Transcript Artificial Intelligence

DSS: Decision Support Systems
and
AI: Artificial Intelligence
In Business
AI in Business
Some Commercial Applications
• Decision Support
• Expert Systems
• Information Retrieval
• Virtual Reality
• Robotics
I’m ready to
do some
business
Overview of AI
• Goal of AI
– develop computer systems that exhibit
intelligence or simulate the ability to think
• AI pioneered by Computer Science
• But, AI involves a combination of
– Computer Science, Biology, Psychology,
Linguistics, Mathematics,Engineering
What really is Intelligence?
• Specifically, what are the
signs of Intelligent
Behavior?
• Think about it and write
down some answers.
What really is Intelligence?
• You are about to start an
online chat (IM) with two
entities:
– One entity is a human
– The other is a computer
• After hours of conversation,
you can not tell which
entity is a computer.
• Does this mean the
computer is Intelligent?
Intelligent Behavior
• What are some of the signs,
attributes, or characteristics of
Intelligent Behavior
Characteristics of
Intelligent Behavior
1. Learn from experience & apply
the knowledge
 Computer can automatically improve
performance based on Experience
 Machine Learning
 Computational Learning
Characteristics of
Intelligent Behavior
2. Handle complex situations
 Computer Systems can often handle
complexity better than humans
 Consider a process control system
that must simultaneous track 100
different system variables.
Characteristics of
Intelligent Behavior
3. Solve problems when important
information is missing
 Computer Systems can find patterns
and deal with all sorts of missing
information
Characteristics of
Intelligent Behavior
4. React quickly & correctly to new
situations; Acquire & Apply
Knowledge
 Here is where computers start to fail.
 Adapting to completely new situations is
a problem for computer systems.
 Its very difficult to design a computer
system that can combine, connect, and
acquire knowledge to solve completely
new problems
Characteristics of
Intelligent Behavior
5.
6.
7.
8.
Determine what is important.
Exhibit creativity and imagination
Process visual information efficiently
Use reason to solve problems

These are some other Characteristics that
humans possess.
Computer systems have a lot of catching up
to do.

AI in Business
• AI continues to improve and evolve.
• Scientists and Engineers are pushing the
envelope of what is possible.
• In Business, there is a better
understanding of the capabilities of
Intelligent Computer Systems
• It is important to know which types of
problems are suited for humans, and
which are suited for Computers.
Human Intelligence vs. AI
Attribute
Human
Intelligence
Artificial
Intelligence
Use a variety of information
sources
High
High
Ability to acquire large amounts
of external info.
Medium
High
Ability to do rapid, accurate,
and complex calculations
Low
High
Ability to transfer information
rapidly
Low
High
Human Intelligence vs. AI
Attribute
Human
Intelligence
Artificial
Intelligence
Ability to use sensors or senses
High
Medium
Creativity or imagination
High
Low
Ability to learn from experience
High
Medium
Ability of be adaptive
High
Medium
AI: Application Domains
AI: Commercial Domains
• Decision Support
– Integrating the advantages of AI with Human
Intelligence.
– More intelligent Interfaces
– More intelligent processing for massive data
• Information Retrieval
– Automatic simplification for massive data
– Natural language technology: computer can
speak our language.
AI: Commercial Domains
• Virtual Reality
– Better training environment from pilots to
doctors
• Robotics
– Bringing the precision and speed of computers
into the physical world
– Goes beyond manufacturing and assembly lines;
Baggage Inspection, Bomb Removal,
Replacement Limbs.
Expert Systems
• The idea is to inject expert knowledge in to
a computer system.
• The primary purpose is to automate
decision making.
• The decision environments have structure
• The alternatives and goals are often
established in advance.
Expert Systems vs. DSS
Expert System
• Inject expert knowledge in
to a computer system.
• Automate decision making.
• The decision environments
have structure
• The alternatives and goals
are often established in
advance.
• The expert system can
eventually replace the
human decision maker.
Decision Support System
• Extract or gain knowledge
from a computer system
• Facilitates decision making
• Unstructured environment
• Alternatives may not be
fully realized yet
• Use goals and the system
data to establish
alternatives and outcomes,
so a good decision can be
made
Some Interesting Applications
of Expert Systems
• Triage – Medical Diagnosis (Medical Expert System)
– User enters symptoms
– System makes diagnosis
– Doctors collective expertise is captured in the system
• Patriot Missile Guidance System
– Radar identifies Scud missile
– System steers Patriot missile to it intercepts Scud
missile
– Laws of physics, expert knowledge about missile
trajectory is captured in the system
• Financial Decision Making – Currency Trading
Expert System Categories
• Decision Making
– buy/sell
– risk/no risk
– rain/ no rain
• Trouble Shooting /
Diagnosis
• Selection/Classification
– Tell me what you see,
expert system figures out
what it really is...
• Process Monitoring and
Control
– Robot control, assemblyline control, missile
control
– Hello welcome to Dell;
how can I help you?
– Suddenly an idiot seems • Design/Configuration
– Specify what you want,
like an expert.
expert system figures out
specifically how to do it.
Expert System Components
Expert System Software
User
Interface
user
Engine
Knowledge
base
Expert System Components
Expert System Software
User
Interface
Engine
Knowledge
base
user
Expert System Development Process
Knowledge
Acquisition
Program
Expert or
Knowledge
Engineer
Raw Data or
Facts
Expert System Components
Nonexpert
Robot
Expert System Software
Interface
Engine
Knowledge
base
Missile
Expert System Development Process
Knowledge
Acquisition
Program
Expert or
Knowledge
Engineer
Raw Data or
Facts
Expert System vs. DSS
Someone
with
Knowledge
Decision
Maker
DSS Software
Model Base
User
Interface
Analytical &
Statistical
Models
Engine
DSS Processes
Data Management
Extraction, Generation,
Validation, etc.
Raw Data or
Facts