Tatiana Gavrilova

1
Tatiana Gavrilova
2. Knowledge Engineering 2.1. Knowledge
Engineering Structure 2.2. Knowledge
Elicitation 2.3. Practical Techniques and
Methods 2.4. Knowledge Structuring
2
Tatiana Gavrilova
2.1. Knowledge Engineering Structure
3
Main Stages of Intelligent Systems Development

• Exploratory design ? requirements analysis
• Data and Knowledge Acquisition
• Informations Structuring (Conceptualisation)
• Formalization and business-processes models
  development
• Pilot Prototype Development
• Testing

4
Knowledge Engineering

• Knowledge Engineering is a branch of Artificial
  Intelligence, which deals with the problems of
  knowledge
• elicitation, structuring and formalization
• targeted at knowledge based systems development.
  It basically deals with knowledge preprocessing

5
Knowledge Engineering Structure
Knowledge Engineering
Structuring
Knowledge capture
Knowledge representation
Object-or. approach
Process approach
Modular
Net
Direct
Computer-aided
Logic
Frames
Acquisition
Semantic Networks
Elicitation
Discovery
Productions
6
Knowledge Engineering

• Knowledge Capture is the process of acquiring
  knowledge from its source/ expert/ group of
  experts/literature (books, documents, papers).
• Knowledge structuring is the process of
  organizing the acquired knowledge into the
  knowledge fields. Knowledge fields shows the main
  concepts of the domain and relationship between
  them.
• Knowledge Formalizing is the process of
  transformation of knowledge fields into knowledge
  base, which are xpressed in knowledgerepresentatio
  n languages.

7
Where can you find data and knowledge?
8
Misinterpretation problem
9
Tatiana Gavrilova
2.2. Knowledge Elicitation
10
Tatiana Gavrilova
????-?????-???????
Knowledge Elicitation Problems
Gnosiological
aspect
Psychological
aspect
Linguistic
aspect
Methodical
aspect
11
Tatiana Gavrilova
Expert Analyst interaction
100
80
70
60
Can present in verbal form
40
30
Has expressed
Has listened
Stored in memory
Knows
Realises
expert
analyst
12
Tatiana Gavrilova
Three levels of psychological aspect
13
Non-verbal Components Consideration

• Pose (straight, openly, small droop, limply,
  visual contact between evasion and stare
  drilling) Egan, 1990
• Mimicry(if it necessary you may be intolerable,
  but you must smile)
• Head movement (nods, back -arrogance, forward
  -dependence)
• Dress-code (better under-dressed, than
  over-dressed, expert corporative standard
  consideration Berceley vs Stanford )

14
Test N 9. Stanislavskys warming-up

• 8 expressions types
• Joy
• Sorrow
• Disgust
• Anger
• Fear
• Interest
• Disregard, boring
• Pain

15
Tatiana Gavrilova
????-?????-???????
Three levels of linguistics aspect
Linguistic aspect
Conceptual Structure
Common Code
User Lexicon
16
Cultural aspect

• Dont speak to the driver (USA)
• It is strictly forbidden for passengers to speak
  to the driver (Germany)
• You are graciously requested to refrain from
  speaking to the driver (England)
• What have you got to gain by speaking to the
  driver (Scotland)
• Dont answer the driver (Italy)

17
Test N10(40 objects properties)
18
Tatiana Gavrilova
2.3. Practical Techniques and Methods
19

Knowledge Elicitation Methods
Communicative
Textual

• documents analysis
• reference books analysis

Group
Individual

• round table
• discussion
• brain storm
• games

Active
Passive

• observation
• verbal reports
• lections

• interview
• questionnaire
• design
• role games

20
Tatiana Gavrilova
????-?????-???????
Passive Methods Comparison
21
Tatiana Gavrilova
????-?????-???????
Training for Interview
Training
Specific
Common
Psychological
Special
System Approach
Branch Study
Interview Skills
Communication Technology
Common Culture
Cognitive Psychology
22
Questions Classification

• On form (open, closed)
• Personal, impersonal
• Direct, indirect
• Extra (buffer, exploratory, control)

23
Test N11(Madrid-Nice)
24
Test N11(Secret service or reconnaissance)
25
Tatiana Gavrilova
????-?????-???????
Active Methods Comparison
26
Test N12(Interview)
27
Tatiana Gavrilova
????-?????-???????
Group Methods Comparison
28
Tatiana Gavrilova
????-?????-???????
Expert games classification
29
Tatiana Gavrilova
????-?????-???????
Expert games comparison
30
Tatiana Gavrilova
Analysts common mistakes
31
Test N13(Mafia)
32
Text Compression

• Look-up for understanding and new words
  selection
• Reading and SKW (Set of Key Words) copying
• Reading and SKW linking in the knowledge field
• Reading and field checking

