Progress toward a device representation that organizes knowledge based on functionality is described. Device representation involves theories about languages for representing structure, commitments for representing behavior, and kinds of causation needed to represent behaviors. The current focus is on the first two issues. The functional representation described provides a package that shows the relationship among structure, function, and behavior. Knowledge of this relationship provides basic, task-independent, intrinsic capabilities: simulation, i.e., given changes in a devices structure, what can be determined about changes in functionality; identification of structural cause. i.e. given changes in function (malfunction or reduced effects), what changes in structure could account for them; and identification of functional components, i.e. given a specific component, what functional purpose it provides. The structure of this functional representation, organized around functional packages, provides the means by which these capabilities can be accomplished.<<ETX>>
[1]
Johan de Kleer,et al.
A Qualitative Physics Based on Confluences
,
1984,
Artif. Intell..
[2]
Pamela K. Fink,et al.
Expert Systems and Diagnostic Expertise in the Mechanical and Electrical Domains
,
1987,
IEEE Transactions on Systems, Man, and Cybernetics.
[3]
D. Bobrow.
Qualitative Reasoning about Physical Systems
,
1985
.
[4]
Randall Davis,et al.
Diagnosis Based on Description of Structure and Function
,
1982,
AAAI.
[5]
Jon Sticklen,et al.
'Deep' models and their relation to diagnosis
,
1989,
Artif. Intell. Medicine.
[6]
Benjamin Kuipers,et al.
Commonsense Reasoning about Causality: Deriving Behavior from Structure
,
1984,
Artif. Intell..
[7]
Balakrishnan Chandrasekaran,et al.
Machine understanding of devices: causal explanation of diagnostic conclusions
,
1989
.
[8]
Jon Sticklen,et al.
Mdx2: an integrated medical diagnostic system
,
1987
.