EXPERT SYSTEMS (contd.)

Download Report

Transcript EXPERT SYSTEMS (contd.)

Smile!
“You haven't lost
your smile at all, it's
right under your
nose. You just forgot
it was there.”
MODULE-2 : UNIT-9
ATIFICIAL
INTELLIGENCE
AND INTELLIGENT
SYSTEMS IN
BUSINESS
TOPICS IN UNIT-9
•ARTIFICIAL INTELLIGENCE
•INTELLIGENT SYSTEMS
•EXPERT SYSTEMS
•OTHER INTELLIGENT SYSTEMS
•INTELLIGENT AGENTS
•VIRTUAL REALITY
•ETHICAL & GLOBAL ISSUES
INTELLIGENT SYSTEMS
OF
NATURE OF HUMAN
INTELLIGENCE
1. Intelligence = Information communicated.
= Ability to understand and
communicate information.
2. Nature of Human Intelligence :
- Found in human beings,
- Can be increased by education
and training.
- Degree varies in persons
ARTIFICIAL INTELLIGENCE (A.I.)
1. Definition : It is a field of science and technology
based on disciplines like computer science,
biology, psychology, linguistics, mathematics and
engineering to provide intelligence artificially.
2. Purpose of A.I. : Developing machines with
intelligent behaviour.
3. Intelligent Behaviour :
(a) Learn from experience and apply knowledge
acquired from experience.
(b) Handle complex situations,
Contd.
ARTIFICIAL INTELLIGENCE (A.I.)
3. Intelligent Behaviour (contd.):
(c) Solve problems when important information is missing,
(d) Determine what is important,
(e) React quickly and correctly to a new situation,
(f) Understand visual images using perceptive
systems (see, hear and feel),
(g) Process and manipulate symbols (symbols and 3-D
objects)
(h) Be creative and imaginative,
(i) Use heuristics (thumb rules from experience).
DIFFERENCE BETWEEN NATURAL INTELLIGENCE
(N. I.) AND ARTIFICIAL INTELLIGENCE (A.I.)
S.
N.
ATTRIBUTES
Ability to :
N.I.
A.I.
(HUMAN)
(MACHINE)
1.
Use sensors (eyes, ears, touch, smell)
HIGH
LOW
2.
Be creative and imaginative
HIGH
LOW
3.
Learn from experience
HIGH
LOW
4.
Be adaptive
HIGH
LOW
5.
Afford the cost of acquiring intelligence
HIGH
LOW
6.
Use a variety of information sources
HIGH
HIGH
7.
Acquire large amount of external information
HIGH
HIGH
8.
Make complex calculations
LOW
HIGH
9.
Transfer information
LOW
HIGH
10
Make series of calculations rapidly and accurately
LOW
HIGH
MAJOR A.I. APPLICATION AREAS
A. I.
COGNITIVE
(KNOWLEDGE)
SCIENCE
APPLICATIONS
NATURAL
INTERFACE
APPLICATIONS
ROBOTICS
APPLICATIONS
COGNITIVE (KNOWLEDGE ) SCIENCE
APPLICATIONS
1. Expert Systems : Computer system that stores
knowledge and makes inferences similar to
reasoning by human expert.
2. Fuzzy Logic : Reasoning that deals with
uncertainties or partial information.
3. Genetic Algorithm : Used for finding optimal
solution from large number of possible solutions.
4. Neural Networks : Computer system that can
simulate the functioning of human brain.
5. Intelligent Agents :
- Percepts (Sensors)
- Acts (Actuators)
ROBOTICS APPLICATIONS
1.
2.
3.
4.
5.
Visual perception.
Tactility (touch).
Dexterity (cleverness).
Locomotion.
Navigation
NATURAL INTERFACE
APPLICATIONS
1. Natural Language Processing: Ability to
communicate with computers in human language.
2. Speech Recognition : Recognition and
understanding by a computer of a spoken
language.
3. Multi-sensory Interface :
4. Virtual Reality System : Enables one or more
users to move and react in a computer-simulated
environment.
INTELLIGENT SUPPORT SYSTEMS
FUNCTIONS AND TYPES
1. Capturing Tacit Knowledge :
- Expert Systems,
- Case-based reasoning,
- Fuzzy Logic
2. Knowledge Discovery :
- Neural Networks
- Data Mining
3. Generating Solutions to Very Large and Complex
Problems :
- Genetic Algorithm
EXPERT SYSTEMS
1. Definition : It consists of hardware and
software that stores knowledge about a
specific area and makes inferences to act
as an expert consultant in providing
decision support to end user.
