Scientists close to creating simulated intelligence

intelligenceScientists have created a computer program that approximately passed the test for intelligence.  In other words, science is on the verge of creating simulated intelligence.

Plan developed by almost managed to pass the well-known Turing test, developed by British mathematician Alan Turing in 1950.  The essence of the test is that one or more people should ask questions to two secret interlocutors, and the answers should show which of them a car was, and who the person.  If the machine can not masquerade as a man assumes that it is reasonable.

In his paper “Computing Machinery and Intelligence,” Turing test called the “imitation game.”  According to scientists, if the machine is able to deceive inspectors in response to 30% of the questions asked, you can say that it has artificial intelligence.

On the occasion of the centenary of the birth of mathematics at the University of British town of Reading was organized by the so-called “Turing marathon”.  The competition involved several teams.  The winning program simulates a 13-year-old boy, Eugene.  The judges took a human response “robot” by 29.2% of all the questions, so before winning features were just 0.8%.

Of course, the program until he was recalled humanized robots from the movies “Artificial Intelligence” and “Terminator”, but successful tests indicate that computers have made progress in the art of conversation – a natural thing for humans, but extremely difficult for computers, writes The Telegraph.

Turing himself considered the question – can a machine “humanize” – meaningless.  In his opinion, it is better to focus on the fact that the computer can do: if he could talk?  Can carry on a conversation well enough to pass for a man?  If can, according to the scientist, on some level, it can be called intelligent.

Leave a Reply

Your email address will not be published. Required fields are marked *