Transcript Slide 1

A Lecture for GCET Students
September 11, 2009
Dr. Priti Srinivas Sajja
Department of Computer Science
Sardar Patel University
Vallabh Vidyanagar
Introduction and Contact Information
•
•
Name: Dr Priti Srinivas Sajja
Communication:
• Email : [email protected]
•
•
•
•
•
•
Mobile : 9824926020
• URL: http://priti.sajja.info
Academic Qualifications: Ph. D in Computer Science
Thesis Title: Knowledge-Based Systems for Socio-Economic
Rural Development
Subject area of specialization : Knowledge-Based Systems
Publications : 84 in International/ National Journals and Conferences
(Including two books and one chapter)
Academic Position : Associate Professor at
Department of Computer Science
Sardar Patel University
Vallabh Vidyanagar 388120
2
Outlines of the Lecture
Part 1: Artificial Intelligence




Natural intelligence and Artificial Intelligence
Nature of AI Solutions
Testing Intelligence
Categories of Application Areas
Part 2: Symbolic Knowledge-Based Systems





Data Pyramid and CBIS
DBMS and KBS
Structure of KBS
Types of KBS
Example KBS
Part 3: Connectionist Systems
 Symbolic and Connectionist Systems
 Example ANN for Course Selection
3
Natural Intelligence
• Responds to situations flexibly.
• Makes sense of ambiguous or erroneous messages.
• Assigns relative importance to elements of a
situation.
• Finds similarities even though the situations might
be different.
• Draws distinctions between situations even though
there may be many similarities between them.
4
Artificial Intelligence
• According to Rich & Knight (1991) “AI is the study of
how to make computers do things, at which, at the
moment, people are better”.
• A machine is regarded as intelligent if it exhibits
human characteristics generated through natural
intelligence.
• AI is the study of human thought processes and moving
towards problem solving in a symbolic and nonalgorithmic way.
• AI is the branch of Computer Science that attempts to
solve problems by mimicking human thought process
using heuristics, symbolic and non-algorithmic
approach in areas where people are better.
5
Make Your Own Definition of AI
human thought process
heuristic methods
where people are better
non-algorithmic
characteristics we
associate with
intelligence
knowledge using symbols
Figure 1.1: Constituents of artificial intelligence
6
Nature of AI Solutions
Acceptable
solution in
acceptable
time
Extreme
solution, either
best or worst
taking 
(infinite) time
time
Figure 1.2: Nature of AI solutions
7
Testing Intelligence
Turing test will fail to test for
intelligence in two
circumstances;
Can you tell
me what is
222222*67344
?
Why
Sir?
The Boss could not judge who was replying,
thus the machine is as intelligent as the
secretary.
Figure 1.4: The Turing test
8
1. A machine may well be
intelligent without being able to
chat exactly like a human; and;
2. The test fails to capture the
general properties of
intelligence, such as the ability
to solve difficult problems or
come up with original insights.
If a machine can solve a
difficult problem that no person
could solve, it would, in
principle, fail the test.
Application Areas of Artificial Intelligence
Rich & Knight (1991) classified and described the different areas that Artificial
Intelligence techniques have been applied to as follows:
Mundane Tasks
Expert Tasks
•
• Engineering - design, fault
finding, manufacturing
planning, etc.
• Scientific analysis
• Medical diagnosis
• Financial analysis
•
•
•
Perception - vision and
speech
Natural language
understanding, generation,
and translation
Commonsense reasoning
Robot control
Formal Tasks
• Games - chess,
backgammon, checkers, etc.
• Mathematics- geometry,
logic, integral calculus,
theorem proving, etc.
9
Data Pyramid and Computer Based Systems
Heuristics
and models
Wisdom
Novelty
Knowledge
Rules
Information
Experience
Concepts
Data
Understanding
Raw Data through
fact finding
Researching
Absorbing
Doing
Interacting
Reflecting
Figure 1.6: Convergence from data to intelligence
10
Data Pyramid and Computer-Based Systems
IS
Strategy makers apply morals, principles,
and experience to generate policies
WBS
Higher management generates knowledge
by synthesizing information
KBS
Middle management uses reports/info.
generated though analysis and acts
accordingly
Basic transactions by operational
staff using data processing
DSS, MIS
TPS
Volume
Wisdom (experience)
Knowledge (synthesis)
Information (analysis)
Data (processing of raw observations )
Sophistication
and complexity
Figure 1.7: Data pyramid: Managerial perspectives
11
Computer-Based Information Systems Tree
Intelligent systems:
21st century
challenge
Software resources
IS
EES
ES
1990
ESS
EIS
Users’ requirements
EES:
Executive expert system, which is a
