Can machines think

 In August this year, Lei feng's network (search for "Lei feng's network", public interest) will be held in Shenzhen, a global influence of unprecedented artificial intelligence and robot Innovation Conference. When Lei feng's network will publish "artificial intelligence and robot Top25 list of innovative enterprises". At present, we are visiting related companies in the area of artificial intelligence, robots, sift through them eventually elected to the list list of companies. If you want to join our list, please contact: 2020@leiphone.com. Cath Kidston iPhone 6 cases

Lei feng's network: the author should be benevolent people from bupt and cognition Lab (male: man machine and cognition lab).

Machine intelligence? Machine will understand? Can machines think? From the 17th century to today, every major breakthrough in machine functions, this question was asked again, but denied it had become increasingly difficult.

Alpha Go tournament with Li Shishi dust settles, chess fans and after the technical discussion, philosophical reflection. Overseas academic circles this calm, perhaps numb to it. Speech conversation technology, image recognition and intelligent service has now entered the daily life of people, thinking as far back as when you start the application of artificial neural network to re-examine the philosophy of AI cognitive. Some people in China also believes that the machines are just puppets, says this game is just a programmer and player, AI is better than a human is a joke. Its philosophical basis, mainly from the John r. Searle's "Chinese room" thought experiment.

Chinese room (Chinese room) is a United States Berkeley philosophy professor John r. Searle (John Searle) was designed in 1980 by a thought experiment (The Chinese Room argument), and his 1999 revision is as follows:

Imagine the following scene. There was a man only know English don't understand Chinese, and text in a box (for databases) and on how to handle Chinese characters using the word English manual (procedures) are locked in a room. Without the knowledge of the people outside the room, tucked a note questioning in Chinese (input). Persons selected according to manual instructions for it in Word, send the correct answer (output). This manual allows the bird-people, adopted the Turing test, understand Chinese, but he in fact knew nothing about Chinese.

Searle thinks the House computers and artificial intelligence, don't have to understand the meaning of the input and output character, albeit only in accordance with the rules mechanically characters, give people outside the impression one can understand Chinese, the Chinese room and computer did not thinking and understanding.

This is one of the most famous in modern philosophy thinking, he is able to pass the Turing test (Turning Test) of machine intelligence, caused wide discussion in philosophy and cognitive science. Now with a Google Scholar search for "Searle's Chinese Room", there are tens of thousands of results.

Chinese also have some discussion of this thought experiment, some people eager to design a machine cannot properly answer the question, trying to prove the machine intelligence will not overcome the human. In fact, this is a misreading of it. Searle's experiment is assuming that the system is able to pass the Turing test, in this discussion based on a philosophy of mind issue. Now machine intelligence technology development, confirms that the premise of this philosophy is nonsense.

Let us investigate in more detail the origin of the famous speculative connotation.

Can machines think?

Historically, Leibniz (1646-1716) with the mill's thought experiment (Leibniz'Mill) study on this issue.

If to thinking and perception of digital and mechanical movement principles to explain, this intelligent machine magnified many times, people just like to check into the mill, we can only find out parts between the drive and there are no parts to be able to explain the perception, such as pain, their combination of machinery must be so. Leibniz to oppose prevailing mechanical determinism, advocates dualism of matter and mind.

From the beginning of the information revolution, Turing thinking activities information flow according to the program. He envisaged a paper machine that can play chess (Turing Paper Machine, 1948).

If you have a piece of paper, write chess responses to all situations, people don't need to know the game of chess, as long as control instructions on the paper can play chess. The contents of the paper are programs, computers follow programs running like humans think chess.

In 1950, Turing published a landmark paper to the computer can think? , First put forward the "thinking machine" concept. He said, if dialogue through telex terminals with the other side, people can't tell there is a machine or a human? Then admit the machine intelligence. This test is called the "Turing test" (Turning Test). His mind in cognitive science known as functionalism.

The 70 's of the last century, computers became commonplace, Roger Schank at Yale University with "concept" technology to break down the statement, AI is considered background knowledge database can understand English. Chinese room Searle speculation against the intelligent interpretation of functionalism.

Searle constructed "Chinese room", imitating the Turing test, but inherited Leibniz's speculative study each component of this intelligence where one has understanding of himself, all the others are not the tools of life and mechanical operation, but it can show understanding of Chinese, and he does not know the Chinese language. So this is just a play.

Chinese room thought experiment think of macro-micro-mechanism and driving force, different micro-mechanism and driver might formed the same macroscopic phenomena, Turing, as long as the macro function is identical to human intelligence, we can use the word to describe this smart feature. As in a tournament in the game to talk about the skill level of the macroscopic tests, said the machine has a smart and necessary language dialogues as well. Sayre officials stressed that their internal mechanisms and indicate whether it has true intention, merely that the Chinese room dialogue function of simulation, does not have the capacity to understand this activity.

