HFE 451/651
Artificial Intelligence and
Expert Systems
-Presented By
Damodar
Kavya
Sogra
10/3/2015
Contents
• Introduction
• Definitions of AI
• Approaches of AI
• History of AI
• Designing an AI system
• Applications of AI
• Expert Systems
• Conclusion
• References
• 10/3/2015
Questions????
Introduction
Artificial Intelligence (AI) is the area of computer science
focusing on creating machines that can engage on
behaviors that humans consider intelligent.
AI is a broad topic, consisting of different fields, from
machine vision to expert systems. The element that the
fields of AI have in common is the creation of machines
that can "think".
10/3/2015
Introduction(contd.)
AI researchers are active in a variety of domains.
• Formal Tasks (mathematics, games),
• Mundane tasks (perception, robotics, natural
language, common sense reasoning)
• Expert tasks (financial analysis, medical
diagnostics, engineering, scientific analysis, and
other areas)
10/3/2015
Some definitions of AI
S y stem s th at th in k lik e h u m an s
S y stem s th at th in k ratio n ally
“T he exciting new effort to m ake com puters “T he study of m ental faculties
think… m achines w ith m inds, in the full and through the view s of com putational
literal sense”(H augeland, 1985)
m odels”(C harniak and M cD erm ott,
1985)
“[T he autom ation of] activities that w e
“ T he study of com putations that
associate w ith hum an thinking, activities such m ake it possible to perceive reason
as decision m aking, problem solving,
and act”(W inston, 1992)
learning… ”(B ellm an, 1978)
S y stem s th at act lik e h u m an s
“T he art of creating m achines that perform
functions that require intelligence w hen
perform ed by people”(K urzw eil, 1990)
“T he study of how to m ake com puters do
things at w hich, at the m om ent, people are
better”(R ick and K night, 1991)
10/3/2015
S y stem s th at act ratio n ally
“A field of study that seeks to explain
and em ulate intelligent behavior in
term s of com putational
processes”(S chalkoff, 1990)
“T he branch of com puter science that
is concerned w ith the autom ation of
intelligent
behavior”(L uger
and
S tubblefield, 1993)
Approaches to AI
Acting humanly: The Turing Test approach
• Alan Turing(1950)
• Designed to provide a satisfactory operational
definition of intelligence
• Intelligent behavior- The ability to achieve human-level
performance in all cognitive tasks, sufficient to fool an
interrogator.
The computer would need to possess
• Natural language processing
• Knowledge representation
• Automated reasoning
• Machine learning
10/3/2015
Thinking humanly: The Cognitive modelling approach
Determine how humans think
• Introspection
• Psychological experiment
Come up with precise theory of the mind and express as a
computer program
• GPS - Newall and Simon, 1961
• Wang
10/3/2015
Thinking rationally: The laws of thought approach
• Aristotle – “Right thinking”
• Laws of thought govern the operation of mind –
initiated the field of logic
• Programs based on laws of thought to create intelligent
systems
Main obstacles
- Informal knowledge in terms of formal terms
- Difference between theoretical and practical approach
10/3/2015
Acting rationally: The rational agent approach
•
•
•
•
Acting so as to achieve one’s goals given one’s beliefs
Agent – perceives and acts
AI is the study and construction of agents
Situational awareness unlike the laws of thought
approach(makes inferences)
• Knowledge and reason to reach good decisions in a
wide variety of situations
Advantages:
- More general than laws of thought approach
- More open to scientific development than approaches
based on human behavior or thought –clearly defined
rationality
10/3/2015
Why Artificial Intelligence??
• Attempts to understand intelligent entities-learn
more about ourselves
• Strives to build intelligent entities as well as
understand them
• Computers with human-level intelligence(or better)
would have a huge impact on our daily life
• Allows less or no human involvement
10/3/2015
History of AI
The beginnings of AI reach back before electronics, to
philosophers and mathematicians such as Boole and
others theorizing on principles that were used as the
foundation of AI Logic.
AI really began to intrigue researchers with the invention
of the computer in 1943
The technology was finally available, or so it seemed, to
simulate
10/3/2015 intelligent behavior
History of AI
• Warren McCulloch and Walter Pitts (1943)
developed a model of artificial neurons.
• Claude Shannon (1950), and Alan Turing (1953)
developed chess programs
• John McCarthy, Marvin Minsky, Shannon and
Nathaniel Rochester - neural networks and the
study of intelligence
10/3/2015
History of AI
• A big contribution to AI, again came from
McCarthy in 1958 when he wrote a high level
programming language called 'LISP'.
• Allen Newell and Herbert Simon developed
'General Problem Solver‘
• Weizenbaum's ELIZA program (1965)
• MYCIN was developed to diagnose blood
infections.
• Many other algorithms
10/3/2015
History of AI(contd.)
AI has grown from a dozen researchers, to thousands
of engineers and specialists; and from programs
capable of playing checkers, to systems designed to
diagnose disease.
Advanced-level computer languages, as well as
computer interfaces and word-processors owe their
existence to the research into artificial intelligence.
