Knowledge Representation

1
Knowledge Representation
2
Knowledge Representation

• A knowledge base can be organised in several
  different configurations to facilitate fast
  inferencing
• Knowledge Representation Schemes
• designed so that facts and other knowledge
  contained within then can be used in reasoning
• Majority are based on models, rules and frames

3
Knowledge Representation Schemes

• Scripts
• Decision Tables
• Lists
• Models
• Semantic Networks
• Production Rules
• Frames

4
Scripts

• Describes a sequence of events
• Elements of a typical script
• Props
• objects that are used in the sequence of events
  that occur
• Roles
• the people involved in the script
• Scenes
• describe the actual sequence of events that occur

5
Scripts

• Useful form of Knowledge representation
• so many stereotypical situations
• Used in case-based reasoning

Decision Tables

• Also called induction tables or knowledge maps
• Knowledge is collected in knowledge acquisition
  sessions

6
Lists

• A written series of related items
• names of peoples
• shopping list
• things to do that week
• Can be used to represent hierarchical knowledge
• objects are grouped, or graded according to rank

7
Modelling

• Decision trees
• Composed of nodes representing goals and links
  representing decisions
• All human decision making are based on models of
  aspects of the world
• complex, incomplete, difficult to ascertain
• Computer models are more precise and consistent

8
Models

• They are written clearly so that they can be
  understood and criticised
• They always draw error free logical conclusions
  from their assumptions
• They can be changed and tested very quickly

9
When to use Models

• To organise thoughts and ideas
• to gain understanding
• to suggest new ideas and perceptions
• to aid in the development of theories
• to predict the consequence of actions
• to aid in the communication of mental methods

10
Expert Builder

• Enables the construction of models based on rules
• The rules are represented in the diagram
• Diagram can be saved as rules and viewed

11
Semantic Networks

• Semantic networks are graphic depictions of
  knowledge composed of nodes and links
• They show hierarchical relationships between
  objects
• Objects can be any physical item
• book, car, desk or even a person
• Objects can also be
• concepts, events, actions
• FIGURE 14.2

12
Production Rules

• Knowledge is presented as production rules
• These take the form of condition-action pairs
• IF condition occurs
• THEN action
• Each rule implements a chunk of expertise
• The quickly become interdependent

13
Rules - Advantages

• Easy to understand
• Inferences are easily derived
• Modifications are relatively easy

Rules - Limitations

• Usually a great number of rules are required
• Difficult to maintain
• Difficult to represent all knowledge in terms of
  rules

14
Frames

• Structure that includes all the knowledge about a
  particular object
• Each Frame describes one object
• what it is
• How it behaves
• An application of object-oriented programming for
  AI and ES
• Widely used

15
Uncertainty

• Refers to a wide range of situations where the
  relevant information is deficient
• information is partial
• information is not fully reliable
• conflicting information
• approximate information
• Probability may be included in the rules
• IF the interest rate is increasing, THEN the
  price of stocks will decline (80 percent
  probability)