Dear thoughts-ga,
Good day!
Introduction
The idea of creating intelligent machines has always fascinated
mankind. The Greek mythology also
dwelt with the concept of intelligent machines when Talos was made by
Hephaestus (as a gift for
Europa from Zeus). Talos was a man of bronze whose duty was to patrol
the beaches of Crete and
thwart invaders by hurling great rocks at them, or by heating himself
red hot and squeezing
trespassers in a warm embrace.
The development of the electronic computer in 1941 meant that
technology was available to create
machine intelligence. The term 'Artificial Intelligence' was
originally coined by John McCarthy on
August 31, 1955 in his proposal to organise a two month summer
workshop at Dartmouth college to
study artificial intelligence.
At a fundamental level, We need to define intelligence to understand
the concept of AI. And this has
been one of the most difficult problems faced by AI researchers as,
even today, there is no single
definition of intelligence. A classic solution was proposed by the
British computer scientist Alan
Turing in his 1950 paper titled "Computing Machinery and
Intelligence". In the "Turing test", a
human being and a computer would be interrogated under conditions
where the interrogator would not
know which was which, the communication being entirely by textual
messages. He argued that if a
computer could successfully pretend to be human to a knowledgeable
observer, then it should
certainly be considered intelligent.
The fledgling field of AI caught people's attention and the
expectations started to build up. The
scientific community also got carried away with the hype. For example,
in the 1960s, scientists
working in the field predicted that machines would be smarter than
people within 20 years. By 1970s,
AI had begun to specialize into areas such as expert systems, language
analysis, knowledge
representation and computer vision.The popular media and entertainment
continued to stoke the common
man's fantasies for intelligent machines by presenting, variously, the
benign as well as the malign
side of thinking machines. The hype curve of the AI had reached its
peak in 1986 and then the boom
went bust! The accumulated weight of unrealistic expectations from AI
and the subsequent "broken
promises " pushed the field into a deep trough of disillusionment.
The Gulf War in the early 1990s were the first testbed for AI, where
it was used for simple tasks
(e.g., loading transport planes) as well as complicated tasks (e.g.,
the timing and co-ordination of
Operation Desert Storm). Advanced weapons such as "cruise missiles"
used AI technologies such as
Robotics and Machine Vision.
Today, in the 21st century, the realization has dawned that while it
may never lead to such
"intelligent" machines, AI has already permeated our lives in small
but significant ways.
Q1)
The American Association for Artificial Intelligence defines AI as
"the scientific understanding of
the mechanisms underlying thought and intelligent behavior and their
embodiment in machines." John
McCarthy defines AI as "the science and engineering of making
intelligent machines, especially
intelligent computer programs. It is related to the similar task of
using computers to understand
human intelligence, but AI does not have to confine itself to methods
that are biologically
observable."
However, before we move forward, one thing needs to be understood very
clearly - Artificial
Intelligence as a branch of science is not only about computers. AI
overlaps with several other
disciplines - from machine vision to expert systems. If we draw
paralles between AI and intelligent
animals, then AI overlaps with several older disciplines, including,
for instance, psychology,
neuroscience, philosophy, logic, and linguistics.
The major branches of AI are as follows:
a) Logical AI - here What a program knows about the world in general,
the facts of the specific
situation in which it must act, and its goals are all represented by
in mathematical logical
language. The program then decides what to do by inferring that
certain actions are appropriate for
achieving its goals,
b) Search - involves examining large numbers of possibilities, e.g.
