home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Shareware Overload
/
ShartewareOverload.cdr
/
educ
/
ai.zip
/
AI.TXT
Wrap
Text File
|
1986-03-24
|
19KB
|
369 lines
Making a Computer Get Smart
appeared in Insight Magazine, March 24, 1986
Scientists are teaching a computer to reason out problems and learn from
experience. In other words, they are trying to give a computer common sense.
They have had limited success, but just to teach a computer how to converse
requires a breakthrough.
----------
Inside a gleaming office complex in Austin, Texas, some of the nation's
brightest scientist, linguists and psychologists are trying to tutor a dumb
student.
The student is a computer, and a 24-member artificial-intelligence team is
spoon-feeding it thousands of scraps of knowledge, as well as giving it
grammar and vocabulary lessons. Their goal is to cram the machine - a
mindless array of thumbnail-sized silicon chips - with enough facts, rules of
thumb and human language skills that it may begin to think and learn on its
own.
Here, at the Microelectronics and Computer Technology Corp. (MCC), a joint
research and development venture backed by America's corporate giants, the
future is being built.
MCC is "pushing back the frontiers of science," says its chairman, retired
Navy Adm. Bobby R. Inman, who previously served as deputy director of the
Central Intelligence Agency.
Article by article, the team of researchers is dissecting an encyclopedia,
then encoding its contents into the computer's memory bank. For example, all
the facts presented in an article on flight are encoded, plus the underlying
knowledge about the world needed to understand the article.
They are feeding the machine thousands of bits and pieces of common sense: If
you're out in the rain, you get wet. If you drop something, it falls to the
ground. An object can't be in two places at once. Each person lives for a
single interval of time.
They also are teaching the computer about itself. "It has to understand that
it is a program," says a scientist. "It needs to know that a human being is
watching it." MCC, which began its high-stakes research in January 1984, is
owned by 21 U.S. companies, including Rockwell International Corp., Honeywell
Inc. and Martin Marietta Corp. MCC's goal is to create a variety of new
computer technologies for the 1990s and beyond - passing along the fruits of
its research to its shareholder companies to give them a head start over
foreign competitors in designing new products and services.
The $65-million-a-year project has resulted in a remarkably high degree of
cooperation between otherwise archrivals. At the Austin headquarters, one-
third of MCC's 410 employees are on loan from the various shareholder firms.
MCC has quickly emerged as one of the new heavyweights of artifical
intelligence (AI), the discipline that has already taught computers to play
chess and to help perform medical diagnoses.
Researchers at MCC and a handful of laboratories are trying to build the
prototype for a fifth-generation computer capable of reasoning its way through
myriad tasks in the home, at the workplace and on the battlefield.
But when asked to explain what makes machines "intelligent," a computer
scientist is likely to talk in circles.
" `Artificial intelligence' is trying to do things we don't know how to do
yet," says Marvin L. Minsky, a pioneer in artificial intelligence at MIT.
"But that's a working definition. It changes every year.
"Twenty years ago, having a machine recognize a picture or play chess or
understand simple language would have been out of reach," he says. "It's sort
of a moving horizon."
Even before the first generation of huge machines powered by vacuum tubes, men
dreamed of building a computer that could mimic human thought. But efforts
over the past 30 years to make such a computer have fallen short.
Powerful, number-crunching computers can analyze vast amounts of data, spit
out amazing mathematical solutions and guide an unmanned probe to the outer
reaches of the solar system. Yet these machines have no inkling of human
goals and beliefs, no sense of the world or their place in it.
Jonathan Slocum, MCC's director of natural language processing, believes that
words are a key to machine intelligence.
His reasoning is simple: A child's ability to learn about the world is closely
tied to his use of words as symbols. Digital computers have no grasp of the
meaning of words or what lies beyond them. And these machines will forever
lack common sense until they are able to communicate with, and learn from,
people.
But what might seem like a straightforward task - teaching English to a
computer by cramming it with grammatical rules, words and definitions - has
proved a monumental endeavor.
"We would be very happy if these machines were as effective as a 4-year-old
child with respect to the grammar," says Slocum.
Home computers can mimic verbal skills by using sentences to display a
problem's solution. But faced with interpreting sentences, advance computers
- which rely on limited vocabularies of narrowly defined words - break down.
Simple conversation, as it turns out, takes an enormous amount of information
processing at incredibly high speeds.
"We rarely perceive ambiguity in something someone says," says Slocum. But
"almost any sentence you hear a human being utter will be ambiguous."
Depending on the context, the word "ball" in a sentence could mean a dance, a
round object used in sports or a good time. Similarly, a simple sentence
might contain 10 words with three definitions each.
" We don't consciously review all the interpretations. Human beings select
one and go with it almost all the time," Slocum says. "If your confidence [in
your first interpretation] is high, you're not going to stop the speaker. If
your confidence is low, you may stop the speaker and ask whether he meant this
or that."
Slocum is writing a computer program in which his "linguist's intuition" is
encoded in plausibility scores: the mathematical probabilities for the
likelihood that a statement is true.
