant by knowledge and a knowledge base as now understood. Knowledge in the pragmatic terms of artificial intelligence is described in terms of its representation. Knowledge is a combination of data structures and interpretive procedures which, if suitably manipulated (as by a suitably programmed computing machine), will lead to what might best be termed “knowledgeable” behavior. A knowledge base is a set of knowledge representations which describes a domain of knowledge. See generally Elaine Rich, McGraw-Hill Book Company, New York, N.Y. (1983) (hereinafter Rich). A knowledge base is to an artificial intelligence environment what a database is to a conventional computer program. Unlike a database, however, a computer knowledge base can include executable program material within a defined record, herein called a slot, and is separate from the inference engine and control strategy used for problem solving within the domain of expertise being modeled.
Knowledge representation techniques and theories are still evolving. Nevertheless, knowledge representation techniques appear to be classifiable into various categories depending on the type of knowledge being represented. One category of knowledge is descriptive knowledge. This category of knowledge representation provides techniques for the collection and organization of facts, ideas or entities which might be acted upon. The basic units of descriptive knowledge are generally called frames, as hereinafter explained. They have also been known variously as units, concepts or objects. The term frame lacks some precision of meaning due to its use in other disciplines. Therefore, hereinafter a basic unit of descriptive knowledge is denoted a knowledge representation frame or KR frame. A KR frame contains one or more slots.
Another category of knowledge representation is that of procedural knowledge in the form of rules or structured reasoning procedures. This category of knowledge representation includes techniques which emulate the human mind's structural capability to make choices. The premise-conclusion (IF THEN) format is a typical representation of a procedural knowledge conditional expression. Procedural knowledge emphasizes action and is encoded into a knowledge base as a rule in conditional expression form. The procedural knowledge may reside in a slot of a KR frame.
The knowledge base has expert rules of thumb (or heuristics) that are extracted from a domain expert. A typical rule is in the form, for example:
If
Condition A is satisfied.
Condition B is satisfied. AND
Condition C is satisfied. AND
Then
Assert D AND
Perform E.
That is, if a plurality of conditions are satisfied in a given problem state, then assert a new condition to the problem state and perform a new step that changes the problem state. Some conditions are satisfied from existing data and some are satisfied after querying the user for additional data. In this example, if a set of conditions A,B and C are satisfied in the given problem state, then condition D should be asserted to the problem state and step E should be performed on existing data.
The inference engine performs the task of executing or applying the rules in the knowledge base to a problem domain. It matches the conditions on the “If” side to the problem state and performs the necessary steps to apply the “Then” side. In contrast to conventional programs, the inference engine of AI systems also selects which rule to apply next, from the set of heuristic rules. Therefore the “knowledge” for the knowledge base is embedded within the rules as well as in the structure of the inference engine. A key feature of the steps followed in the process is the iterative “reasoning” process.
The third category of knowledge representation is that of logic programming wherein knowledge required to derive facts logically from a set of statements is represented with first order predicate calculus statements. Examples of languages using logic programming are the language of the so-called fifth generation computers of the Japanese, called PROLOG and the language MRS employed at Stanford University.
Often domain knowledge, represented with various techniques such as those described above, can be organized naturally in a hierarchical structure. The key to the use of hierarchical structures is the concept of connecting relations between structures of data or knowledge through which information about attributes may pass to other structures of data or knowledge. One of the major contributions of artificial intelligence is the concept of inheritance to provide the connecting relations in a hierarchical structure. The concept of inheritance has a number of advantages. First, an inheritance mechanism allows the specification of many components of a data structure or knowledge structure through reference to other data structures or knowledge structures. As used herein, high-level data structures or knowledge structures refer to organized collections of simpler data structures or knowledge structures, such as a collection of various relations in a relational database sense, or a collection of logical assertions as in the predicate calculus sense. Second, an inheritance mechanism can assure consistency among high-level data or knowledge structures. That is, the inheritance mechanism can be used to specify that a given data or knowledge structure must obey restrictions placed on characteristics from other data or knowledge structures. Third, the inheritance mechanism allows the implementation of semantics. That is, the inheritance mechanism is a technique for combining higher level concepts and specifying meaning.
The concept of representing knowledge as hierarchical data structures with inheritance was first referred to in terms of “frames” by its most prominent early supporter, Marvin Minsky of the Massachusetts Institute of Technology. Professor Minsky gave the first general description of the concept and laid the intellectual
Higgins Kevin
Pandit Nitin S.
Samant Nikhil
Sletten Tim
Stunder Mark
Chid Subramanian
Ensys, Inc.
Starks, Jr. Wilbert L.
LandOfFree
Method for representing synoptic climatology information in... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for representing synoptic climatology information in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for representing synoptic climatology information in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3053690