Artificial Intelligence and Modeling the Human State

1
Artificial Intelligence and Modeling the Human
State

• Are computers smart enough to replace people?
• In this chapter
• Does looking intelligent mean that intelligence
  is present?
• How does the human brain differ from a computer?
• How does a computer gain and retrieve knowledge
  as compared to how a human gains and retrieves
  knowledge?
• How is it that a computer can recognize text,
  speech, or a human face?
• How are computer scientists making computers
  smarter?

2
What is IntelligenceArtificial or Not?

• Chapters Purpose
• To show the attempts at creating intelligent
  systems using the computer.
• To get some very small insight into the working
  of the human brain.

3
What is IntelligenceArtificial or Not?

• The search for intelligence
• Plato (400 BC) - This Greek philosopher believed
  that ethereal spirits were rained down from
  heaven and entered the body.
• Aristotle (Platos student) - The heart must
  contain the soul and the brains function was to
  cool the blood.
• Galen - Treated fallen gladiators with spinal
  cord injuries. Noted that feeling lost in certain
  limbs sometimes came back.
• Galvani - Used Benjamin Franklins findings about
  static electricity to show that static
  electricity stimulated the nerves causing a frog
  to jump.
• Subsequently - Human nervous system found to be a
  complex network of billions of neurons.

4
What is IntelligenceArtificial or Not?

• Does looking intelligent mean that intelligence
  is present?
• Maillardets Automaton (1805)
• Object having human form.
• Disguised as a young boy.
• Complex drawing machine containing levers,
  ratchets, cams and other mechanical devices.
• Could draw several complex images.
• Because it had human form and could draw complex
  images, a certain feeling of intelligence was
  ascribed to the machine.

5
What is IntelligenceArtificial or Not?

• Sailing vessel drawn by Maillardets Automaton.

6
What is IntelligenceArtificial or Not?

• Alan Turing (1912 - 1954)
• Proposed a test - Turings Imitation Game
• Tests the intelligence of the computer.
• Phase 1
• Man and woman separated from an interrogator
• The interrogator types in a question to either
  party
• By observing responses, the interrogators goal
  was to identify which was the man and which was
  the woman

7
What is IntelligenceArtificial or Not?

• Phase one of the Turings test in operation.

Interrogator
Honest Woman
Lying Man
8
What is IntelligenceArtificial or Not?
Interrogator

• Phase 2 of the Turings test
• The man was replaced by the computer.
• If the computer could fool the interrogator as
  often as the man did, it could be said that the
  computer had displayed intelligence.

Human
Computer
9
What is IntelligenceArtificial or Not?

• Claude Shannons comparison of the human brain
  and the computer
• Difference in size The brain has a million more
  parts.
• Difference in structural organization The
  seemingly random local structure of nerve
  networks differ vastly from the precise wiring of
  a computer.
• Differences in reliability The brain can operate
  reliably for decades.
• Differences in logical organization The brain is
  largely self-organizing. Digital computers do
  only a few narrowly defined tasks well.
• Differences in input-output equipment Brain is
  designed with input organs and output muscles and
  glands. Computers operate in an abstract
  environment of numbers and operations on numbers.

10
Fundamental Concepts in Artificial Intelligence

• How do humans keep the vast amounts of knowledge
  and how do they access it?
• One way to study complex systems is to build a
  working model of the system, and observe it in
  action.
• Approaches used to model the human knowledge
  system
• Semantic networks - Designed after human
  associative memory.
• Frames and Scripts -
• Frames attempt to create descriptions of objects
  and events.
• Scripts describe activities and supply possible
  outcomes.
• Rule-based or Expert systems - Consists of rules
  of the form IF (condition) THEN (action).

11
Fundamental Concepts in Artificial Intelligence

• Semantic networks - Designed after human
  associative memory.

Is a
Is a
John
Plumber
Worker
Owner
Is a
Ownee
Owner
Ford
Car
Is a
Start-time
May 97
Time
Is a
End-time
Oct 98
Is a
Ownership
Situation
12
Fundamental Concepts in Artificial Intelligence

• Frames attempt to create descriptions of objects
  and events.