Dissecting a sentence, his computer program assigns plausibility scores for
the possible meaning of each word, and then applies rules for combining
plausibility factors as it examines each element. Future computers will
recognize, he says, when to accept at face value its first interpretation of a
sentence, when to ask for clarification and when to say, "I'm confused."
"Four-year-olds are quite good. They know most of the grammar that an adult
does," he says. "They don't know all the grammatical structures that exist in
the language, but they know a great majority of them."
It will take a major scientific breakthrough, he says, for computers to use
metaphors, idioms and similes. After all, how does a literal-minded machine
catch the meaning of phrases such as "cry a river of tears," "kick the bucket"
or "she is like a rose"?
What Slocum's computer program lacks in grammar skills, he hopes to bolster
with a working vocabulary of 20,000 words. Future computer programs, using
complete dictionaries of words and multiple interpretations, will have "vast
proficiency, out-stripping any human being," he says.
Meanwhile, MCC's artificial-intelligence team is bringing up its baby by
feeding the computer with more facts about humans, the world and itself.
The computer is a tabula rasa, a blank slate, says Douglas B. Lenat, an
artificial intelligence project director at MCC. "We're bootstrapping it up
to the point where it will be a reasonable student.
"The more you know, the more easily you can learn," he says. "If you start
out a [computer] program that knows next to nothing, it's hard for it to
assimilate new pieces of information.
"But children already know so much about the world that it's very likely that
they'll have something they can hook new experience onto and thereby relate,"
he says.
Future computers, he says, will examine a problem - for example, a battlefield
situation - and decide what problem-solving strategy to employ, "introspect"
to see if they are making progress and take another tack if needed.
"That kind of behavior leads to something that appears very much like
consciousness," he says, which he defines as "largely the ability to
introspect on what your doing."
Computer programs "have a form of consciousness," he says. "They have to be
conscious of why the do what they do. You can stop it at any point and ask,
`Why did you do that?' and it will tell you, after a fashion."
But a new generation of superfast computers with huge memory banks will be
required if machines are to learn English and become smarter.
Researchers at MCC and a few other U.S. and Japanese laboratories are
developing "parallel processors": networks of tens, hundreds and even
thousands of computer chips, each with a separate memory bank, that work in
concert to solve a problem.
Others are creating new computer languages for parallel computers that can
correctly divide problems into subproblems - for instance, examining different
parts of a sentence.
Lenat hopes that 10 years from now, MCC will have taken a giant step toward
building a new generation of machines with some degree of common sense. "But
until computers are smarter than they are now," he says, "most questions they
ask will be stupid."
It is only a matter of time before the smart machines arrive.
As the pace of new technology accelerates, scientists at the world's top
research laboratories are locked in a high-stakes computer race.
We are attempting to make machines capable of enormous amounts of knowledge,
more than any human can muster," says MCC's Slocum. "And these machines will
have enormous amount of power."
According to various experts, "thinking" machines of the 21st century may
range from huge data banks to crude vision, hearing and speech systems
installed in factory robots.
At home, future computers may act as general problem solvers, time managers
and consultants on matters ranging from taxes to medical problems to assisting
with a person's daily schedule. Truly user-friendly, these machines may speak
English and respond to spoken commands.
Industrial and government computers may be used for everything form handling
air traffic control systems at busy airports to finding ways to pay off the
national debt.
Commercial robots operated by computer programs may scour the ocean floor for
manganese nodules, clean radioactive spills and toil in hazardous coal mines.
Military robots may be ready to serve in other hostile situations like
chemical or nuclear war zones. Unmanned fighter aircraft, tanks, ships and
submarines may be activated simply by voice commands.
Powerful defense computers may help contain and resolve small conflicts
between nations by automatically interpreting 24-hour-a-day surveillance data
from various land, sea, air and space sensors - foiling the possibility of
suprise attacks.
In addition, spaceborne computers able to instantaneously detect, target and
defend against thousands of enemy warheads may provide the linchpin for the
Strategic Defense Initiative, The so-called "star wars" antiballistic missle
shield.
These computers would have to distinguish in seconds between real attacks and
routine space launches, as well as discriminate between decoys and warheads,
before automatically launching particle beams to neutralize enemy warheads.
No computer system currently exists that is capable of handling this awesome
task, but some defense experts believe that breakthroughs in "intelligent"
software and high-speed parallel computers could make the difference.
"over the next 100 years there will be fantastic progress," says Minsky, the
artificial-intelligence pioneer. "Machines will do miraculous things. I'm
sure they'll be flying airplanes better than people.
"To [use] a gruesome image, it's like slavery. Machines will do whatever
people want them to do."
The trick to building the most helpful computers is to endow them with common
sense and learning skills, he says, "so you don't have to program them
yourself. Getting them to think well is the problem. It's probably the
hardest problem science has ever faced. And it won't happen overnight."
The international race for computer supremacy may determine which nations hold
scientific, economic and military dominance well into the next century.
