Explanatory diagnosis: conjecturing actions to explain observations
From Semantic Portal Wiki
{{#vardefine:category|Publication}}{{#vardefine:templatename|i.publication}}{{#vardefine:package|smwbp_instance_templates}}
| Edit |
Reference: {{#vardefine:pagename|explanatory diagnosis: conjecturing actions to explain observations }}
- [[]]
bibtex
{{#vardefine:pagename|Explanatory diagnosis: conjecturing actions to explain observations }}{{#vardefine:key| }}
abstract: In this paper we present contributions towards a logical theory of diagnosis for systems that can be affected by the actions of agents. Specifically, we examine the task of conjecturing diagnoses to explain {\it what happened} to a system, given a theory of system behaviour and some observed (aberrant) behaviour. We characterize what happened by introducing the notion of explanatory diagnosis in the language of the situation calculus. Explanatory diagnoses conjecture sequences of actions to account for a change in system behaviour. As such, we show that determining an explanatory diagnosis is analogous to classical AI planning with state constraints and incomplete knowledge. The representation scheme we employ provides an axiomatic solution to the frame, ramification and qualification problems. Exploiting this representation, we show that determining an explanatory diagnosis can be achieved by regression followed by theorem proving in the database describing what is known of the initial state of our system. Further, we show that by exploiting features inherent to diagnosis problems, we can simplify the diagnosis task.
download:
- paper:
- slides:
| Abstract | In this paper we present contributions tow … In this paper we present contributions towards a logical theory of diagnosis for systems that can be affected by the actions of agents. Specifically, we examine the task of conjecturing diagnoses to explain {\it what happened} to a system, given a theory of system behaviour and some observed (aberrant) behaviour. We characterize what happened by introducing the notion of explanatory diagnosis in the language of the situation calculus. Explanatory diagnoses conjecture sequences of actions to account for a change in system behaviour. As such, we show that determining an explanatory diagnosis is analogous to classical AI planning with state constraints and incomplete knowledge. The representation scheme we employ provides an axiomatic solution to the frame, ramification and qualification problems. Exploiting this representation, we show that determining an explanatory diagnosis can be achieved by regression followed by theorem proving in the database describing what is known of the initial state of our system. Further, we show that by exploiting features inherent to diagnosis problems, we can simplify the diagnosis task. blems, we can simplify the diagnosis task. |
| Author | Sheila A. McIlraith + |
| Bibtype | techreport + |
| Institution | Knowledge Systems, AI Laboratory + |
| Key | KSL-98-21 + |
| Month | July + |
| Number | KSL-98-21 + |
| Tag | Computer science + |
| Title | Explanatory Diagnosis: Conjecturing actions to explain observations + |
| Tr id | KSL-98-21 + |
| Year | 1998 + |
