ARTIFICIAL INTELLIGENCE 

A field where the applications generally involve many distinct specialties. In Expert System case, the techniques includes symbolic representation, symbolic inferences and heuristic searching

EXPERT SYSTEM

Special Computer Program, developed through the science of Artificial Intelligence, which executes complex tasks in a way similar a human expert would do in a particular field and containing the knowledge extracted from the expert usually structured with rules and instances, allowing the system a reasoning process

KNOWLEDGE ENGINEER

The Knowledge Engineer extracts the relevant knowledge from the domain expert and creates the Knowledge Base.  He then designs the structure of the Knowledge Tree and formulates the rules which will govern the Expert System

KNOWLEDGE ACQUISITION

Process of extracting knowledge or expertise from a human expert in order to formalize it in machine-interpretable code, generally in the form of decisions rules

KNOWLEDGE BASE

A computerized collection of facts, rules and heuristics on which the Expert System operates

INFERENCE ENGINE

The part of an Expert System that contains the mechanisms for solving problems: An interpreter decides how to apply rules, facts and heuristics in order to infer knowledge and a scheduler decides the order in which those rules, facts and heuristics should be applied

KNOWLEDGE TREE

A decision tree is the graphical representation of the factors and rules of the Expert System being developed. It relates the different application factors and their hierarchical links

A goal is the factor that is found at the root of a decision tree. An Expert System can have one or more goals and sub goals

A factor is a data representation of any physical or abstract entity. A factor has properties and these properties have values

A rule is a logical link between factors and has the form IF... THEN. The IF part contains the conditions of rules, while the THEN part contains the conclusion reached when the conditions are met

While working on a factor and its properties, values are represented in the following manner:

Types of Values can be represented by: Quantitative, Qualitative, Multivalued Qualitative, True/False

Determination of Values can be represented by: Equation, Question, Database

EXPERT SYSTEM DEVELOPMENT STAGES

First, you have to establish a team which usually consists of the Knowledge Engineer, the Domain Expert(s) and a team of programmers

Second, you need to establish the king of purpose you wish to develop the Expert System: Knowledge Acquisition, Knowledge Production, Control of Operations, Predictive Scenarios, Evaluation of Contexts or Individuals

Third, you need to have an Expert System Shell, which will serve to develop the Expert System Program, using the Artificial Intelligence Science

Fourth, the Knowledge Engineer has to establish the Knowledge Acquisition Process which is extracting knowledge or expertise from a human expert in order to formalize it in machine-interpretable code, generally in the form of decisions rules

Fifth, the Knowledge Engineer, after he has gathered all the knowledge in the domain of expertise for which the Expert System will be developed, has to construct the Knowledge Tree and program the rules, "IF-THEN" rules, which will manage the functioning of the system. During this stage are also built all the interfaces which will be available to the user, once the Expert System will be completed and operational

Sixth, once the Knowledge Tree and all the decisions rules have been completely achieved, the Knowledge Engineer has to formulate all the "conclusion messages" which will be given when the system, when utilized, will provide to the user

And the last phase is the testing of the system. To give you an idea of such a serious test would be to give you a possible scenario of a System. Let’s just say that an Expert System, developed in a domain of expertise, would contain a small quantity of 500 questions or situations to be tested

The final test of the system consists of asking all the 500 questions, all at once, to test the knowledge tree, it’s structure for any mistake or error. If there is one, the system will crash. If we were to ask 500 questions to an expert at the same time, we know that it would be impossible for a human being to do so. That gives you an idea of the immense power of an Expert System constructed under Artificial Intelligence

Send mail to jeanlambertpersonalcoach@hotmail.com with questions or comments about this web site

For real time communication you may reach Jean at Windows Live Messenger: jeanlambertpersonalcoach@hotmail.com

Email sent to our support team is answered the next business day