Expert Systems - MYCIN

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

Expert Systems

(DS 322)
Lecture Objectives
• Major Features of MYCIN
• A review of MYCION features provides insight into developing a rule-based
system
• Main objective is not primarily concerned with what MYCIN is, but rather
why it is important to expert system development in general

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Background
• Developed at Stanford University
In the year of 1976
To aid physicians in diagnosing and testing patients
With infectious blood diseases caused by bacteremia(Bacteria in blood)
Meningitis(Bacterial disease that causes inflammation of the membrane surrounding
the brain and spinal cord)
MYCIN took approximately 20 person-years to complete
• MYCIN is a rule-based expert system that uses backward chaining
“The MYCIN is a computer-based consultation system designed to assist physicians
in the diagnosis of and therapy selection for patients with bacterial infections”
 Incorporates approximately 500 rules
 Written in INTERLISP
 A dialect of the LISP programming language

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Motivation
• Need to react quickly to a possibly life-threatening situation caused
difficulties for the physician
• System to be unobtrusive and accommodating to physicians
• Avoid overdose of antibiotics
• Shortage of expertise in diagnosing and treating infectious blood disease

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN
• Design Consideration s for Consultation Programs
Useful
Educational when appropriate
Able to explain its advice
Able to understand and respond to simple questions stated in natural language
Able to acquire new knowledge, either through experience or by being told
Easy modified
• Program should be useful
 The need for the assistance that the program provides
 The reliability of the advice
 The mechanics for accessing the machine and retrieving the desired information.
• Program should be Educational when appropriate

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN
• Program Should be Able to Explain its Advice
• Program Should be Able to Understand Questions
• Program Should be Able to Acquire Knowledge
• Program Knowledge-base should be easily modified
• Acceptability Criteria
Modularity to insure straightforward modification(LISPENGLISH)
Ability to acquire new knowledge from experts
Ability to understand the Questions
Educational Capabilities
General Usefulness
Easy modified

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN

Physician
Expert

Data about
Advice Explanations
patient

Physician
Non Expert

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN Physician
Expert

Facts Production
About Rules
The For Making
Domain Inferences
Static
Knowledge
General Reasoning
Question Rule Interpreter Status
Answerer Checker

Data Conclusions Record of


About About Conclusions
Patient Patient
Dynamic
Knowledge

Advice Knowledge –based


Explanations at any time Explanations during consultation Production System
MYCIN (Natural Language) (“Why”, “How”, “Explain” )

Physician
User
Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN Physician
Expert

Facts Production
About Rules
The For Making
Domain Inferences
Static
Knowledge
General Reasoning
Question Rule Interpreter Status
Answerer Checker

Data Conclusions Record of


About About Conclusions
Patient Patient
Dynamic
Knowledge

Advice Knowledge –based


Explanations at any time Explanations during consultation Production System
MYCIN (Natural Language) (“Why”, “How”, “Explain” )

Physician
User
Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN Physician
Expert

Facts Production
About Rules
The For Making
Domain Inferences
Static
Knowledge
General Reasoning
Question Rule Interpreter Status
Answerer Checker

Data Conclusions Record of


About About Conclusions
Patient Patient
Dynamic
Knowledge

Advice Knowledge –based


Explanations at any time Explanations during consultation Production System
MYCIN (Natural Language) (“Why”, “How”, “Explain” )

Physician
User
Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN Physician
Expert

Facts Production
About Rules
The For Making
Domain Inferences
Static
Knowledge
General Reasoning
Question Rule Interpreter Status
Answerer Checker

Data Conclusions Record of


About About Conclusions
Patient Patient
Dynamic
Knowledge

Advice Knowledge –based


Explanations at any time Explanations during consultation Production System
MYCIN (Natural Language) (“Why”, “How”, “Explain” )

Physician
User
Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN Physician
Expert

Facts Production
About Rules
The For Making
Domain Inferences
Static
Knowledge
General Reasoning
Question Rule Interpreter Status
Answerer Checker

Data Conclusions Record of


About About Conclusions
Patient Patient
Dynamic
Knowledge

Advice Knowledge –based


Explanations at any time Explanations during consultation Production System
MYCIN (Natural Language) (“Why”, “How”, “Explain” )

Physician
User
Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN Physician
Expert

Facts Production
About Rules
The For Making
Domain Inferences
Static
Knowledge
General Reasoning
Question Rule Interpreter Status
Answerer Checker

Data Conclusions Record of


About About Conclusions
Patient Patient
Dynamic
Knowledge

Advice Knowledge –based


Explanations at any time Explanations during consultation Production System
MYCIN (Natural Language) (“Why”, “How”, “Explain” )

Physician
User
Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN

Consultation
program

Dynamic Static Factual


Patient Explanation
and judgmental
data program
knowledge

Knowledge
acquisition
program

Infectious Diseases
Expert

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN

Consultation
program

Dynamic Static Factual


Patient Explanation
and judgmental
data program
knowledge

Knowledge
acquisition
program

Infectious Diseases
Expert

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN

Consultation
program

Dynamic Static Factual


Patient Explanation
and judgmental
data program
knowledge

Knowledge
acquisition
program

Infectious Diseases
Expert

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
Design Considerations for MYCIN

Consultation
program

Dynamic Static Factual


Patient Explanation
and judgmental
data program
knowledge

Knowledge
acquisition
program

Infectious Diseases
Expert

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.
References
• Expert Systems Design and Development, John Durkin, Macmillan Publishing Company, 1994.
• Giarratano J., Riley G., Expert Systems, Principles and Programming, Course Technology; 4th
edition, 2004.

Expert Systems [DS 322] Dr. P. Pavan Kumar, Asst. Prof., Dept. of DS & AI, IcfaiTech, Hyd.

You might also like