Generic RESTAURANT Frame Specialization of
Business-establishment Types Range (Cafeteri
a, Seat-yourself, Wait-to-be-seated) Default Wa
it-to-be-seated If needed IF
plastic-orange-counter THEN fast-food IF
stack-of-trays THEN cafeteria IF
wait-for-waitress-sign OR reservations-made
THEN wait-to-be-seated OTHERWISE
seat-yourself Location Range an address If
needed (Look-at-the-menu) Name If
needed (Look-at-the-menu) continues...
13
Fundamental Concepts in Artificial Intelligence

• Scripts describe activities and supply possible
  outcomes.

EAT-AT-RESTAURANT Script Props (Restaurant,
Money, Food, Menu, Tables, Chairs) Roles (Hungry
-persons, Wait-persons, Chef-persons) Point-of-v
iew Hungry-persons Time-of-occurrence (Times-of
-operation-of-restaurant) Place-of-occurrence (L
ocation-of-restaurant) Event-sequence First Ent
er-restaurant script Then if
(Wait-to-be-seated-sign or Reservations) then
(Get-maitre-ds-attention script) Then Please-b
e-seated script Then Order-food
script Then (Eat-food script) unless
(Long-wait) then (Exit-restaurant-angry
script) Then Pay-for-it script Finally Leave
-restaurant script
14
Fundamental Concepts in Artificial Intelligence

• Rule-based or Expert systems - Consists of rules
  of the form IF (condition) THEN (action).
• IF (it is raining AND you must go outside)
• THEN (put on your raincoat)

15
Fundamental Concepts in Artificial Intelligence

• For any of these models of the human knowledge
  system to work, it must be able to make use of
  this knowledge in three different ways
• Knowledge acquisition - Must be some way of
  putting information or knowledge into the system.
• Knowledge retrieval - Must be able to find
  knowledge when it is wanted or needed.
• Reasoning with knowledge - Must be able to use
  that knowledge through thinking or reasoning.

16
Fundamental Concepts in Artificial Intelligence

• Knowledge acquisition
• Chair A thing with four legs, a back, and a flat
  surface that you can sit on.

17
Fundamental Concepts in Artificial Intelligence

• Chair A thing with four legs, a back, and a flat
  surface that one person can sit on at a time.
• Chair A piece of furniture consisting of a seat,
  legs, and a back, and often arms, designed for
  one person.

18
Fundamental Concepts in Artificial Intelligence

• Knowledge retrieval (by searching)
• Brute-force search - Searching all possible
  moves, and then selecting the best.
• Looking for a museum in a small town example
• Drive around, down every street, until you find
  one!
• Heuristic search - Uses rules of thumb,
  intuition. (The solution is not always
  guaranteed.)
• Looking for a museum in a small town example
• Look for the museum down the towns main street
  (museums are usually on the main street in
  small towns!)

19
Fundamental Concepts in Artificial Intelligence

• Reasoning with knowledge (what we humans do to
  solve problems)
• Two types usually used in the field of Artificial
  Intelligence
• Shallow reasoning - based on heuristics
  (intuition), or rule-based knowledge.
• Deep reasoning - analyzing the structure and
  function of component parts of the problem.
• Humans commonly apply deep reasoning.
• Computers, for the most part, use shallow
  reasoning.

20
Fundamental Concepts in Artificial Intelligence

• Learning systems For computers to become truly
  intelligent, they must be capable of learning on
  their own.
• A commonly accepted classification scheme for
  learning
• Rote learning - memorization of facts.
• Learning by instruction - similar to
  student/teacher relationship found in classrooms.
• Learning by deduction - drawing conclusions from
  certain premises (This is a cat. All cats are
  animals. Therefore, this is an animal.)
• Learning by induction - Includes subcategories
  learning by example, experimentation,
  observation, and by discovery.
• Learning by analogy - combines both deductive and
  inductive learning. (Being bitten by a teased dog
  may make an individual not tease bees.)