10/3/2015
Designing an AI System
1. Top Down Approach
2. Bottom Up Approach
Bottom Up Approach is most widely used
10/3/2015
Some Facts about the Human
Brain
• Human Brain is made up of Billions of cells
called neurons
• Neurons work when grouped together
• Decisions are made by passing electrical signals
• Neurons are devices for processing Binary digits
10/3/2015
How Binary processing works
• Binary numbers are represented as 0 and 1or T
and F
• A decision is made from a given input in terms of
0 and 1
– Apples are red-- is True
– Apples are red AND oranges are purple-- is
False
– Apples are red OR oranges are purple-- is True
– Apples are red AND oranges are NOT purple-is also True
10/3/2015
•
Relevance to the Human Mind
• The Human Mind works on the principle of
Binary processing
• Information is transmitted via impulses
– Presence of impulse – True
– Absence of impulse –False
*Logical Operation is based on two or more such
signals
10/3/2015
Network of Neurons
10/3/2015
Decision Making Process
• Identify a Bird
10/3/2015
Applications Of AI
• Banking System
- Micro Bankers High Tech Banking System
- Internet Banking
• Medicine
- MYCIN
- INTERNEST
• Eliza
- The Psychotherapist
10/3/2015
ELIZA- computer therapist
http://www.manifestation.com/neurotoys/eliza.php3
10/3/2015
Expert Systems
Expert systems are computerized advisory
programs that attempt to imitate the
reasoning process and knowledge of
experts in solving specific types of
problems.
10/3/2015
History
• 1960s
• 1970s
Renaissance Age
10/3/2015
What can Expert Systems do?
•Diagnosis
•Instruction
•Monitoring
•Analyzing
•Interpretation
•Debugging
•Repair
•Control
•Consulting
•Planning
•Design
10/3/2015
Knowledge Engineering
-the discipline of building expert systems
•Knowledge Acquisition
•Knowledge Elicitation
•Knowledge Representation
10/3/2015
How does it work?
• Knowledge Base
• Inference Engine
• A generalized Interface
10/3/2015
When Expert Systems are
applicable to the
Nature of the task?
• Expert systems can do much better
• Task involves reasoning and knowledge
and not intuition or reflexes
• Task can be done in minutes or hours
• Task is concrete enough to codify
• The task is commonly taught to novice in
the area.
10/3/2015
When expert systems are
applicable
Nature of the knowledge
• Recognized expert exist
• There is general agreement among
experts
• Experts are able and willing to
articulate the way they approach
problems.
10/3/2015
How the system works?
•
•
•
•
Use AI techniques
Knowledge component
Separate knowledge and control
Use inference procedures - heuristics uncertainty
• Model human expert
10/3/2015
Comparison of conventional and
expert systems
Conventional System
Information and processing are
combined in one program
Expert System
Knowledge base is separated from processing
mechanism
May make mistakes
Does not make mistakes
Changes are tedious
Changes are easy
System operates only when completed
System can operate even with few rules
Data processing is a repetitive process
Knowledge engineering is inferential process
Algorithmic
Heuristic
Representation
10/3/2015 and use of data
Representation and use of knowledge
How do people reason?
•
•
•
•
•
10/3/2015
They create categories
They use specific rules, a priori rules
They Use Heuristics --- "rules of
thumb"
They use past experience --- "cases"
They use "Expectations"
How do Computers Reason?
Computer models are based on models
of human reasoning
• They use rules A--->B--->C
• They use cases
• They use pattern
recognition/expectations
10/3/2015
Features of Expert Systems
• Deal with complex subject which normally
require a considerable amount of human
expertise.
• Exhibit performance and high reliability
• Capable of explaining and justifying solutions
and recommendations.
10/3/2015
Features of Expert Systems
(contd.)
• Incorporate some form of Inferential
reasoning.
• Be flexible, capable of accomodating
significant changes without necessary
programming
• Be user friendly
10/3/2015
Examples of Expert Systems
•
•
•
•
•
Dendral-Identify organic compounds.
Mycin-diagnosing medical problems.
Prospector-identifying mineral deposits
XCON-customized hardware configuration.
Expert Tax- accrual and tax planning
10/3/2015
Advantages of Expert Systems
•
•
•
•
•
•
•
•
10/3/2015
Permanence
Reproducibility
Efficiency
Consistency
Documentation
Completeness
Timeliness
Differentiation
Disadvantages of Rule-Based Expert
Systems
•
•
•
•
Creativity
Learning
Sensory Experience
Degradation
•
Common sense
10/3/2015
Conclusion: Computers Think--and
Often Think Like People
10/3/2015
References
• Artificial Intelligence – A Modern Approach-Stuart J. Russell and Peter
Norvig
• http://library.thinkquest.org
• http://www.ai.mit.edu/people/minsky/minsky.html
• What is Artificial Intelligence? by John McCarthy, Computer Science
Department, Stanford University
• What is Artificial Intelligence? by Aaron Sloman, Computer Science
Department, University of Birmingham, UK
• Expert Systems: A Quick Tutorial - by Schmuller, Dr. Joseph, Journal
of Information Systems Education 9/92, Volume 4, Number 3
• Artificial Intelligence a Modern Approach --- Chapter 1 Introduction
by Stuart Russell and Peter Norvig.
• AI Tutorial by Eyal Reingold, University of Toronto
• AI Education Repository - links to classes, tutorials etc.
10/3/2015
Descargar

Artificial - Home | College of Engineering and Computer