Searle artificial intelligence fall into two categories: that must be able to understand natural language, and mimic human behavior, machines with similar mental activity, known as the "strong artificial intelligence (Strong AI)" intelligent machines that only imitates functions not thought of as "weak artificial intelligence (Weak AI)". His model of strong artificial intelligence is enough to explain the mechanisms of mental activity. He then believes that research brain and mind-independent, the Turing test is not enough to establish psychological state without understanding or intention (intentionality), you can't say the machine will think, so there is no common sense thinking ability.

Searle reasoning directly to the conclusion that "realization of the Chinese room is not a strong artificial intelligence", General reasoning is "impossible through character actions, from the language of the form (Syntax), the semantic content (semantic)".

Philosophy is a concept, and its judgment in the name of truth or impractical from the authority. Philosophy is not like mathematics, the latter expression is the logical result of the agreement, by careful reasoning, can distinguish right from wrong, nor does it express like a scientific fact, his claim can be proved or denied. Philosophical significance is thinking of through polemical analysis to clarify ambiguities, inspired on the perception of the world.

Since 1980, philosophers, scientists began to publish in academic journals reviews, participate in discussions, Chinese room has been widely referenced. Most scholars do not agree with Searle. BBS editor Stevan Hamad said: "the vast majority of people think that the Chinese room thought was totally wrong. "No Chinese room thought, does not agree with the broadly divided into three categories: System response, robotic responses and negative responses. Searle decades compiling different opinions, sharp retort, promoted the soul-searching. Cath Kidston iPhone

In an earlier reply, is the largest "systems reply" point of view. They admit Chinese rooms of the operators do not know Chinese, but said that does not mean that one system does not have a understanding of the Chinese. Characteristics of the system are not able to refine the distinction to a part, it is produced by the overall synergy. Searle argued, the machine can speak Chinese, does not mean that it understands Chinese. I put everything in the room include programs, databases, manuals into mind, I was the system. I like the computer according to the procedure in my answer, act like Chinese, but I know that I still do not understand Chinese.

"Robot replies" agreed that Chinese room scenario, unable to speak or understand Chinese, like Westerners to never eat bread, database how to explain Chinese steamed buns, are not themselves seen, smelled, ate, did so with understanding, steamed bread can truly understand the semantics of the term. So they suggested that machine brains, wrapped in a have hearing, vision, smell, taste and touch sensor robot, let it walk in our world, then the robot feels, and will have a understanding. Searle thinks that attitude is no better than the "systems reply" clever, because all these sensors senses is just extra inputs, just as in the Chinese room, tucked into some other note on world knowledge, or the knowledge has been compiled into a database, in any case, me in the room, is still not understand Chinese.

"Denied replies" is completely not identity Chinese room of speculative enough to denied strong artificial intelligence, said Searle put room in the of all are loaded into mind, reference Chinese database in of knowledge, can fluent with Chinese answered problem, this case on like physical, and chemical most experiment we are didn't do had, most of vocabulary of meaning are is from books in the others of experience, but we are think himself can understanding these vocabulary and related statement, why put in also practices of AI Shang on no does? There are those who even believes that Searle's thinking from the wrong intuition, is like stepping into Leibniz's Mill, the intuitive does not reflect the macro-phenomena in nature. For example rotating magnets studied people can't see the light, but we can not deny that the light was electromagnetic waves. This thought experiment is like to slow the clock into the atomic scale, let us see the light. To micro-scale and slow process to understand "intelligent" and "understand" the macroscopic phenomena, human intuition is often wrong in the House. Searle to such criticism with "prisoner of ideology" in a nutshell. He said that does not clearly spell out what kind of system are, by what mechanism can export the mind, became convinced that the mind can from the system "emerges", is an entirely intuitive blindly confident about yourself.

Chinese room really is about the speculation on the philosophy of mind. It is Turing was the most extensive discussion of the issues, it is clear and pure cognitive science to philosophy on the classic.

Although the "Chinese room" thought is for artificial intelligence, its philosophical thinking very deep, most computer scientist and artificial intelligence expert, but I don't think that their work-related, not find any limitations in AI. They worked hard, developed with the input of human language and AI image recognition, which combined with humans in chess than winning machine. Artificial intelligence, whether classified as weak, like the zombies in the philosophy without a soul, or better than humans have their own consciousness, which in many intellectual activities will be stronger than that of humans is no longer speculation.

Guidelines for these smart ideas in the study of, what is it?

Lei Feng network Note: reprints please contact the authorized and retain complete information may not delete, modify articles.

"Promotional" bird cloud cloud computing market changes across all product lines 0 Yuan/half a year free

2851 votes

Lumia 940XL

Lumia 940XL as Microsoft Windows 10 ultimate models will be equipped with mycophenolate mofetil 810 processor, with 3GB and 32GB storage space, 20 million + 5 million before and after camera configuration and 2K screen means 940XL will become the next generation of hardware monster, while supporting the 3D interactive is its unique highlights.

View details of the voting >>