Artificial Intelligence

1
Artificial Intelligence
Chester
2
What is AI?

• Artificial Intelligence is the mental capacity
  exhibited by an artificial (non-natural,
  man-made) entity to reason, plan, solve problems,
  think abstractly, comprehend ideas and language,
  and to learn.
• Intelligent Agent (IA) When an AI system
  interacts and learns from its working
  environment, a software agent that shows a kind
  of artificial intelligence.

3
Characteristics of An Intelligent Agent

• When interacting with the environment it should
  be able to learn and improve.
• To be able to adapt when it is connected to a
  larger network and also to adapt in real time,
  this means that there is a time deadline for the
  event to occur.
• To be able to learn quickly after absorbing a
  large amount of data.
• To accommodate new rules and ideas for problem
  solving.
• Has an ideal memory capacity that can remember
  and retrieve data and knows what to do with it
  without needing human input.
• Has parameters and variables to represent things
  like short and long term memory, age, and
  forgetting among others.
• To be able to analyze itself in things like how
  it behaves, errors its made and successes it has
  achieved and is able to adapt accordingly.

4
History of AI1950s

• Began in the 50s as an experimental field.
• Allen Newell and Herbert Simon founded first AI
  lab at Carnegie Mellon University in Pittsburgh,
  Pennsylvania.
• 1956 John McCarthy coins the term Artificial
  Intelligence at the Dartmouth Conference.
• 1956 Allen Newell, J.C. Shaw and Herbert Simon
  demonstrate first AI program called Logic
  Theorist.
• 1958 John McCarthy invents Lisp programming
  language.
• 1959 John McCarthy and Marvin Minsky founded the
  Massachusetts Institute of Technology (MIT) AI
  Lab.

John McCarthy
5
1960s

• 1961 James Slagle from MIT used Lisp to write
  SAINT which could solve college freshman level
  Calculus problems.
• 1963 ANALOGY is written by Thomas Evans. It
  showed that computers could solve analogy
  problems found on IQ tests.
• 1964 Danny Bobrow from MIT showed that computers
  could understand natural language good enough to
  solve algebra word problems.
• 1967 Richard Greenblatt from MIT writes
  MacHack the first knowledge-based program that
  plays chess and could achieve a class-C rating
  playing in tournaments.
• 1969 Shakey the Robot, the first mobile robot
  is built at the Stanford Research Institute (SRI).

6
1970s

• 1971 Terry Winograd from MIT combines his
  SHRDLU program with a robotic arm. Demonstrated
  that computers can understand and act on English
  sentences in a world restricted to childrens
  blocks.
• 1972 Alain Colmerauer and Robert Kowalski
  develop the Prolog programming language as an
  alternative to Lisp.
• 1973 Freddy Robot was built at the University
  of Edinburgh, it used visual perception to locate
  and assemble models.
• 1973 Lighthill report is published, it gave a
  negative verdict on AI research which prompted
  the British government to no longer support
  research except for at two universities.

7
1970s continued

• 1974 Mycin is presented by Edward Shortliffe,
  it could diagnose infectious blood diseases and
  recommend treatment.
• 1978 Mark Stefik and Peter Friedland demonstrate
  MOLGEN, showed that object-oriented programming
  representing knowledge could plan gene-cloning
  experiments.
• 1979 The Stanford Cart is built by Hans Moravec.
  It is the first autonomous vehicle and
  successfully crosses a room filled with chairs
  acting as obstacles and circumnavigates the
  Stanford AI Lab all without human guidance.
• 1979 The American Association for Artificial
  Intelligence (AAAI) is created. Its first
  national conference is held at Stanford in the
  early 80s.

The Stanford Cart
8
1980s

• 1982 The Fifth Generation Computer System
  project begins. It is a initiative by Japans
  Ministry of International Trade and Industry
  aimed at making the first fifth generation
  computer with AI attributes.
• 1989 Dean Pomerleau creates ALVINN (An
  Autonomous Land Vehicle in a Neural Network), it
  managed to drive a car coast-to-coast in the
  United States under computer control for 2800 out
  of 2850 miles.

9
1990s and Beyond

• Major advances in every field of AI.
• Gerry Tesauro creates TD-Gammon a backgammon
  playing program that showed that reinforcement
  learning is a good enough to challenge
  world-class championship players.
• 1993 Fifth Generation Computer System project
  is cancelled.
• 1997 Deep Blue, a chess playing program from
  IBM beat world champion Garry Kasparov. First
  RoboCup robot soccer tournament is held with 40
  teams entered.
• Late 90s Web crawlers become essential for
  search engines to be effective by allowing fast
  searches responses.
• 2000 Robopets are introduced, they behaved
  like real pets and could interact with people.

10
Example of AI TodayThe Sims 2

• The AI must keep the needs of each sim satisfied
  by having them interact with objects that will
  fill that need.
• AI determines who gets along with who based on if
  they share the same interests and if their
  relationship is ready for the social interaction
  being done.
• AI reacts to important events appropriately by
  being happy during happy events and sad for sad
  events, etc.

11
Social Issues

• If we push the development of technology too far
  we may find ourselves in a situation similar to
  what the humans in movies like The Matrix, I,
  Robot, Terminator, and Stealth are in. We
  may have to deal with rogue and hostile robots
  who have developed enough to be as smart if not
  smarter than we are.
• Sometimes it is better to have the human touch in
  some situations instead of the calculating
  precision of machines. It will be very difficult
  and possibly even impossible to program in human
  traits such as compassion, pity, charity,
  remorse, etc.

12
Conclusion

• AI has the potential to do great good for us or
  it may overcome and destroy us.
• We should be careful and aware of how far we are
  willing to go in advancing AI and understand what
  the implications might be, because playing God
  should not be taken lightly.