"The U.S hasn't lost its lead in the ability to create new technologies. But
our lead is somewhat diminished," says MCC's Inman.
MCC was created in the wake of Japan's announcement in 1980 that it would
launch a major, national effort to surpass the United States' technological
edge. Japan's Fifth Generation Project is aimed at building a "supercomputer"
for high-speed numerical calculations, as well as artificial-intelligence
software for a supersmart computer capable of solving non-numerical problems.
Joining the race, European technology firms also recently began their own
cooperative research project called Espirit. But already, some 500 "expert
systems" are at work outside the laboratory in the marketplace - performing a
wide range of tasks from monitoring nuclear reactors to pinpointing the
location of mineral deposits.
These primitive computer programs mimic human reasoning by quickly solving
routine yet highly complex problems involving expert knowledge in a particular
discipline.
Expert systems are helping to diagnose bacterial infections, to predict crop
irrigation needs, to design electronic circuits and to control the treatment
of patients in intensive care.
And expert software is available for personal computers to help people make
financial planning decisions and to negotiate successfully at home and on the
job.
By 1990, the industrial and home market for expert systems is expected to grow
to $2 billion in sales. Yet despite their success, expert systems have narrow
applications. Faced with anything outside their limited field of expertise,
these fragile programs get confused.
These systems emulate only a slice of human intelligence. They are unable to
learn from experience or to handle nonrecurring problems. The bulk of their
expert knowledge depends on simple sets of "if-then" rules, such as: "If a
child has blisters and if it has a fever, then it may have chicken pox."
An electronic cardiologist, for example, can offer near-human advice on
medical treatment. But it has no idea of what a patient is, let alone a human
being.
"If a patient has a disease that's outside the expertise of the expert system,
there's no telling what diagnosis it will come up with," says MCC's Slocum.
"if the doctor follows the diagnosis, some disastrous things could happen.
"Any reasonable person with a small amount of modesty knows when he's out of
his depth," he says. "We're trying to program computers to have a general
knowledge of the world, coupled with self-knowledge, so the machine can gauge
its limitations."
This "intelligent" software, he says, is designed to bolster expert systems
with "the common sense that a human being would bring to bear" on a problem.
A few experimental computer programs already have been built that analyze
novel problems, make guesses by using rules of thumb learned from analogous
situations and test their hypotheses by conducting experiments.
Meanwhile, the Defense Department is funding a variety of artificial-
intelligence projects, including the development of an unmanned vehicle.
Using machine vision, it may act as a mobile sentry or ferry supplies and
weapons across rough terrain.
The Defense Department also has funded research into voice-activated command
and control systems, vision systems to guide missiles to targets, and computer
aids for fighter pilots, tank drivers and battlefield commanders to help them
quickly make complex decisions.
Artificial-intelligence researchers, often voracious readers of science
fiction, envision a future in which man reaches new heights with the aid of
brilliant and not-so-smart machines. They see a future in which cheap
industrial robots perform mundane and hazardous jobs - working in unmanned
factories, repairing satellites in space and disposing of nuclear wastes.
Private industry is already using robots as spray painters, spot welders and
assembly-line inspectors. General Motors Corp. is developing machine vision
systems that may enable robots to move more like people in the workplace.
The new technology will not be without a human price. Imagine the
psychological impact on a highly skilled, middle-aged welder who is taken off
his job because a machine does it better.
A proliferation of smart computers and related technologies in advance nations
also may spark new political tensions with the Third World.
"The pace of change by emerging technologies is likely to be somewhat faster
over the next 10 to 15 years that it's been over the past 30 years," says
Inman. "In the early years, there's clearly going to be an even wider gap
between the haves and the have-nots."
MCC's Lenat forsees the eventual development of a "global knowledge base": a
vast information bank that people can tap into with portable links. Computers
will serve as "intelligence amplifiers to help us think faster and better than
we were ever able to," he says. "They will provide an opportunity to enhance
ourselves."
As progress in artificial-intelligence research accelerates, people may face
som mind-boggling questions. Could machines ever be capable of inspiration or
insight - that sudden flash of recognition you experience when you finally
solve a tough problem?
"I believe the process we are embarking on will demystify intuition," says
Lenat. "Intuition is often just analogizing to other situations to come up
with an approximate answer very quickly."
Will computers ever attain any sort of self-awareness? Could they be
programmed to know the meaning of love or honor? Could they ever appreciate
or even recognize humor? And finally, to what extent should we rely on these
machines?
"No more and no less than people whose insides we don't have access to,
either," says Lenat. "We [should] use performance to measure the degree of
trust."
"The net effect of [artificial-intelligence] will be positive. It will save
lives," he says. "It will increase human productivity, including the amount
of food in the world...all sorts of positive things and a few negative ones."
Computer systems that increase the warning time of military attack, he says,
will prolong the time for political response and negotiation.
"[And] if you're going to send out a tank," he says, "it's better that a human
is not in it."
- J.H. Doyle