2. Purpose : Overcomes limitations / nonavailability of human expert by :- Coping with new challenges
- Handling many decision variables,
- No ‘Information Fatigue Syndrome’,
EXPERT SYSTEMS (contd.)
1.
2.
3.
4.
5.
6.
7.
8.
9.
APPLICATION AREAS
Decision Management (make or buy, credit
limit, incentive scheme, customer query,
production query, investment counseling)
Diagnostic / troubleshooting,
Maintenance / scheduling
Design / configuration,
Selection / classification,
Process monitoring,
Product development / evaluation,
Performance evaluation,
New product launch
EXPERT SYSTEMS (contd.)
1. Success Factors in BES Implementation
- Cost effectiveness,
- Selective in scope
- User friendliness
2. Limitations of BES :
- Poor availability of domain human expert,
- Lack of flexibility required by dynamic
environment,
- Suited only for limited applications,
- Cannot replace human experts
COMPONENTS OF EXPERT SYSTEMS
A.I. SHELL
E.S. SOFTWARE
EXPLANATION
FACILITY
USER
INFERENCE
ENGINE
PROGRAMS
KNOWLEDGE
RULE BASE
USER
INTER
-FACE
USER
INTERFACE
PROGRAMS
BASE
SYMBOL
(FACTS)
DATABASE
R1 : IF () THEN()
1. Hari has a
R2 : IF() THEN ()
bank balance of
Rs 50000
2. Hari has an
urban
background.
KNOWLEDGE
BASE
ACQUISITION
FACILITY
EXPERT
How expert systems work?
1. Functions of A.I. Shell :
- Be user friendly,
- Quickly generate user interface screens
- Capture the knowledge base,
- Manage the strategies for searching the
rule base.
2. Inference Engine : Strategies used to search
the rule base :(a) Forward chaining,
(b) Backward chaining
contd.
How expert systems work? (contd.)
Forward Chaining
The inference begins with the information (facts)
entered by the user and searches the rule base
to arrive at a conclusion.
Backward Chaining
1. Process of starting with conclusion and working
backwards to supporting facts.
2. If facts do not support the conclusion, another
conclusion is selected and tested as per 1
above.
3. Continue the process until correct conclusion is
identified.
CASE-BASED REASONING
1. Purpose : To adapt successful solutions used in the
past (cases) to solve new problems.
2. Method : Find the solutions that solved problems
similar to current problem. Then adapt the previous
solution to fit the current problem.
3. Steps in Finding Relevant Cases :
(a) Characterizing the input problem,
(b) Retrieving from memory the cases with those
features,
(c) Picking the case or cases that best match the
input.
FUZZY LOGIC
1. Fuzzy Logic : Reasoning that deals with uncertainties
or partial information.
2. Method :
(a) Deals with uncertainties by simulating the process
of qualitative human reasoning.
(b) Allows computers to behave less precisely /
logically.
3. Rationale : Decision making is not always precise,
there are grey areas where terms approximately,
possible and similar are used.
4. Example : In developing marketing strategy, it helps
managers handle uncertainties and fuzziness of data
and information.
NEURAL NETWORKS
1. Definition : A complete system
that can simulate the functioning
of a human brain.
2. Like human brain, a neural net
has a large number of sensing
and processing nodes that
continuously interact with each
other.
NEURAL NETWORKS (contd.)
•
•
•
•
•
FEATURES OF NEURAL NETWORKS
Ability to retrieve information even if
some neural nodes fail.
Fast modification of data as a result of
new information,
Ability to discover relationships and
trends in large databases.
Ability to solve complex problems for
which all information is not available.
Have generalized capability to learn.
NEURAL NETWORKS (contd.)
APPLICATION AREAS OF NEURAL NETWORKS
1.
2.
3.
4.
Medicine : For screening of patients (e.g. Papnet)
Financial Industry :
(a) Periodic stock performance, bond ratings or corporate
bankruptcies,
(b) Detecting credit card frauds,
Other Business Applications of Neural Nets:
(a) Pattern classifications (e.g. patterns in sales data)
(b) Predictions, (c) Control,
(d) Optimization,
(e) Decision on credit / mortgage applications
(f) Pick duplicate names in mailing lists
Scientific Applications :
(a) Hand written character recognition,
(b) Image compression,
(c) Electronic nose.
