UNIT-5
1. With a neat sketch explain the architecture of expert system.
Definition:
An expert system is an AI software that is designed to solve complex
problems by using knowledge from its knowledge base to provide
decision-making ability like a human-expert.
Architecture of expert system:
The process of building an expert system is often called knowledge
engineering.
Key Components of Expert System:
The components of an expert system typically include the knowledge
base, inference engine with explanation module, user interface, and
learning module.
a) Knowledge Base:
The knowledge base stores/represents facts and rules.
Factual Knowledge − Based on facts
for example, Ramesh is an analyst.
Heuristic Knowledge (rules of thumb) − Based on experience.
e.g. IF Ramesh is an analyst, THEN he needs a workstation.
b) Inference Engine (Rules of Engine):
It applies inference rules to the knowledge base to derive
conclusions or recommendations.
With the help of an inference engine, the system extracts the
knowledge from the knowledge base.
The following modes are used by the inference engine to generate
solutions:
1) Forward Chaining: This is a data-driven reasoning approach
where the system starts with available data and applies rules to
deduce new information until a conclusion is reached.
2) Backward Chaining: This is a goal-driven reasoning approach
where the system starts with a goal or conclusion and works
backward to find evidence or data supporting that goal. It is
useful for diagnostic systems that need to identify the causes of
observed symptoms.
c) User Interface
It is an interface that helps a non-expert user to interact with the
expert system to find a solution. Example: Text-Based interface,
Graphical User Interface (GUI)
d) Knowledge Acquisition:
The function of this component is to allow the expert system to
acquire more and more knowledge from various sources and store it
in the knowledge base.
e) Learning Module (Optional):
Some expert systems incorporate a learning mechanism to improve
performance. Examples include supervised, unsupervised,
reinforcement learning.
f) Explanation Module:
This module helps the expert system to give the user an explanation
about how the expert system reached a particular conclusion.
2. How do meta-knowledge and heuristics work together in real-world
scenarios.
Relationship Between Meta-Knowledge and Heuristics:
In expert systems, meta-knowledge and heuristics often work hand in
hand.
Meta-knowledge can be used to decide when and how to apply
certain heuristics to solve a problem.
Meta-knowledge monitors and optimizes the whole process,
ensuring correct usage of both rule-based knowledge and heuristic
shortcuts.
Real-world scenarios:
a) MYCIN (Medical Diagnosis):
An early expert system used in medicine, MYCIN employed
heuristic rules for diagnosing bacterial infections and prescribing
treatments.
Its meta-knowledge allowed it to assess the certainty of its
conclusions, providing users with confidence scores based on the
available evidence.
b) XCON (Computer Configuration):
XCON was used by Digital Equipment Corporation to configure
computer systems.
It utilized heuristics for selecting compatible hardware
components based on user requirements and meta-knowledge
to manage conflicts between different rules.
3. Explain the key components of an expert system shell and discuss the
benefits and limitations.
Definition:
Expert system shell in AI provides a user-friendly software
environment/toolkit to knowledge engineers for building an expert
system.
Key Components of Expert System Shells:
1. Knowledge Base: Stores facts and rules related to the domain of
expertise.
2. Inference Engine: The core component that applies logical rules to
the knowledge base to deduce new information or make decisions.
Two main reasoning methods:
a) Forward chaining: Starts with known facts and applies rules to
infer new facts until a goal is reached.
b) Backward chaining: Starts with a goal and works backward to see
if there is evidence to support that goal.
3. User Interface: Provides interaction between the user and the
system.
4. Explanation Facility: Explains how the system reached a particular
conclusion
5. Knowledge Acquisition Tool: Assists in adding or updating the
knowledge base.
Advantages of Using Expert System Shells:
o Pre-built Components: Saves time by providing ready-made
modules
o No Need for Low-Level Programming
o Flexibility: Can be adapted to various domains like medical
diagnosis.
Limitations:
o Requires expert knowledge to define rules or facts.
o Performance Issues: Large knowledge bases can slow down
processing.
o Lack of Creativity: Shells are limited to the rules and knowledge.
5. Examine the architecture and functionality of the MYCIN expert system
in medical diagnosis.
Architecture of MYCIN:
MYCIN was developed in the early 1970s at Stanford University by
Edward Shurtleff as a rule-based expert system focused on medical
diagnosis and treatment recommendations for bacterial infections.
Key Components of MYCIN Architecture:
Knowledge Base: A set of rules encoding domain expertise.
Inference Engine: Performs backward chaining to apply rules to the
current problem.
User Interface: Interacts with the user through questions and
answers.
Explanation System: Provides the reasoning behind the system's
conclusions.
Certainty Handling: Manages uncertain data using certainty factors.
Purpose:
It was designed to assist physicians by providing recommendations on
which antibiotics to prescribe for patients, based on a set of symptoms
and laboratory results.
Key Features:
Rule-based inference engine (over 600 rules).
Certainty factor-based reasoning to handle incomplete or uncertain
data.
Natural language processing for user interaction
6. Discuss the role of the DART in expert system
DART (Decision Analysis and Resolution Tool):
DART is designed for decision support within expert systems, focusing
on multi-criteria decision analysis and resolution.
It is more of a methodology or tool than a standalone expert system,
often integrated into broader decision support systems or expert
system shells.
Functionality:
a. Provides structured decision-making by analysing different
alternatives based on weighted criteria.
b. Supports decision-making processes in domains where multiple
options need to be evaluated against performance, risk, and other
factors.
7. Compare and contrast MYCIN, DART, and XCON expert systems
Expert system shells like:
1. DART (Decision Analysis and Resolution Tool) is designed for multi-
criteria decision analysis and resolution within expert systems.
2. MYCIN, a rule-based expert system, designed to assist physicians by
providing treatment recommendations and antibiotics for bacterial
infections.
3. XCON (eXpert CONfigurer): XCON was designed to help DEC’s sales and
manufacturing teams configure the complex hardware components of
their VAX computers.
Comparison of DART, MYCIN, and XCON is given by:
Feature DART MYCIN XCON (R1)
Rule-Based Rule-Based
Decision Analysis
Type Medical Expert Configuration Expert
and Resolution Tool
System System
Primary General Decision Medicine Manufacturing and
Feature DART MYCIN XCON (R1)
(Infectious Sales (Computer
Domain Support
Diseases) Systems)
Backward
Inference Multi-Criteria Forward Chaining
Chaining (Rule-
Method Decision Making (Rule-Based)
Based)
Handling Probabilistic and Limited (Focuses on
Certainty Factors
Uncertainty multi-criteria Models Configuration)
Diagnosis and
Primary Use Decision VAX Computer
Treatment of
Case Optimization Configuration
Infections
Enhanced decision-
Pioneering Successful commercial
Impact making in various
medical AI AI system
domains
