EXPERT SYSTEMS

1
EXPERT SYSTEMS
GROUP F
2
Overview
EXPERT SYSTEMS

• What are Expert Systems?
• Object-Based Knowledge (Rhonda Riggins)
• Frame-Based Knowledge (JoAnna Kim)
• Rule-Based Knowledge (Abbie Hoffman)
• Case-Based Reasoning (Mary Mabie)
• Advantages/Disadvantages (Dan Andrzejewski)

3
What Are Expert Systems?
EXPERT SYSTEMS

• Def Knowledge-based info. system that uses its
  knowledge about a specific, complex application
  area to act as an expert consultant to end users
• Components
• Knowledge Base
• 1) Information
• 2) Reasoning Procedures
• Software Resources
• 1) Inference Engine

4
Object-Based Knowledge
EXPERT SYSTEMS
5
Object-Based Knowledge
EXPERT SYSTEMS

• Definition
• Knowledge represented as a network of objects
• Comprised of

Attribute
Value
Object
Skin
Smooth
Texture
Rough
6
Object-Based Knowledge
EXPERT SYSTEMS
High
Normal
Temperature
Smooth
Low
Texture
SKIN CONDITION
Clammy
Generalized
Pale
Rough
Color
Measles
Chicken Pox
Patchy
Reddish
7
Object-Based Knowledge
EXPERT SYSTEMS

• Application Category
• Design/configuration- Systems that help
  configure equipment components, given existing
  constraints.
• Selection/classification- Systems that help
  users choose products or processes, often from
  among large or complex sets of alternatives.

8
Frame-Based Knowledge
EXPERT SYSTEMS
9
Frame-Based Knowledge
EXPERT SYSTEMS

• represents knowledge in a hierarchical structure
  or a network of frames.
• A frame describes all the knowledge about one
  particular object or concept.
• Composed of
• Slots- set of attributes or characteristics that
  describe the object represented by the frame
• Facets (or subslots)- describe some knowledge or
  procedure about the attribute in the slot.

10
Frame-Based Knowledge
EXPERT SYSTEMS
Physical Condition Frame
Skin Condition Frame
Heart Condition Frame
Texture Condition Frame
SLOT Temperature -High Facet
-Low Celsius -Normal SLOT Texture
-Rough general, patchy -Smooth
Range
Values
Facet
11
Frame-Based Knowledge
EXPERT SYSTEMS

• Application Category Selection/Classification
• Frames also allow users to choose processes from
  among large or complex sets of alternatives
• Material selection
• Delinquent account identification
• Information classification

12
Rule-Based Knowledge
EXPERT SYSTEMS
13
Rule-Based Knowledge
EXPERT SYSTEMS

• Bases decisions on rules and facts typically
  represented in the form of IF/THEN statements.
• These rules are a type of knowledge known as
  heuristics or "rules of thumb" that experts would
  use in their day-to-day work.
• By combining the applicable rules, the expert
  system reaches a conclusion or makes a diagnosis.

14
Rule-Based Knowledge
EXPERT SYSTEMS

• Example
• IF the following conditions are true
• ltPulse Oximetry is gt 92gt and
• ltRespiratory Rate gt 12gt and
• ltFIO2 lt 28gt
• THEN consider the following
• ltCPAP Trial for 1 Hour at 21 FIO2gt
• ltFailure to Weangt

15
Rule-Based Knowledge
EXPERT SYSTEMS

• Application Category Decision Management
  Systems that appraise situations or consider
  alternatives and make recommendations based on
  criteria supplied during the discovery process.
• Criteria Supplied fraction of inspired oxygen,
  airway passages, respiratory rate, heart rate,
  blood pressure, body temperature, etc.
• Recommendation Mechanical Ventilation
  Weaning Strategy

16
Case-Based Reasoning
EXPERT SYSTEMS
17
Case-Based Reasoning
EXPERT SYSTEMS

• Definition
• Representing knowledge in an expert systems
  knowledge base in the form of cases, i.e.,
  examples of past performance, occurrences, and
  experiences (textbook, p. 487)

18
Case-Based Reasoning
EXPERT SYSTEMS

• Case base - set of cases
• Index library - search retrieve most similar
  cases
• Similarity metrics - measures how similar current
  problem to past cases
• Adaptation module - creates solution
• modifies the solution (structural adaptation)
• creates new solution (derivational adaptation)
• (Carol Brown Daniel OLeary, 2000, Introduction
  to Artificial Intelligence and Expert Systems,
  IV-3, http//www.bus.orst.edu/faculty/brownc/es_tu
  tor/es_tutor.htm)

19
Case-Based Reasoning
EXPERT SYSTEMS

• Application Category
• Decision Management
• systems that appraise situation or consider
  alternatives and make recommendations based on
  criteria supplied during discovery process
• Diagnostic/troubleshooting
• Systems that infer underlying causes from
  reported symptoms and history

20
Advantages vs Human Experts
EXPERT SYSTEMS

• Permanent Information
• Consistent Recommendations
• Easily Reproducible
• Efficient Operation/Development
• Complete Information
• Timeliness
• Other Advantages

21
Disadvantages vs Human Experts
EXPERT SYSTEMS

• Lack of Common Sense
• Lack of Creativity
• Manually Updated Learning
• Sensory Experience
• Problem Recognition

22
ANY QUESTIONS?
EXPERT SYSTEMS