|
|
 |
|
| |
|
|
|
 |
REFRAIN
It’s
only a matter of time before machines will be able to download
software that would turn them smarter than humans, writes Raj
Kaushik
In Spielberg’s film A.I.: Artificial Intelligence,
there’s a young robot that doesn’t just follow instructions but also
thinks independently. Now, most scientists feel that in the near
future, it might be possible to create an artificial brain endowed
with the power, emotion and flexibility of humans.
“At this
moment, computers show no sign of intelligence. This is not
surprising, because our present computers are less complex than the
brain of an earthworm. But it seems to me that if very complicated
chemical molecules can operate in humans to make them intelligent,
then equally complicated electronic circuits can also make computers
act in an intelligent way,” predicted Stephan Hawking in 1998. Dr
Hugo de Garis, head of Starlab’s Artificial Brain Project, Belgium,
believes that an artificial human-like, even smarter, brain is an
attainable goal, thanks to Moore’s law.
In 1965 Gordon Moore
observed an exponential growth in the number of transistors per
integrated circuit and predicted that this trend would continue.
Interestingly the electronic performance of chips is doubling every
year, whereas it took a million years for the human brain to double
its capacity. Who imagined that one day a computer would beat the
best human brain in chess?
In 1997, Deep Blue, an IBM
supercomputer, defeated Garry Kasparov, then world champion. Deep
Blue won two games, lost one and drew the other three. As
developments in the area of artificial intelligence are going on at
a predicted rate, it’s only a matter of time before machines will be
able to download software that would turn them smarter than humans.
Unsuccessful attempts to develop an artificial brain go as far
back as 1940s when scientists tried to create a machine using a
model of the neuron taken from the spinal cord. But the science of
artificial intelligence experienced a radical breakthrough in 1950
when Alan Turing invented a test that could be used to designate a
machine as conscious or non-living.
In the Turing test, you talk
to two terminals: one hooks to a computer and the other interfaces
with a person who types responses. You can ask questions, make
assertions and explore feelings and motivations for as long as you
wish. If you fail to determine whether a computer or a human is
working behind the scene, the computer passes the test and can be
designated as “conscious”.
In 1993, ATR Human Information
Processing Research Laboratories in Japan, under the leadership of
Hugo de Garis, embarked on an ambitious project called the “CAM
(Cellular Automata Machines)-Brain Project”. The goal was to build
an artificial brain with a billion artificial neurons by 2001.
Cellular Automata is a type of logical system composed of simple
objects whose state is determined by following simple rules about
the state of fellow objects; for example, high school girls will
wear tomorrow the same kind of revealing dresses that actresses wear
today.
In 1999, the ATR lab developed a brain for a robot cat
called Robokoneko (Robot Kitten in Japanese). The brain consisted of
37.7 million artificial neurons and 32,000 cellular automata (CA)
based neural circuit modules. Initially, the connections between
neurons were random, but in a dynamic, Darwinian-like process of
evolution, the robocat was able to discard nonsense processes and
develop only the successful ones. In 2001, Russian scientists
claimed to have developed the first artificial brain, a
“neuro-computer” with the same intellectual potential as its human
counterpart.
Vitaly Valtsev, a member of the International
Academy of Information Science, attributed the success of the
Russian attempt to the fact that they used a model of the neurons in
the brain in building the computer. Not all scientists agree with
this. Professor Paul Coleman of the University of Rochester, NY,
believes that Russian claim is overstated, particularly as we don’t
yet know how the real brain works. True, new technologies such as
MRI scanning have enabled scientists to understand the neurological
processes in ever finer detail, laying a theoretical foundation to
explain precisely how the brain work, however, remains a mystic and
insurmountable task.
In a recent development, a California-based
company Artificial Development activated a computer persona named
“Kjell” (after Alan Turing’s lover). Though still lacking the visual
and auditory cortex areas, Kjell realistically simulated the human
brain’s cortex, motor and somatosensory areas. The hippocampus,
basal ganglia and thalamic systems are under development. The system
is composed of 20 billion neurons capable of making 20 trillion
connections. It runs on a Linux-based hardware cluster of 500
dual-CPU nodes with a total of 1 terabyte of RAM and 200 terabytes
of storage.
In his next logical step, Hugo de Garis undertook a
project of constructing a 75 million neuron artificial brain. The
massively parallel neural network consists of roughly one million
modules of cellular automata, which grow and evolve at electronic
speeds inside a CAM Brain Machine (CBM). The CBM updates cellular
automata cells at a rate of 130 billion per second, and can evolve a
fully-operational neural net module in about one second.
The
first generation of artificial brain is already on display. The
second generation of artificial brains equipped with 10 billion
neurons will be completed around 2006; the third generation with
1000 billion neurons is projected to be completed by 2011. The
artificial brains of third generation are supposed to surpass the
complexity of the average human brain. The Defense Advanced Research
Projects Agency (DARPA) in the USA spends $29 million per year on
its Perceptive Assistant That Learns programme. The goal of the
programme is to develop thinking robots by 2030. But the agency is
likely to exhibit elements of artificial intelligence in robots,
such as reasoning and speech recognition, sooner.
Let’s face it,
super brainy machines are around the corner and they’re here to
stay. Both scientists and politicians are not comfortable with the
idea of machines subjugating humans. In June 2001, the French Senate
brought together leading scientists to ponder over the question,
“The Robot: The Future of Man or Man of the Future?” The focus was
on learning about the potential impact artificial intelligence will
have on future society. Emerging advances in artificial intelligence
are likely to bring society at a crossroads. The most active
scientists in the area are also not worry-free. Hugo de Garis, for
example, reveals he is haunted by the prospect that his creations
might “swat me like a fly”.
Stephan Hawking says that there is no
need to panic. His suggestion is to improve human intelligence
through genetic engineering, or maybe by interfacing ourselves to
computers directly. “We must develop as quickly as possible
technologies that make possible a direct connection between brain
and computer, so that artificial brains contribute to human
intelligence rather than opposing it,” said Hawking. The future of
humans lies in nurturing harmonious relationship with brainy
computers. It is mandatory to teach them the value of peace and
friendship. If they were engaged in negative activities – like
spying, bomb building and terrorism, they will decide our destiny.
(A former project coordinator with the National Council of
Science Museums, India and exhibits manager at the Discovery Centre,
Halifax, the author now works as a senior server developer in
Toronto.)
|
|
| |
 |
|
|
