1. Knowledge Representation

2. Introduction Knowledge is organized in knowledge base in one or more configurations. A good knowledge representation ‘naturally’ represents the problem domain A variety of knowledge representation schemes exist Can be programmed with programming languages Contents of knowledge base can be manipulated by inference mechanism 2

3. Knowledge Representation Schemes:  Logical Form  Propositional Logic  Predicate Calculus  Production Rules  Semantic Network  Scripts  Frames  List  Decision Table & Decision Tree  O-A-V Triplet 3 Introduction

4. A statement that can be either true or false Once known, it becomes a premise that can be used to derive new propositions or inferences Rules are used to determine the truth (T) or falsity (F) of the new proposition To create compound propositions use connectives, such as AND, OR, NOT, IMPLIES Elementary propositions or statements represented with statement letters A, B, … 4 Propositional Logic

5. Simple Proposition:  Statement: A = The mail carrier comes Monday thru Friday  Statement: B = Today is Sunday  Conclusion: C = The mail carrier will not come today Propositional logic: limited in representing real-world knowledge 5 Propositional Logic

6. A proposition can be broken down into a predicate and the subject of the predicate A predicate is a logic function; possible values are T and F Predicate calculus uses variables and functions of variables in a symbolic logic statement A proposition Smith is a worker will be written as: is-a(smith, worker) where is-a is the name of the predicate 6 Predicate Calculus

7. Predicate calculus is the basis for Prolog (PROgramming in LOGic) Prolog Statement Examples  comes_on(mail_carrier, Monday).  likes(jay, chocolate). (Note - the period “.” is part of the statement) The values of the argument of a predicate depend on the context Arguments of a predicate could take both qualitative and quantitative values 7 Predicate Calculus

8. One of the most common modes of knowledge representations in expert systems Condition-Action Pairs  IF this condition (or premise or antecedent) occurs,  THEN some action (or result, or conclusion, or consequence) will (or should) occur IF the stop light is red AND you have stopped, THEN a right turn is OK Two types of rules: declarative and procedural Metarules: Rules about rule 8 Production Rules

9. Each production rule in a knowledge base represents an autonomous chunk of expertise When combined and fed to the inference engine, the set of rules behaves synergistically Rules can be viewed as a simulation of the cognitive behavior of human experts Rules represent a model of actual human behavior 9 Production Rules

10. Forms of Rules  IF premise, THEN conclusion  IF your income is high, THEN your chance of being audited by the IRS is high  Conclusion, IF premise  Your chance of being audited is high, IF your income is high  Inclusion of ELSE  IF your income is high, OR your deductions are unusual, THEN your chance of being audited by the IRS is high, OR ELSE your chance of being audited is low 10 10 Production Rules

11. More Complex Rules  IF credit rating is high AND salary is more than $30,000, OR assets are more than $75,000, AND pay history is not "poor," THEN approve a loan up to $10,000, and list the loan in category "B.”  Action part may have more information: THEN "approve the loan" and "refer to an agent" 11 11 Production Rules

12. Production Rules: Knowledge and Inference Rules Common Types of Rules  Knowledge rules, or declarative rules, state all the facts and relationships about a problem  Inference rules, or procedural rules, advise on how to solve a problem, given that certain facts are known  Inference rules contain rules about rules (metarules)  Knowledge rules are stored in the knowledge base  Inference rules become part of the inference engine 12 12

13. Production Rules: Major Advantages Easy to understand (natural form of knowledge) Easy to derive inference and explanations Easy to modify and maintain Easy to combine with uncertainty Rules are frequently independent 13 13

14. Complex knowledge requires many rules Builders like rules (hammer syndrome) Search limitations in systems with many rules 14 Production Rules: Major Limitations

15. Production Rules: Major Characteristics First Part (IF …) Second Part (THEN …) Names1.Premise 2.Antecedent 3.Situation 1.Conclusion 2.Consequence 3.Action SizeCan have many IFsUsually has one conclusion StatementsORIf any condition is true, the conclusion is true ANDIf all conditions are true, the conclusion is true 15

16. A simple and graphic depiction of knowledge consisting of nodes and links Nodes represent objects, concepts, events, actions, attribute of an object Nodes are interconnected by links or arcs; two frequent associations are is-a and has-a Semantic network can show inheritance; is-a association shows the specialization 16 16 Semantic Network

17. Shortcomings  Does not distinguish between the class of an object and a particular object  Does not have appropriate structure for accommodating the complex associations of the attributes of an object with other objects  The presentation and structure for complex systems could become unmanageable 17 17 Semantic Network