21
Fundamental Concepts in Artificial Intelligence

• Machine learning Writing intelligent computer
  programs that are capable of learning.
• Example Teaching a computer to play a game. The
  more the computer plays, the more strategies it
  will learn.
• Common sense
• The computer must be able to make inferences from
  the knowledge base.
• Answers to problems might not be listed.
• The computer will need to come up with its own
  answers!
• This has been a very difficult area in Artificial
  Intelligence.

22
Pattern Recognition

• Humans have the ability to understand or
  recognize the relationship among various parts of
  patterns in visual object, sound patterns,
  smells, and taste.
• Pattern recognition using computers has been
  applied in many areas including
• Robot vision
• Speech recognition
• Fingerprint identification
• Handwriting identification
• Optical character recognition (OCR)
• Weather data analysis and satellite data analysis

23
Pattern Recognition

• Recognizing printed letters
• Problems - The letters needs to be in some
  standard position and size.
• Some commercial products - fax modems and
  scanners.

Letter positions and orientations
Letter fonts and styles
24
Pattern Recognition

• Speech-pattern recognition
• Problems - Accents, continuous speech, words that
  sound the same but are spelled differently.
• Some commercial products - Some commercial
  products do recognize continuous speech. (not yet
  perfect)
• He walked to the store.
• He walked to this door.

25
Game Playing

• Early days of AI - Researchers thought that
  teaching computers to play games such as chess
  would enable them to understand something about
  human intelligence.
• Found it easy to have computers play games.
• Found it difficult to go beyond game playing and
  into the realm of human intelligence.

26
Game Playing

• Major problem in game playing How to make the
  moves?
• Involves
• Processing or searching many moves (see game
  tree)
• By adding heuristics (rules of thumb /
  intuition), decisions can be made by the computer
  as to what is the best move.

27
Game Playing

• The Hexapawn game tree.

28
Game Playing

• A particular Hexapawn game situation.

Black Wins
White Wins
29
Game Playing

• Chess
• MacHack (1974) by Richard Greenblatt of MIT -
  First program to compete in a chess tournament.
• Deep Blue (1996) by IBM - Garry Kasparov,
  world-champion chess player, won over Deep Blue 4
  points to 2.
• Deep Blue (1997) by IBM - Garry Kasparov conceded
  victory to Deep Blue, 3.5 points to 2.5.
• The fastest chess-playing program processes
  200,000 positions per second.
• (Thousands of moves more than a human is capable
  of.)
• Still, the best chess players beat most computer
  programs.

30
Natural Language Communication

• The semantics (meaning of the words) is the goal
  in communication with the computer.
• Eliza (1965) by Joseph Weizenbaum of MIT
• Simulation of a therapist mirroring a patients
  conversation in a positive manner.
• The program fooled people into believing it was a
  therapist.

31
Natural Language Communication

• All men are alike.
• IN WHAT WAY
• Theyre always bugging us about something or
  other.
• CAN YOU THINK OF A SPECIFIC EXAMPLE
• Well, my boyfriend made me come here.
• YOUR BOYFRIEND MADE YOU COME HERE
• He says Im depressed much of the time.
• I AM SORRY TO HEAR YOU ARE DEPRESSED
• Its true, I am unhappy.
• DO YOU THINK COMING HERE WILL HELP YOU NOT TO BE
  UNHAPPY
• I need some help, that much seems certain.
• WHAT WOULD IT MEAN TO YOU IF YOU GOT SOME HELP
• Perhaps I could learn to get along with my
  mother.
• TELL ME MORE ABOUT YOUR FAMILY
• Eliza conversation

32
Natural Language Communication

• Early failures in natural language translation
• Translation of English to Russian
• Typed in English (sentence to be translated...)
• The spirit is willing, but the flesh is weak.
• Translated by the program into Russian
• The vodka is strong, but the meat is rotten.
• Translation programs have come a long way since
  then!

