1. Page 1
Arthur C. Clark and Stanley Kubrick made one and only mistake in their piece
of work “2001: A Space Odyssey”, and that was the date.
Marvin Minsky, the great intellectual and guru in the field of artificial
intelligence, gazes the future of machines with optimism. “One day”,
he maintains, “machines will master the ability of thinking, whatever
the term “think” might mean.” When interviewed by the New
York Times he did not hide his enthusiasm. He stated that he is a cinema fan,
and that he never misses any science fiction films: “They introduce ideas”,
he says. “The Terminator’ and the ‘Total Recall’, presented
ideas about memory transplant. Quite crude ones, but I loved the mechanics.” He
also reads science fiction literature. “There are dozens of really rich
sources of inspiration in these books”, he claims. “Gregory Benford,
David Brin and Larry Niven are the best writers of our times. In their writings,
there is always a new brilliant idea about something. I’ve also borrowed
ideas from older science fiction writers like Robert Meimlein and Isaac Asimov…” His
view about the first female science fiction writer, Mary Shelley, ( Frankenstein),
boils down to the fact that “she was right in her forecast that ‘humans
would never understand the poor creature.’ It is such a sad story! By
the way, I read the book once again, more thoroughly this time, to trace any
hints on how that robot functioned. Unfortunately there was nothing in it and
the funny thing is that when you read the novel you don’t care…”
Marvin Minsky has been amongst the pioneers of artificial intelligence. At
the end of the 1950s, when the term did not exist, he originated a research
program along with John McCarthy, a professor of Informatics at Stanford University,
which was to become the famous MIT Artificial Intelligence Laboratory. Nowadays,
almost forty years later, he is looking back into the field he helped to shape:
“We got stuck”, he says. “Artificial Intelligence has made
it into creating a bunch of wonderful things…programs that can do better
than an average stockbroker, programs that can amend certain things…
2. Page 2
Around 1980, progress ceased and mankind took several paths in an effort to
find new directions.
A good example is a student of ours-now the vice-president of the Xerox Corp.-called
Daniel Bobrow, who in 1964-65 made up a program that could read a question
from an algebra schoolbook and in some occasions come up with solutions to
problems. So, it was able to detect and comprehend a bit of language along
with some algebra. It could not solve most problems because it was unable to
grasp the meaning of words. Those days, we’d been trying to make a program
that could read a passage from a first or second grade elementary school textbook
But the whole thing went like this: For every individual text, we had to include
in the program all the knowledge it needed to read the specific text. We did
not come up with many problems connected with grammar, but each time there
was reference to something unknown to the program, the system collapsed…
An MIT student had a text about someone’s daughter who had been abducted
by the Mob, ransom being asked. He asked the program: ‘What should we
do?’ The program could not comprehend the question. He finally asked ‘Why
must he pay money to get back his daughter?’ Then, the program was able
to understand bits from the language and pieces from algebra. It did not understand
texts as it was unable to comprehend the meaning of words.”
In the end, what really computers miss is the thing we call common sense.
Yet, what is the so-called “common sense” in the final analysis?
“It is the ability to know about 30 or 50 million facts about the world”,
says Marvin Minsky, “and have them arranged in such a way that will allow
you to find analogies. Common sense won’t let you arrange things in a
rigid order. You store them (in the brain) in accordance with their utility
or what they might remind you of. For example, one can interpret a suitcase
as an object that can be used to step up and change a light bulb instead of
a means to carry things in…
A stick can be used to push things rather than pulling them. Using a line,
one can pull things, not push them. This is what common sense is all about,
but there is not a single computer in this world that is capable of mastering
this piece of knowledge…”
3. Page 3
A big issue having to do with the field of artificial intelligence, or, even
more, with artificial intelligence in the domain of science fiction is emotions.
Since emotions constitute a function of the brain, will computers ever be able
to experience them? The dispute between Marvin Minsky and the other outstanding
figure in artificial intelligence, John McCarthy, involves the emotional part
of intelligence. The latter claims that we are in a position of creating clever
machines, not necessarily bestowed with human feelings. Minsky, on the other
hand, believes that such a hyper-computer like HAL, the one in “Space
Odyssey” which will be capable of envy is attainable. “The solution
to all problems”, he says, “demands resources. When resources are
limited…”
According to Minsky each and every emotion is nothing but a specific array
in our brains: “If you take a look at any of the books referring to the
human mind you will see that it looks like a huge arrangement of switches…
Love means nothing else than 20-30 of these switches being turned on and
some others being switched off. It is a specific arrangement…Anger is
another one…”
MARVIN MINSKY…
…is a member of the faculty at the Massachusetts Institute of Technology.
He also teaches Mathematics at the University of Princeton and is a pioneer
in the field of artificial intelligence. He has been awarded with dozens of
prizes in Science and has written Society of Mind, which has been a bestseller.
Today he is working at the MIT Media Lab and is writing the sequel to the Society
of Mind under the title: Emotion Machine…