Knowledge acquisition is a critical aspect of the knowledge engineering approaches that are used in complex diagnostic problem-solving. Experience with a machine diagnostic problem pointed to difficulties in mapping the acquired knowledge to the fault tree representation and with obtaining an overall picture of the system. We also found that use of the prototype for interactive knowledge elicitation required extremely structured interviews to allow the knowledge engineer time for on-line changes. A knowledge acquisition methodology was developed that works in conjunction with a fault-tree diagnostic tool. The methodology employs task analysis, process analysis, structured interviews, and analogies in a process that focuses on the relationships between symptoms and faults.To correct the problems encountered during task analysis of the fault tree, we used a systems approach to modify the original knowledge acquisition methodology. In the modified methodology, knowledge comprising the symptom and component trees is acquired independently in an almost mechanical fashion. Information to link the symptoms to the components is then obtained from the most capable experts using standard knowledge acquisition techniques. This process has the advantage of providing a framework of symptoms and components that focus the knowledge acquisition activities toward identifying the relationships between these objects as well as assisting in verification and validation during later knowledge acquisition sessions. The new methodology also helps to structure the placement of faults, and questions about faults, to avoid replication and to fit easily into the tool logic. We believe that the symptom-component approach to knowledge acquisition for component diagnostic problems is a significant improvement on our original approach and that it can have general applicability to fault-tree diagnosis.
What is AI?"AI is the science which studies the intelligent behavior of systems"Which are the main 81eas of AI?The AI can be divided into three main areas: (i) knowledge-based and expert systems; (ii) natural languages; and (iii) perception and learning systems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.