33
Test N14(Knowledge acquisition from text)
34
2.4. Knowledge structuring
35
Three Modelling Levels
Goals and Requirements Level
Conceptual Model Level
Specification Level
36
Minimal Structuring Algorythm

• R
• X Y
• Goals or outputs Y
• Glossary of input factors X
• Conceptual Structure (Sc - ontology)
• Functional Structure R( Sf - decision table or
  reasoning model R)

37
Test N14(Structuring)
38
Test N15(Associative chains)
39
Constructing of Associates

• Cloud-table (cloud-sky, sky-ground, ground
  -house, house-furniture, furniture-table)
• Table - moon
• Mushroom- aeroplane
• Ice hair-dryer
• Firewood dance
• Boat - watch

40
Object-structural approach bases on
Synthesis of several approaches
OOP
Knowledge management
Structural analysis


41
Object-Structured Analysis

• s1 WHAT_FOR-Knowledge Strategic Analysis of the
  System, its Intention and Functioning.
• s2 WHO-Knowledge Organisational Analysis of
  System Developers Team.
• s3 WHAT-Knowledge Conceptual Analysis of Subject
  Domain Revealing Concepts and Relationships
  between them.
• s4 HOW-TO-Knowledge Functional Analysis
  Hypotheses and the Models of Decision Making.
• s5 WHERE-Knowledge Spatial Analysis
  Environment, Communications, etc.
• s6 WHEN-Knowledge Temporal Analysis Schedules,
  Time Constraints, etc.
• s7 WHY-Knowledge Causal Analysis Explanation
  System.
• s8 HOW-MUCH-Knowledge Economical Analysis
  Resources, Losses, Incomes, Revenue, SWAT, etc.

42
Object-Structured Analysis Matrix
43
Test N16(OSA on the base of interview)
44
Ontology
45
Ontology
46
How to draw an ontology?
Ontology as
Traditional hierarchy
Hypertext structure
List of properties
Relational table
Contents (catalogue)
47
Ontology
System
SubSystem 1
SubSystem 2
SubSystem 3
Unit 1
Unit 2
Unit N
Object 1
Object 1
Object 2
Object 2
Object N
Value 1
Value 2

48
Living-room conceptions ontology
49
(No Transcript)
50
Cimabue
XIII
P. Venetiano
Giotto
XIV
Masaccio
Jac.Bellini
Donatello
Fra Angelico
Mantegna
XV
Veroccio
Carpaccio
Girlandaio
Lorenzo
Boticelli
Signorelli
Giov.Bellini
Pinturiccio
Leonardo
Perujino
Titian
Giorgione
XVI
Michelangelo
Veronese
Rafael
Tintoretto
PERUGIA
VENICE
FLORENCE
51
Test N17(Departments ontology)
52
Ontologies in Internet
Ontolingua Project KA2 Initiative and Ontobroker
project SHOE project
53
Ontolingua

• Developed in the Knowledge System Laboratory -
  KSL - http//www.ksl.stanford.edu
• Program implementation of the Ontolingua system
  - Common Lisp.
• Main purpose support of the users formal
  tasks specification on the basis of formal
  descriptions library of tasks, models and
  concepts fragments and introduction of the
  fragments library.
• Knowledge Interchange Format (KIF) - language of
  Ontolingua system formal descriptions. Developed
  in KSL and proposed as a standard of
  intercomputer knowledge exchange. KIF language -
  the language of second order predicate calculus
  for representation of non procedural knowledge in
  which procedural knowledge is represented by
  rules of terms copying.
• http//www-ksl-svc.stanford.edu

54
Technology of work with Ontolingua
Internet
55
??2 Initiative and Ontobroker Project
Main research activities
Ontological Engineering, Web-pages
Annotation, Inquiries for information on
Web-pages and answers inference on the basis of
ontology knowledge.
http//ontobroker.aifb.uni-karlsruhe.de/
56
??2 Initiative and Ontobroker Project
Sections of Common Ontology

• organization ontology,
• project ontology,
• person ontology,
• research-topic ontology,
• publication ontology,
• event ontology,
• research-product ontology,
• research-group ontology.

57
??2 Initiative and Ontobroker project
Ontology
58
SHOE (Simple HTML Ontology Extensions) Project
Main Research Trends
Development of reusable ontologies for concepts
that are most frequent for Web-resources Devel
opment of knowledge annotators.
http//www.cs.umd.edu/projects/plus/SHOE/
59
Knowledge management
1. ACCUMULATION Spontaneous and unsystematic 2.
ACQUISITION Transfer process of specialists
competence on analyst. 3. STRUCTURING Main
concepts are selected, structure of information
presentation is worked out, descriptions and
models of business-processes and structure of
information streams are created. 4. DESIGNING OF
THE SYSTEM Subject statement and development of
architecture and specifications on
programming. 5. PROGRAM IMPLEMENTATION Software
development 6. SERVICE
60
Analyst's Motto (triad)

• From acquisition
• Via structuring
• To formalization !