Definitions and Example:
But what exactly is an expert system? Rosenman defined it as:
an automated reasoning system that attempts to mimic the performance of
the human expert.
This, of course, poses the question of what exactly is an expert. Michie avoided the
comparison with the human:
Expert System = Knowledge Base + Inference Engine
They have been successful in many fields. In general there are two types of problems:
(a) select one of several hypotheses (e.g. diagnosis and advice); and
(b) make a solution
that meets requirements (e.g. design and planning). MYCIN was used for medical
diagnosis (for a specific group of diseases). PROSPECTOR evaluated geological sites
for commercial development. DENDRAL helped to infer a molecule's structure given
its chemical formula and other data. XCON was developed to customize a network
system to meet the customer's needs
Design of An Expert System
Like a human expert, an expert system is expected to
o be specialist : know facts and procedural rules
o use heuristics: interpolate from known facts
o justify its conclusions: to establish credibility and confidence. The user can ask:
how do you know a particular fact? why do you ask a particular question?
o be able to learn: be able to absorb new knowledge and apply it
o estimate the reliability of its answer.
There are many advantages to the separation and modularity. For example, the
knowledge base and inference engine being separate allows one to inspect the knowledge of the system, not just its responses to particular questions.
But what exactly is an expert system? Rosenman defined it as:
an automated reasoning system that attempts to mimic the performance of
the human expert.
This, of course, poses the question of what exactly is an expert. Michie avoided the
comparison with the human:
Expert System = Knowledge Base + Inference Engine
They have been successful in many fields. In general there are two types of problems:
(a) select one of several hypotheses (e.g. diagnosis and advice); and
(b) make a solution
that meets requirements (e.g. design and planning). MYCIN was used for medical
diagnosis (for a specific group of diseases). PROSPECTOR evaluated geological sites
for commercial development. DENDRAL helped to infer a molecule's structure given
its chemical formula and other data. XCON was developed to customize a network
system to meet the customer's needs
Design of An Expert System
Like a human expert, an expert system is expected to
o be specialist : know facts and procedural rules
o use heuristics: interpolate from known facts
o justify its conclusions: to establish credibility and confidence. The user can ask:
how do you know a particular fact? why do you ask a particular question?
o be able to learn: be able to absorb new knowledge and apply it
o estimate the reliability of its answer.
There are many advantages to the separation and modularity. For example, the
knowledge base and inference engine being separate allows one to inspect the knowledge of the system, not just its responses to particular questions.
|
|