moves in a chess game, and
learning to do this more efficiently in various domains
c) Pattern recognition - involves comparing observations with a
pattern; e.g., a vision program may
try to match a pattern of eyes and a nose in a scene in order to find a face
d) Representation - representing facts about the world using languages
of mathematical logic
e) Inference - arriving at conclusions from some facts
f) Common sense knowledge and reasoning - the area in which AI is
farthest from human-level;
incorporating common sense into AI programs
g) Adaptive learning - the approaches to AI based on connectionism and
neural nets specialize in
this branch; once the program performs a specific task, it embodies
the knowledge gained to better
itself
h) Planning - these AI programs generate a strategy for achieving the
goal from a knowledge of
general facts about the world (especially facts about the effects of
actions), facts about the
particular situation and a statement of a goal
i) Epistemology - a study of the kinds of knowledge that are required
for solving problems in the
world
j) Ontology - study of the kinds of things that exist
k) Heuristics - a way of trying to discover something or an idea
imbedded in a program
l) Genetic programming - a technique for getting programs to solve a
task by mating random LISP
(Limited Instruction Set Programming - one of the most popular
languages to develop AI applications)
programs and selecting fittest in millions of generations
Today, AI is a crucial ingredient of the extraordinary performances of
a large number of the biggest
companies. An indicative use of AI by businesses is given below:
a) to develop better jet engines,
b) to monitor equipment on the shopfloor and signal when preventive
maintenance is needed,
c) to coordinate immense logistics operations,
d) to gain new insights into the tremendous amount of data on the
human genome, etc.
Q2)
Expert systems are computer programs that apply AI techniques to the
problem solving process and
perform decision-making tasks based on a programmed set of rules
(rules base) and logic (knowledge
base) within specific subject areas (facts base). In common parlance,
expert systems are
computer-based systems designed to mimic the performance of human
experts. These systems were first
developed around 1975.
Some of the common examples of the use of expert systems are as follows:
a) insurance underwriting,
b) investing in stocks,
c) study the effects of the depletion of ozone layer on the global environment,
d) simulate the effects of a nuclear bomb, etc.
There are two basic approaches to AI - bottom-up and top-down. The
researchers in the first "camp"
believe that the best way to achieve AI is to create electronic
replicas of human brain's complex
network of neurons, while the researchers using the top-down approach
attempt to mimic the brain's
behavior with computer programs. Expert systems take the top-down
approach. They use the huge data
storage capacities of modern computers to store facts, then interpret
these statistics to formulate
rules and then apply these rules to the problem. They ork very much
like a detective who solves a
mystery by applying logic and rules (to the facts of the current case)
that he has developed from
his previous experiences.
Q3)
Today, A.I. is gradually moving more and more into people's everyday
lives, especially as the
interest in computers and computer games grows.
The major application areas of AI in society are as follows:
a) Game playing - programming computers to play games such as chess.
This is one of the most popular
branches today
b) Speech recognition - recognizing speech inputs and taking the necessary action
c) Natural language processing - programming computers to understand
natural human languages, i.e.,
understanding the context in which a word is spoken
d) Machine vision - digitisation, manipulation and analysis of images;
increasingly, the images are
being visualized in all the three dimensions
e) Expert systems - Programming computers to make decisions in
real-life situations using
information and logic (or rules)
f) Heuristic classification - putting some information in one of a
fixed set of categories using
several sources of information, e.g., advising whether to accept a
proposed credit card purchase.
g) Neural Networks - Systems that attempt to simulate intelligence by
reproducing the types of
physical connections that occur in animal brains
h) Robotics - Attempts to program robots to act intelligently, e.g.,
to sense and react to sensory
stimuli such as sight, sound, etc.
Some of the most famous implementations of AI applications are as follows:
a) SmartAirport Operations Center, a logistics program created by
Ascent Technology, uses adaptive
learning (with genetic algortithms) to handle the entire logistics of
airports. It has raised
productivity by up to 30 percent at every airport where it's been
implemented. In fact, the Pentagon
used this system to manage the logistics of Operation Desert Storm!
b) Monster.com uses an intelligent Web crawler called FlipDog to find
new customers. Wandering the
Web, Using text parsing technology, FlipDog develops a sense for which
parts of sites are more
likely to contain jobs, then parses the pages to pull out the relevant
information (company, salary,
kind of work, address for sending a résumé) and files it in a
database. The first time the crawler
ran, it came back with more than half a million jobs. The real feat
was not that FlipDog found the
postings, but that it was able to organize them
c) The Falcon program, a pattern recognition system designed by San
Diego-based HNC, is used by 9 of
the top 10 US credit card companies; they claim that Falcon has
improved fraud detection rates from
30 to 70 percent. The program maintains a perpetually microadjusting
profile of how, when, and where
customers use their credit cards.
e) An after-hours tech support program developed by Handspring uses
speech processing to handle
customer queries. The program identifies phonemes, or letter sounds,
within a spoken sentence and
assembles them into a variety of possible words using statistical
analysis. It discards "noise"
words ansd retains keywords.
