ALAN TURING LÀ AI

Computer pioneer & artificial intelligence (AI) theorist Alan Turing would have sầu been 100 years old this Saturday. To mark the anniversary the bloginar.net has commissioned a series of essays. In this, the fourth article, his influence on AI retìm kiếm and the resulting controversy are explored.

Bạn đang xem: Alan turing là ai


*

image captionDespite advances in computer giải pháp công nghệ, scientists have sầu not been able khổng lồ create a "thinking machine" that can pass the Turing Test

Alan Turing was clearly a man ahead of his time. In 1950, at the dawn of computing, he was already grappling with the question: "Can machines think?"


The term artificial intelligence had not even been coined. John McCarthy would come up with the term in 1956, two years after Alan Turing's untimely death.


Yet his ideas proved both to have sầu a profound influence over the new field of AI, và to lớn cause a schism amongst its practitioners.


One of Turing's lasting legacies lớn AI, & not necessarily a good one, is his approach to lớn the problem of thinking machines.


Instead, he turned the tables on those who might be sceptical about the idea of machines thinking, unleashing his formidable intellect on a range of possible objections, from religion khổng lồ consciousness.


*

image captionTuring outlined his AI experiment while working at the University of Manchester, where a memorial statue has since been erected

With so little known about where computing was heading at this time, the approach made sense. He asserted correctly that "conjectures are of great importance since they suggest useful lines of research".


But 62 years on, now that we have sầu advanced computers khổng lồ thử nghiệm, it seems wrong that some proponents of AI still demand the onus be put on sceptics lớn prove sầu the idea of an intelligent machine impossible.


The philosopher Bertrand Russell ridiculed this type of situation, likening it lớn asking a sceptic to lớn disprove there is a china teapot revolving around the sun while insisting the teapot is too small to be revealed.


This can be seen as wrong-footing the scientific process of hypothesis testing & evidence collection.


In fact, Turing well understood the need for empirical evidence, proposing what has become known as the Turing Test to lớn determine if a machine was capable of thinking. The kiểm tra was an adaptation of a Victorian-style competition called the imitation game.


It involves secluding a man and woman from an interrogator who has to guess which is which by asking questions và studying written replies.


In the Turing Test, a computer program replaces the man. Turing asked: "Will the interrogator decide wrongly as often when the game is played lượt thích this as he does when the game is played between a man and a woman."


Effectively, the kiểm tra studies whether the interrogator can determine which is computer và which is human (although Turing did not explicitly say that the interrogator should be told that one of the respondents was a computer it seems clear to me from his example questions that this was what he intended).


The idea was that if the questioner could not tell the difference between human và machine, the computer would be considered khổng lồ be thinking.


*

image captionTuring was inspired by a parlour game in which an interrogator puts questions to a man và woman (A and B) in a separate room who reply with typewritten notes. The ayên is to lớn determine which is the man & which is the woman.
*

image captionTuring's demo replaces the man with a computer running a program designed to lớn deceive sầu the questioner about its true identity. Will he still be able to determine which is the woman?
*

image captionThe idea was that if the person asking the questions could not tell the difference between human & machine, the computer would be considered lớn be thinking and have sầu artificial intelligence.

Xem thêm: " Contra Account Là Gì, Nghĩa Của Từ Contra Account, Nghĩa Của Từ Contra Account Trong Tiếng Việt


Turing suggested that by the year 2000 the average interrogator would have sầu less than a 70% chance of making the right decision after five sầu minutes of questioning.


My iPhone has more than 500 times the storage capađô thị he thought would be required & orders of magnitude more processing power, yet passing the thử nghiệm still seems a long way off.


In 1990, Thành Phố New York businessman Hugh Loebner phối up the annual Loebner Prize competition with a prize of $100,000 (£63,500) khổng lồ the creator of a machine that could pass the Turing Test.


image captionLoebner Prize judges have sầu five sầu minutes to ask questions khổng lồ determine which respondent is a computer & which a person

The AI aristocracy strongly supported the conthử nghiệm until it became clear how badly the machines were performing.


Marvin Minsky, one of the fathers of AI, wrote in 1995: "I vì hope that someone will volunteer to lớn violate this proscription so that Mr Loebner will indeed revoke his stupid prize, save sầu himself some money, & spare us the horror of this obnoxious & unproductive sầu annual publicity campaign."


No-one in AI seems to take the failure of the Turing Test as an argument against the possibility of thinking machines.


Because Turing had talked about a future "imaginable machine", some of the proponents say that we will have sầu them at a future date. But others now argue that the Turing Test simply is not the best way lớn measure machine intelligence.


Despite the failure of machines khổng lồ deceive us inlớn believing they are human, Turing would be excited by the remarkable progress of AI.


It is flourishing in so many spheres of activity, from robots investigating the progress of climate change khổng lồ computers running the world's finances.


I expect that Turing would have sầu danced for joy in 1997 when Deep Blue defeated world champion Gary Kasparov at chess.


image captionIBM's Watson beat two Jeopardy champions in a special edition of trò chơi in 2011

I can also imagine hlặng cheering in the wings of the TV trò chơi Jeopardy when the program Watson beat the two best human opponents in the history of the American game.


It is difficult to lớn tell how any of these achievements would have been possible without the continued inspiration from Turing's original và radical ideas.


In my opinion, the Turing Test remains a useful way khổng lồ chart the progress of AI và I believe sầu that humans will be discussing it for centuries lớn come.


Noel Sharkey is professor of artificial intelligence and robotics at the University of Sheffield và co-founder of the International Committee for Robot Arms Control.