18. Age 22 years Person Teacher Debra IS-AIS-A HAS HAS IS-A 18

19. Frames Frame: Data structure that includes all the knowledge about a particular object Knowledge organized in a hierarchy for diagnosis of knowledge independence Form of object-oriented programming for AI and ES. Each Frame Describes One Object 19 19

20. Provide a concise, structural representation of knowledge in a natural manner Frame encompasses complex objects, entire situations or a management problem as a single entity Frame knowledge is partitioned into slots Slot can describe declarative knowledge or procedural knowledge Hierarchy of Frames: Inheritance 20 20 Frames

21. Terminology for Frames:  Default  Demon  Facet  Hierarchy of frames  If added  Instance of  Instantiation  Master frame  Object  Range  Slot  Value (entry) 21 21 Frames

22. Capabilities of Frames:  Ability to clearly document information about a domain model; for example, a machines and the associated attributes  The related ability to constrain the allowable values that an attribute can take on  Modularity of information -> Ease of system expansion and maintenance  More readable and consistent syntax for referencing domain objects in the rules  Access to a mechanism that supports the inheritance of information down a class hierarchy Frames

23. Frames OBJECT: Marketable Securities PARENTCLASS: MEMBERSLOTS: MEMBERSLOTS: OWNSLOTS: OWNSLOTS: Issued-by Issued-by Risk Risk Marketable Marketable LOCALSLOTS: LOCALSLOTS: Total_Portfolio_Value Total_Portfolio_Value OBJECT: Bonds PARENTCLASS: Marketable Securities MEMBERSLOTS:METHOD: MEMBERSLOTS:METHOD: RISK: LowRisk_Eval RISK: LowRisk_Eval MARKETABLE: Yes MARKETABLE: Yes OWNSLOTS: OWNSLOTS: Face_Value:=1000 Face_Value:=1000 Face_Interest_Value Face_Interest_Value Maturity Maturity LOCALSLOTS: LOCALSLOTS: Bonds_Market_Value Bonds_Market_Value OBJECT: 9245631 IBM Class C MEMBERCLASS: Bonds MEMBERSLOTS: MEMBERSLOTS: Issued-by: IBM Issued-by: IBM Face_Interest_Value:=6.875 Face_Interest_Value:=6.875 Maturity:=31 DEC.2007 Maturity:=31 DEC.2007 LOCALSLOTS: LOCALSLOTS: Market_Value:=1276 Market_Value:=1276

24. Scripts Knowledge Representation Scheme Describing a Sequence of Events Elements include  Entry Conditions  Props  Roles  Tracks  Scenes

25. Lists Written Series of Related Items Normally used to represent hierarchical knowledge where objects are grouped, categorized or graded according to  Rank or  Relationship

26. Decision Tables (Induction Table) Knowledge Organized in a Spreadsheet Format Attribute List Conclusion List Different attribute configurations are matched against the conclusion

27. Decision Trees Related to tables Similar to decision trees in decision theory Can simplify the knowledge acquisition process Knowledge diagramming is frequently more natural to experts than formal representation methods

28. O-A-V Triplet Objects, Attributes and Values O-A-V Triplet Objects may be physical or conceptual Attributes are the characteristics of the objects Values are the specific measures of the attributes in a given situation O-A-V triplets

29. O-A-V Triplet ObjectsAttributesValues Class-roomSize12 X 15, 10 X 12 etc. Admissions to MBA program GMAT score500, 510, 520 …. Admissions to MBA program GPA3.0, 3.3, 3.5 …

30. Considerations for Evaluating a Knowledge Representation Naturalness, uniformity and understandability Degree to which knowledge is explicit (declarative) or embedded in procedural code Modularity and flexibility of the knowledge base Efficiency of knowledge retrieval and the heuristic power of the inference procedure

31. No single knowledge representation method is ideally suited by itself for all tasks (Table 14.5) Multiple knowledge representations: each tailored to a different subtask Production Rules and Frames works well in practice Object-Oriented Knowledge Representations  Hypermedia Considerations for Evaluating a Knowledge Representation

32. Knowledge Representation Schemes (Turban and Aronson, 1998) SchemeAdvantagesDisadvantages Production Rules  Simple syntax  Easy to understand  Easy to modify  Modular  Inefficient for large systems  Difficult to follow hierarchies  Can not be used to represent all knowledge Semantic Networks  Easy to follow hierarchies  Easy to trace associations  Meaning attached to nodes may be ambiguous  Difficult to program Frames  Easy to set up slots for new properties and relations  Easy to include default information  Easy to detect missing values  Difficult to program  Difficult to make inference Formal Logic  Precision  Completeness  Inefficient with large knowledge bases