Q4)
An expert system offers the following advantages over a human:
a) it can do the work of a professional
b) it can be trained quickly
c) it can also learn intuition
d) it has virtually no operating cost
e) it does not forget what it learns
f) it never reports sick
g) it never retires
h) it never goes on vacation
Most importantly, expert systems can consider a large amount of
information that may not be
considered by humans.
In some fields such as weather forecasting, bug tracking, etc., expert
systems are sometimes more
accurate than humans. But for other fields, such as medicine, the
human doctor should not be
replaced. Expert systems have the power and range to aid to benefit,
and in some cases replace
humans, and computer experts, if used with discretion, will benefit human kind.
Some key examples of the use of expert systems are given below:
a) Wal-Mart consolidates point-of-sale data from its 3,000 stores and
uses data-mining systems to
sift instantly through the data and uncover patterns and
relationships. Data-mining systems include
neural nets, statistical analysis, and expert systems with if-then
rules that mimic the logic of
human experts. The results enable Wal-Mart to predict sales of every
product at each store with
uncanny accuracy, translating into huge savings in inventories and
maximum payoff from promotional
spending.
b) FocalPoint, an expert system, examines Pap smears for signs of
cervical cancer. Built by TriPath
Imaging, the system screens 5 million slides each year or about 10
percent of all slides taken in
the US.
Hope the above answers are to your satisfaction!
Thanks and regards,
reeteshv-ga
Google Answers researcher
Additional links:
A good collection of resources on AI is available at the website of
the American Association for
Artificial Intelligence:
http://www.aaai.org/AITopics/html/overview.html
An introductory article on the applications of AI (mainly robotics)
can be found on BBC's Hot Topics
section:
http://www.bbc.co.uk/science/hottopics/ai/
"21st Century AI - Proud, Not Smug" by Tim Menezies (May/June 2003)
IEEE Computer Society
http://www.computer.org/intelligent/ex2003/x3018.pdf?SMSESSION=NO
"It's Alive!" by Jennifer Kahn (Mar 2002)
Wired
http://www.wired.com/wired/archive/10.03/everywhere.html
"Smart Tools" by Otis Port with Michael Arndt and John Carey (Spring 2003)
Business Week
http://www.businessweek.com/bw50/content/mar2003/a3826072.htm
"What is Artificial Intelligence" by John McCarthy (March 29, 2003)
Stanford University
http://www-formal.stanford.edu/jmc/whatisai/whatisai.html
Open University
In the debate on March 15, 2002 in The Next Big Thing, the panelists
looked at the issue of AI
http://www.open2.net/nextbigthing/ai/ai.htm
"An Introduction to the Science of Artificial Intelligence" by Tim
Dunn, Adam Dyess, Bill Snitzer
Thinkquest.org
http://library.thinkquest.org/2705/
The Alan Turing Internet Scrapbook describes the Turing test in short
http://www.turing.org.uk/turing/scrapbook/test.html
An exhaustive introduction to machine vision can be found here:
http://www.eeng.dcu.ie/~whelanp/resources/r_preface.html
Search Strategy:
"Artificial Intelligence"
://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&safe=off&q=Artificial+Intelligence&spell=1
define: expert systems
://www.google.com/search?sourceid=navclient&ie=UTF-8&oe=UTF-8&q=define%3A+Expert+systems
"Machine Intelligence"
://www.google.com/search?hl=en&lr=&ie=UTF-8&oe=UTF-8&safe=off&q=machine+vision |
