Artificial intelligence and neural networks - The Computing Universe: A Journey through a Revolution - page 266

Since 1991, a New York businessman, Hugh Loebner ( Fig. 13.2 ), has
sponsored an annual Turing Test competition; and in 2012, the centenary of
Turing's birth, the contest was held at Bletchley Park. In more than twenty
years of competition, no chatbot program has come close to deceiving a sophis-
ticated judge.
An everyday demonstration of a computer's inability to pass something like
a Turing Test is a reverse version based on recognizing distorted letter shapes.
To pass this reverse Turing Test, a computer would need highly developed per-
ceptual abilities that are currently beyond the capability of the most advanced
computer vision algorithms. These puzzles were called CAPTCHAs ( Fig. 13.3 )
by Luis von Ahn ( B.13.3 ), an acronym standing for “Completely Automated
Public Turing test to tell Computers and Humans Apart.” Humans can easily
recognize the distorted letters, so CAPTCHAs enable websites to distinguish
between human users and automated “robot” programs trying to access the
site. It is estimated that more than two hundred million CAPTCHAs are solved
every day.

Fig. 13.2. The Loebner Prize for $100,000
was established in 1990 for the AI sys-
tem that first passes the Turing Test.

Fig. 13.3. CAPTCHAs can be easily
read by a human, but not by a com-
puter. This is one commonly used
mechanism to distinguish between
human visitors to websites and robotic
crawlers.

From logic theorist to DENDRAL

The term artificial intelligence was coined by John McCarthy ( B.13.4 ) in a
workshop at Dartmouth College in New Hampshire in 1956. McCarthy and
fellow AI pioneers Marvin Minsky, Claude Shannon, and Nathaniel Rochester
wrote a proposal for the workshop stating:

The study is to proceed on the basis of the conjecture that every aspect of
learning or any other feature of intelligence can in principle be so precisely
described that a machine can be made to simulate it. An attempt will be made
to find how to make machines use language, form abstractions and concepts,
solve kinds of problems now reserved for humans, and improve themselves.
We think that significant advance can be made in one or more of these
problems if a carefully selected group of scientists work on it together for a
summer. 8

The highlight of the Dartmouth workshop was a reasoning program developed
by Allen Newell and Herbert Simon ( B.13.5 ) from Carnegie Tech, now Carnegie
Mellon University. Their “Logic Theorist” program was able to prove theorems
using simple symbolic logic. It represented each problem as a tree structure
with the root being the initial hypothesis , a tentative explanation that could
be tested by further investigation. Each branch of the tree was a deduction
based on the rules of logic. To prevent the tree from growing uncontrollably,
Newell and Simon needed a way to remove unwanted branches. To do so, they
introduced heuristics , rules of thumb that enabled the program to select only
those branches of the overall search tree that seemed most promising. They
said, “Logic Theorist's success does not depend on the 'brute force' use of the
computer's speed, but on the use of heuristic processes like those employed by
humans.” 9
In their monumental work Principia Mathematica , Alfred Whitehead and
Bertrand Russell had attempted to systematize the principles of mathematical

B.13.3. Luis von Ahn is an asso-
ciate professor at Carnegie Mellon
University in Pittsburgh. He is per-
haps best known for his invention of
CAPTCHAs, those annoying distorted
characters that only humans, not
computers, are supposed to be able
to read.