DIFFERENCE BETWEEN EXPERT
SYSTEMS AND NEURAL NETWORKS
EXPERT SYSTEMS
NEURAL NETWORKS
1. Seeks to emulate a 1. Do not model human
reasoning. Designed to
or model human
imitate the physical
expert’s way of
thought process of
solving problem.
biological brain.
2. Expert system is
highly specific to a 2. Do not aim to solve
specific problems.
given problem.
3. Expert Systems
cannot be easily
retrained.
3. Intelligence is put in to
hardware in the form a
generalized capability to
learn.
DATA MINING
1.
Definition : A means of extracting
previously unknown, predictive information
from the base of accessible data in data
warehouse.
2.
Purpose : Sophisticated
algorithms are used to :
/
automated
(a) To discover hidden patterns, correlations,
and relationships among organizational
data,
(b) To predict future trends and behaviors,
allowing businesses to make proactive,
knowledge-driven decisions.
DATA MINING (contd.)
3. Functions: Five main functions :
(a) Classification,
(b) lustering,
(c) Association
(d) Sequencing,
(e) Forecasting
4.
Applications : Customer retention;
campaign management; market, channel &
pricing analysis; customer segmentation
etc.
INTELLIGENT AGENTS
1. Definition : It is a software entity that senses its environment
and then carries out some operations on behalf of a user (or a
program), with certain degree of autonomy, and in so doing
employs knowledge or representations of user’s goals or
desires.
2. Characteristics :
(a) Autonomy – capability to work on their own,
(b) Exhibition of goal oriented behavior,
(c) Mobility – transportable over networks,
(d) Dedication to a single repetitive task,
(e) Ability to interact with humans, systems and other agents,
(f) Inclusion of a knowledge base,
(g) Ability to learn
INTELLIGENT AGENTS (contd.)
3.Applications of Intelligent Agents :
(a) Information access using search engines,
(b) Decision support and empowerment,
(c) Repetitive office activities,
(d) Mundane personnel activities,
(e) Database search and retrieval agents
(f) Electronic commerce agents
(g) Domain experts,
(h) Management activities
VIRTUAL REALITY
1. Definition : It is interactive, uses computer generated, threedimensional graphics, and is delivered to the user through a
head mounted display.
2. Benefits : Many can share and interact in same environment. It
is powerful medium for communications, collaborative
entertainment, and learning.
3. Business Applications :
(a) Manufacturing : Training, designing, testing, simulation of
assembly and production,
(b) Transportation : Virtual aircraft mockup, new car design &
virtual accidents, air travel simulation,
(c) Finance : View stock prices
(d) Marketing : Store and product display; electronic shopping.
ETHICAL & GLOBAL ISSUES OF
INTELLIGENT SYSTEMS
ETHICAL AND SOCIAL ISSUES :
1.Possibility of power misuse and harm to
people from the use of intelligent
systems,
2.Privacy in knowledge bases,
3.Intellectual
knowledge
property
–
4.Misuse of robotic capabilities.
expert’s
ETHICAL & GLOBAL ISSUES OF
INTELLIGENT SYSTEMS (contd.)
LEGAL ISSUES
1. Responsibility in case of incorrect judgement leading
to damage or disaster,
2. Liability for wrong advice provided by ES?
3. Who owns the knowledge in knowledge base?
4. Who is an expert?
5. Can management force the experts to contribute their
expertise?
6. Should royalties be paid to expert and how much?
7. Value of expert opinion in court when the expertise is
encoded in a computer?
ETHICAL & GLOBAL ISSUES OF
INTELLIGENT SYSTEMS (contd.)
GLOBAL ISSUES : Global applications :1. Foreign Trade : Online ES to advise how to exploit opportunities,
2. Foreign Exchange Transactions : Intelligent systems for FE
transactions,
3. Employee Training : Online training to reduce time,
4. Weather Forecasting : Climatic ES to provide long-range climate
forecasts for global commodity traders.
5. Automatic Language Translation : Very important in global ecommerce.
6. Others : ES to fight money laundering, ES to provide expert
advice in areas of medicine, safety, agriculture, and crime
fighting.
MODULE-2 : UNIT-9
ATIFICIAL
INTELLIGENCE
AND INTELLIGENT
SYSTEMS IN
BUSINESS