Chapter 4 Artificial Intelligence

1
Chapter 4 Artificial Intelligence
2
AI -- Artificial Intelligence

• AI is the science of making machines imitate
  human thinking and behavior.
• IT can further augment business brainpower by
  using AI.
• What is the difference between a car (mechanical
  device) and a robot (intelligent device?)
• The robot has the capability of taking action on
  its own, whereas the car does not.

3
AI and Business Systems

• AI that businesses use most can be broken into
  four categories
• Expert systems
• Neural networks
• Can be trained to recognize patterns.
• Genetic algorithms
• Can generate increasingly better solutions to
  problems by generating many solutions then
  choosing the best one.
• Intelligent agents
• Adaptive systems that work independently,
  carrying out specific, repetitive or predictable
  tasks.

4
Expert Systems

• Sometimes referred to as a knowledge-based
  system.
• Reason through problems and offer advice in the
  form of a conclusion or recommendation.
• Excellent for diagnostic and prescriptive
  problems.
• Work well with problems with clear rules and
  procedures.
• Examples
• Accounting for auditing and tax planning
• Medicine to prescribe antibiotics where many
  considerations have to be taken into account.

5
Expert Systems

• An expert system can
• Handle massive amounts of data.
• Reduce errors.
• Aggregate information from various sources.
• Improve customer service.
• Provide consistency in decision making.
• Provide new information.
• Decrease personnel time spent on tasks.
• Reduce cost.

6
Problems in Creating Expert Systems

• The domain expert may have a hard time explaining
  exactly how they want the software to work.
• The process (set of rules) may be too complex or
  too vague.
• Can only deal with problems for which it was
  designed.
• Has no common sense or judgement.

7
Neural Networks

• An AI system which is capable of learning
  different patterns.
• Simulates the human ability to classify things.
• Most successful when there is a vast amount of
  information available.
• In business neural networks are popular for
  securities trading, fraud detection, credit
  evaluation, and real estate appraisal.

8
Neural Network Example

• Chase Manhattan Bank uses neural network
  internally to evaluate commercial loans.
• Master Card and many other credit card companies
  use neural networks to detect credit card fraud.
• Case in England where neural networks are
  analyzing behavioral patterns in children with
  the hope of detecting criminal patterns early.
• Is this an infringement on civil liberties?
• What would happen if this system were implemented
  in the United States?

9
Genetic Algorithms

• An AI system that mimics the evolutionary,
  survival-of-the-fittest process to generate
  increasingly better solutions to a problem.
• Best suited to decision making environments in
  which thousands, or perhaps millions of solutions
  are possible and each solution must be carefully
  evaluated.

10
Genetic Algorithm Examples

• US West uses a genetic algorithm to determine the
  optimal configuration of fiber-optic cable in a
  network that may include as many as 100,000
  connections.
• Evolution of what foods taste good together has
  taken place over centuries and varies throughout
  the world.

11
Genetic Algorithm

• Genetic algorithms are base on three concepts of
  evolution.
• Selection
• Survival of the fittest.
• Crossover
• Combining portions of good outcomes in hopes of
  creating a better solution.
• Mutation
• Randomly trying combinations and evaluating the
  successes or failures of the outcome.

12
Intelligent Agents

• Software that assists you in performing
  repetitive tasks
• It adapts itself to your preferences.
• Four types of intelligent agents.
• Find and retrieve agents.
• User agents.
• Monitor and surveillance agents.
• Data-mining agents.

13
Find and Retrieve Agents

• Buyer agents and shopping bots
• Intelligent agents that look on the web for the
  product of your choice.
• Retrieves prices.
• Example www.mysimon.com

14
User Agents

• Intelligent agents that help an individual
  perform computer-related tasks.
• Usually work in the background.
• Example
• Programs that automatically check for your email
  and alert you when it is something you want.
• Agents that fill out forms on the web
  automatically for you.
• Agents that scan web sites and highlight text
  that you are interested in.

15
Monitor and Surveillance Agents

• Perform continual and repetitive tasks that keep
  working in the background.
• Used in push technology
• Sending information to your computer without a
  special request for it.
• Email does not come at a pre-arranged time.
• Example
• Network monitoring agents that examine the load
  on a network to make sure it has ample memory.
• Sun Microsystems new idea for network computing.

16
Data Mining Agents

• Operates in a data warehouse discovering
  information.
• May detect a major shift in a trend or another
  key indicator.