The classic ideal of an expert system is to be capable of carrying on an intelligent conversation with the user of the system, asking only those questions that are relevant to the particular context of solving a present problem. For example, if the subject of a health diagnosis is male, an expert diagnostic system should not ask questions relevant to a woman's health.
Backward chaining is the generally accepted inferencing strategy used to construct such conversational systems. The reason for this is that backward chaining pursues subgoals only within the specific context of the main goal that the inferencing engine is pursuing. When a subgoal cannot be satisfied through firing other rules, it is typical for the inference engine to raise a question so that the system user can provide the value of the subgoal.
CA Aion BRE supports constructing conversational systems using backward chaining; however, the inference engine does not automatically generate questions to the user. The programmer must write code that tells the interfacing layer how to raise the questions. Generating the Interface describes various strategies and architectures for generating conversational systems with CA Aion BRE in different environments.
| Copyright © 2009 CA. All rights reserved. | Send E-mail to CA about this topic |