Robotics Reference
In-Depth Information
Figure 45.
The Case Based Reasoning Cycle, from “Case-Based Reasoning: Founda-
tional Issues, Methodological Variations, and System Approaches” by A. Aadmodt and
E. Plaza (
AI Communications
vol. 7, no. 1, 1994, pages 39-59) (Courtesy of Agnar
Aamodt)
domain, which is often useful knowledge in the Case Based Reasoning
process. For example, when a program is diagnosing a medical condition
in a patient by retrieving and reusing cases of previous patients, a model
of anatomy is clearly useful, together with a knowledge of relationships
indicating how various pathological states are connected.
Case Based Reasoning is now the most popular form of reasoning
employed within the field of AI. One research project currently showing
great promise is the Sherlock Holmes program being developed by Jeroen
Keppens at the Joseph Bell Centre for Forensic Statistics and Legal Rea-
soning at Edinburgh University. A known problem faced by investigators
is that police officers often jump to premature conclusions. In general,
they have a tendency to decide at a very early stage of an investigation
on the most likely suspects, decisions that bias the officers' subsequent
investigations. In the case of a suspicious death, for example, it is human
nature for police officers to latch onto a particular theory and then at-
tempt to confirm it, an approach that might well divert them from what
actually happened.
The Sherlock Holmes program uses Case Based Reasoning to decide
on the most likely cause in instances of suspicious death. The program is
designed to highlight less obvious lines of inquiry that police detectives
might overlook, taking an unbiased overview of all the available evidence
