Inference Rules In Expert Systems

Expert systems capture rules, which associate known values of some attributes with assertions that can be made about other attributes.

The orderly processing of these rules dictates the user dialog. This process includes:

1. Knowledge acquisition
2. Inference

Knowledge Acquistion

The knowledge acquisition component of the expert system serves to input the characteristics known to be appropriate to a good inference technique.

A good inference technique is independent of the problem domain. Instead, it must strip away the domain specific expertise to realize the benefits of explanation, knowledge transparency, and reusability of the programs in a new problem domain.

Inference Rule

The "inference rule" concept is fundamental to understanding expert systems. An inference rule is a statement that has two parts:

IF [clause] and THEN [clause]

Using this rule, the expert system has the ability to find solutions to diagnostic and prescriptive problems.

In a travel scenario:

1. If the destination choice includes French speaking countries, and the occasion is romantic, then the destination is Paris.
2. If the destination choice includes French speaking countries, and the occasion is boating, then the destination is Nice or Monaco.
3. If the destination choice also includes the Grand Prix, then the final choice of destination could only be Monaco....and it would have to be the beginning of May [the time the Grand Prix is held in Monaco each year]

An expert system's rulebase is made up of many such inference rules. They are entered as separate rules and it is the inference engine that uses them together to draw conclusions.

• Each rule is a unit
• Rules may be deleted or added without affecting other rules (though it will affect which conclusions are reached).
• Programming using inference rules over traditional programming provides a solution that more closely resembles human reasoning.
• When a conclusion is drawn, it is possible to understand how this conclusion was reached.

Furthermore, because the expert system uses knowledge in a form similar to the expert, it may be easier to retrieve this information from the expert.

Inference techniques may be specific to a particular task, such as diagnosis of hardware configuration. Other techniques may be committed only to a particular processing technique.

Inference techniques are always specific to the knowledge structures.

There are two main methods of reasoning when using inference rules:

1. Forward chaining
2. Backward chaining

Forward Chaining

1. Forward chaining starts with the data available and uses the inference rules to conclude more data until a desired goal is reached.
2. The inference engine, using forward chaining, searches the inference rules until it finds one in which the if-clause is known to be true.
3. It then concludes the then-clause and adds this information to its data. It would continue to do this until a goal is reached.

The data available determines which inference rules are used. For this reason, this method is also called data driven.

Backward Chaining

1. Backward chaining starts with a list of goals and works backwards to see if there is data which will allow it to conclude any of these goals.
2. An inference engine, using backward chaining, searches the inference rules until it finds one which has a then-clause that matches a desired goal.
3. If the if-clause of that inference rule is not known to be true, it is added to the list of goals.

Whilst an expert system consists primarily of a knowledge base and an inference engine, it must also manage reasoning with uncertainty, and be capable of explaining the line of reasoning used.

Confidences

Knowledge is almost always incomplete and uncertain. Rules may therefore possess a confidence factor or a weight.

Conclusions in expert systems are awarded confidences designed to imitate the confidences humans use in reasoning; rather than the mathematical definitions used in calculating probabilities.

Uncertainty

The set of methods for using uncertain knowledge in combination with uncertain data in the reasoning process is called reasoning with uncertainty. This subclass of methods for reasoning with uncertainty is called 'fuzzy logic', and the systems that use them are commonly known as 'fuzzy systems'.

Rationale

Expert system based on uncertain knowledge, are often required to rationalize their decision. If the rationale seems plausible, we tend to believe the answer.

Most expert systems have the ability to answer questions of the form: "Why is the answer X?" Explanations can be generated by tracing the line of reasoning used by the inference engine.

As the rationale is continually tested, the knowledge domain expert may decide that certain pieces of knowledge are wrong. Thus expert systems are organic in nature, and continually evolving. With each refinement, the expert system gradually gains competence, and confidence factors improve.

NEXT: Expert System Architecture

Back To Top