Important Names and Aspects of Artificial Intelligence

1
Important Names and Aspects of Artificial
Intelligence

• Created by Brett Traylor

2
Herbert Simon

• Interested in Cognitive Psychology with
  particular emphasis on computer use to emulate
  human thinking processes
• developed the first AI software with Allen Newell
  called Logic Theorist (1955)
• Logic Theorist was constructed under the premise
  that the computer can solve problems by searching
  through a mass of programmed possibilities or
  theorems and subsequently seek to find a proof.
  As is the case with logicians, Logic Theorist did
  not always guarantee a solution.
• Theorems derived from Whitehead and Russell's
  work entitled Principia Mathematica
• In 1957, the invention of the General Problem
  Solver enabled the computer to learn from
  examples.
• Research, in part, revolved around trying to
  understand how problem solvers can discover more
  effective problem representations.
• All research involves comparison of computer
  simulations of human behavior with actual human
  behavior observed in lab experiments (comparison
  of the two with respect to performance of the
  same task).

3
John McCarthy

• Established first AI Laboratory in 1957
• Created AI programming language called LISP
• LISP allows computing to be done with symbolic
  expressions instead of numbers
• Symbolic expressions are represented by list
  structure in the memory of computers
• Information is represented in external media
  mostly by multi-level lists
• LISP is truly remarkable in that it has been in
  existence for over 20 years
• Many futile attempts to replace LISP

4
Moores Law

• Founder of Intel
• Moores Law posits that the amount of data that a
  microchip can store doubles annually.
• Over the last 20 years, Moores projection has
  become prophetic as we have witnessed the reduced
  cost and increased performance of computers.
• The advancement of deep submicron manufacturing
  technology implies that Moores Law should bear
  validity indefinitely.
• Information handling capacity in computers has
  evolved approximately ten million times faster
  than did our nervous system. The power doubled
  every two years in the 1950s, 1960s, and 1970s,
  doubled every 18 months in the 1980s (Moores
  Law), and is now doubling each year.

5
Neural Networks

• Work under the assumption that computers can
  mimick the operation of the human brain
• Approach to understanding the human brain
• Attempt to recreate the environment of the human
  brain
• First artificial neuron constructed by Warren
  McCulloch and Walter Pits (1943)
• Network is composed of a large number of highly
  interconnected processing elements
  (BrainneuronsNeural Networksnodes) working in
  parallel to solve a specific problem
• Neural networks learn by example
• Unlike conventional computers, Neural networks
  cannot be programmed to perform a specific task
• Like the neurons of the brain, nodes have many
  inputs having different weightings and one output
  which depends on the inputs. Both biological
  neurons and artificial neurons can either fire
  or not fire
• Now used successfully in speech, language, and
  image recognition
• Neural networks are faster and more adaptive than
  expert systems.
• While expert systems have been programmed with
  specific outputs for sets of inputs, neural
  networks have the power to generate outputs from
  inputs that are similar to those seen before but
  that are not exactly the same. (Ex Medical
  Diagnoses)

6
AI in Games

• Perhaps the oldest facility for AI
• Good example of strong AI
• Most popular
• Chess
• Checkers
• Othello
• Half-Life

7
Sources

• http//peace.saumag.edu/faculty/kardas/courses/CS/
  Student20Pages/AI/ComputersAI99
• Herbert A. Simon-- http//www.psy.cmu.edu/psy/facu
  lty/hsimon/hsimon.html
• LISP http//www-formal.stanford.edu/jmc/lisp20th.
  html
• Moores Law http//jeffsutherland.org/objwld98/fu
  ture.html
• Neural Networks http//www.doc.ic.ac.uk/nd/surpr
  ise_96/journal/vol1/ds12/article1.html
• Neural Networks http//www.fortunecity.com/skyscr
  aper/millenit/262/neural.html
• Computer Games-- http//www.ai-center.com/projects
  /excalibur/documentation/intro/aigames.html