33
Expert Systems

• Expert systems are commercially the most
  successful domain in Artificial Intelligence.
• IF (some condition) THEN (some action)
• These programs mimic the experts in whatever
  field.

Auto mechanic Telephone networking Cardiologist De
livery routing Organic compounds Professional
auditor Mineral prospecting Manufacturing Infectio
us diseases Pulmonary function Diagnostic
internal medicine Weather forecasting VAX
computer configuration Battlefield
tactician Engineering structural
analysis Space-station life support
Audiologist Civil law
34
Expert Systems

• Bots
• Bots are precursors of Intelligent Agents that
  roam the World Wide Web looking for things like
  new web pages and email addresses.
• Different forms of Bots

Web robots Userbots Clonebots Gossipbots Spiders
Taskbots Crashbots Gamebots Wanderers Chatterbots
Floodbots Conceptbots Worms Knowbots Annoybots Rov
erbots Cancelbots Mailbots Hackbots Skeletonbots M
odbots Bolo bots Vladbots Spybots Softbots Warbots
Turing bots Spambots
35
Expert Systems

• Intelligent Agents
• Computerized agents that might...
• respond to verbal commands as if it were human.
• be a personal assistant that would access
  electronic communications.
• take phone calls.
• make appointments.
• locate individuals by phone.
• find research material.

36
Neural Networks

• Neuron Basic building-block of the brain.
• There are several specialized types, but all have
  the same basic structure
• The basic structure of an animal neuron.

37
Neural Networks

• Artificial models of the brain are of two
  distinct types
• Electronic Has electronic circuits that act like
  neurons.
• Software This version runs a program on the
  computer that simulates the action of the neurons.

38
Neural Networks

• Artificial neurons
• Commonly called processing elements.
• Modeled after real neurons of humans and other
  animals.
• Has many inputs and one output.
• The inputs are signals that are strengthened or
  weakened (weighted).
• If the sum of all the signals is strong enough,
  the neuron will put out a signal to the output.

Output
Inputs
Artificial Neuron
39
Neural Networks

• Neural Network
• A collection of neurons which are interconnected.
• The output of one connects to several others with
  different strength connections.
• Initially, neural networks have no knowledge.
  (All information is learned from experience using
  the network.)

Neuron 1
Input 1 Input 2 Input 3
Output from Neuron 1
Output from Neuron 2
Neuron 2
40
Neural Networks

• Training a Neural Network
• Supervised training
• Occurs when the neural network is given input
  data.
• The resulting output is compared to the correct
  input.
• The strengths of the connections are then
  modified so as to minimize errors in succeeding
  input/output pairs.

41
Complex Adaptive Systems

• Complex adaptive systems
• They are nonlinear systems. Very small changes
  can result in different outcomes.
• They are parallel rather than serial. They have
  many things happening at the same time that
  affect outcomes.
• They are evolutionary with natural selection
  involved.
• They have emergent behavior. Totally
  unpredictable results can occur.
• The basis of the complex system contains some
  very simple rules.
• They are self-organizing.
• Examples Ant colonies, economies of nations, the
  world economy, political systems, cultural
  systems, the ecological system.

42
Complex Adaptive Systems

• Chaos
• Described as a situation where things seem
  unorganized and unpredictable.
• Tiny changes in the starting point produce
  solutions to a problem that seem to have almost
  random results.
• Butterfly affect A tiny flip of a butterflys
  wings could start a hurricane.

43
Complex Adaptive Systems

• Artificial life (a-life)
• A phenomena in computers that has attributes of
  life.
• Some argue that computer viruses are a form of
  a-life.

44
Complex Adaptive Systems

• Genetic Algorithm (simulated evolution)
• Mimics the processes in the genetics of living
  systems.
• Created by John Holland (mid-1960s) U. of
  Michigan
• The human puts together the system and specifies
  the desired results, but the details on how it is
  done are left to evolve.
• Genetic Programming
• A technique that follows Darwinian evolution.
• The evolution takes place directly on the
  programs in the population that are striving to
  reach the goal specified by the programmer.