hybridization of an expert system ,
executive information system, and
decision support system
DSS
MIS
OAS
TPS
Hardware base/technology
Figure 1.8: CBIS tree
12
1970
1950
Comparison of KBS with Traditional CBIS Systems
Traditional Computer-Based Information
Systems (CBIS)
Knowledge-Based Systems (KBS)
Gives a guaranteed
solution and
concentrate on efficiency
Adds powers to the solution and concentrates on
effectiveness without any guarantee of solution
Data and/or information
Knowledge and/or decision processing approach
processing approach
Assists in activities related to decision making
and routine transactions; supports
need for information
Transfer of expertise; takes a decision based on
knowledge, explains it, and upgrades it, if
required
Examples are TPS, MIS, DSS, etc.
Examples are expert systems, CASE-based
systems, etc.
Manipulation method is numeric
Manipulation method is primarily
symbolic/connectionist and nonalgorithmic
and
algorithmic
These systems do
These systems learn by mistakes
not make
mistakes
Need complete
data
information
and/or
Works for complex, integrated, and wide
areas in a reactive
13
manner
Partial and uncertain information, data, or
knowledge will do
Works for narrow domains in a reactive and
proactive manner
Objectives of KBS
KBS is an example of fifth-generation computer technology. Some of its objectives
are as follows:
•
Provides a high intelligence level
•
Assists people in discovering and developing unknown fields
•
Offers a vast amount of knowledge in different areas
•
Aids in management
•
Solves social problems in better way than the traditional CBIS
•
Acquires new perceptions by simulating unknown situations
•
Offers significant software productivity improvement
•
Significantly reduces cost and time to develop computerized systems
14
Components of KBS
Knowledge base is a repository of domain
knowledge and meta knowledge.
Enriches the
system with
self-learning
capabilities
Inference engine is a software program,
which infers the knowledge available in the
knowledge base
Explanation
and
reasoning
Provides
explanation and
reasoning
facilitates
Knowledge
base
Inference
engine
Selflearning
User interface
Figure 1.10: General structure of KBS
15
Friendly
interface to
users working
in their native
language
Categories of KBS
According to the classifications by Tuthhill & Levy (1991),
five main types of KBS exists:
Expert systems
Linked systems
Intelligent tutoring systems
CASE-based systems
Database in conjunction with an intelligent user interface
16
Difficulties with the KBS
• Completeness of Knowledge Base
• Characteristics of Knowledge
• Large Size of Knowledge Base
• Acquisition of Knowledge
• Slow Learning and Execution
17
An example of a Multi-agent KBS on Grid
Knowledge Utilization
Knowledge Management
Knowledge Discovery
and
Grid FTP ReplicaLocation
Services
Learning Mgt.
Resources
Drills and Quizzes
Knowledge Mgt.
Question/Answer
Meta knowledge
Conceptual system
Tutorial Path
Content knowledge
Explanation
Learner’s ontology
Information
Discovery
Services
Documentation
Security Services
Documents
Local
DataBases
Semantic Search
Mail
E-mail & Chat
Resource Management
Distributed
databases
18
Middleware
Services and
Protocols
User Interface Agent
Internet
Grid Middleware Services
Resource
Management (Grid
Resource
Allocation
Protocol-GRAM)
Agents
Users
Experts
Communication Between Agents
• Agents developed here are communicating with a tool
named KQML.
• Knowledge based Query Management Language.
(register
Action intended for the message
: sender agent_Lerning_Mgt
Agents name sharing message
: receiver agent_Tutorail-Path
: reply-with message
: language
common_language
: ontology
common_ontology
: content
)
19
“content.data”
Action intended for the message
Language of both agents
Ontology of both the
agents
Context-specific information
describing the specifics of this
message
Knowledge Representation of a Tutorial Topic: Array
20
Prototype Screen Designs for the KBS
21
Prototype Screen Designs for the KBS
22
Result from the System
23
An Example of a Connectionist System
Availability of expertise
BioInformatics
Availability of hardware/based
technology
Content /length of the course
Degree of assistance required
suggeste
d
decision
for
Current
Trends
[[[
Knowledge level required for the
course/ depth of the course
Market trend towards
technology/course
Personal interest
Hidden
Layers
Success history if any (last few
years result in%)
Time taken to complete
(revision)
Input
Layer
24
Output
Layer
Wireless
Tech.
Acknowledgement
Thanks to
GCET and Charutar Vidya Mandal
Reference
“Knowledge-Based Systems”
Rajendra Akerkar and Priti Srinivas Sajja
Book published by Jones and Bartlett
Publishers, Massachusetts (MA), USA.
25