Grasp the limitations of artificial intelligence from logic dimension
September 06, 2022 09:37 Source: "China Social Sciences", September 6, 2022, Issue 2486 Author: Sun Wen

Artificial intelligence is the simulation of human intelligent activities by humans using computers,Therefore, some things that computers complete human heart intelligence。Usually,"Smart" is not one -dimensional,but have a clear level、Different dimensions with rich structure,Have various information processing capabilities。Current,Artificial intelligence can use multiple technologies to complete multiple tasks,It has brought great convenience to human life。But with the large -scale and long -term application of artificial intelligence,Because artificial intelligence cannot be self -regulating according to the specific situation,and the lack of the existence of smart machines and the limitations of the intelligent system,Decision error、Algorithm bias、Data leaks and other issues are endless。Therefore,This article explores the limitations of artificial intelligence development from logical dimension,I hope to bring some inspiration to the relevant challenges。

  Bet365 lotto review

The idea of ​​human work and calculation of humans with machines has a long history。As early as the 17th century,Labnitz proposed the idea of ​​creating "general language",That is to create a symbol system that creates the meaning of the meaning and no ambiguity,Express human knowledge and information with these symbols。1936,Alan Turing (Alan Turing) put forward the theoretical model of modern computers in the article "On the Calculating Calculation Bet365 lotto review and Application of the Judgment Issue"。From this,Labnitz's conception can be continued and realized on the Turing machine,Human reasoning is transformed into digital operations,and processed by the computer。

Turing machine is the prototype of modern computers,The structure is very simple,That is a long enough paper belt that can print with symbols and a read -write head that can obtain a symbol。​​Then,Greed up the Turing Machine for Performing Works from a set of instructions,In essence, it can be understood as a limited four -dollar group (current state、paper belt symbol、Operations to be executed、New Status)。From this,We can see,There is a gap between the Turing machine although there is a modern computer,But its basic function is not different from modern computers,Storage and calculation。

Actually,Artificial intelligence is the darling of the information age,Computer processing speed is very fast,and I don’t know tired、Don't worry about annoying,Human expectations will give the computer to computer to complete the heavy task of processing information and knowledge。From the basic function of the computer,Its storage、The computing power is far beyond human individual。From this point of view,Artificial intelligence based on computer,"smart" than humans、More "intelligent"。So far,It also prompted and stimulated us to think about another question: Can artificial intelligence have "real" intelligence?

  Incantance theorem

1931,Incompleteness theorem gives,"Any formal system that contains first -order predicate logic and elementary theory,There is a proposition,It can not be proven as true in this system,It cannot be proved to be pseudo "。1951,Wang Haogen Bet365 app download pointed out according to Gothic's Gibbes speech,Human intelligence does not format all human mathematics intuition,It will also generate new intuition knowledge,This fact can be called "incompleteness"。Therefore,In a sense,,In the proposition logic,Send it from the appropriate premise,Through reasoning can interpret all the truths in this field。but,When the proposition logic is expanded to the predicate logic or an early arithmetic system,There is always the truth in this field that cannot be performed through interpretation,Fullness cannot be established。

Simply put,Artificial intelligence is formalized,It is an executable program composed of data sets and instruction sets。The instruction set of artificial intelligence execution is equivalent to the axiom。When the instruction set is rich enough,Some procedural sentences within this range will always be concentrated from these instructions。Therefore,No matter how powerful artificial intelligence,It is impossible to complete all instructions。There will always be some complicated and important calculations,Form system cannot help us realize it。At this time,Artificial intelligence will not be able to execute these instruction sets。

  Qiu Qi -Turing Theory

1934,Alonzo Church proposed,"The calculated function that can be done is generally recursive"。1936,Turing proposed,"The functions that can be calculated can be a function that can be calculated by Turing machines"。It can be said,Any recursive function is a Turing calculated,collectively referred to as Qiu Qi-Turing-Turing thesis,CTT)。So far,People have determined that the machine can only calculate the general recursion bet365 live casino games function,that is, a function that can be calculated by Turing。The machine cannot complete all function operations。For functions outside the general recursive function,No matter how smart and powerful artificial intelligence is,It is also impossible to calculate。Qiu Qi — Turing The Title clearly explained,Artificial intelligence is not universal,It is necessary to solve the problem with artificial intelligence,The requesting problem is calculated,and also have a reasonable complexity。

other,Qiu Qi -The recursive thinking in Turing Theory is also the basis for the formalization of artificial intelligence。Obviously,The transform sequence connected by the ring is easy to achieve on the machine,The formalization of logic is the foundation of formalization of artificial intelligence。Formalization of artificial intelligence,Generally defined according to the formalization of the logic system。When artificial intelligence handles tasks or problems,The premise is to formally form the task or problem logic。but,The formalization of the logical system is applicable。Current,Artificial intelligence cannot achieve the complete formalization of human intelligence。It is not difficult to see,When facing the task that cannot be characterized by formalization,There will be many limitations and difficulties in artificial intelligence。

  Bols Mind -Machine Theory Questions

1986,A. W. Burks (A. W. Burks) in the book "Robot and Human Mind",Discussing a possibility of all natural functions that can be achieved by a poor motivation,and called it "Mind-Machine Thesis"。From Bolgs,,There is no essential difference between people and robots,The machine will always realize the function of human mind。,This view has been questioned and challenged by many questions。bet365 best casino games From the perspective of philosophy,Can artificial intelligence truly understand the "instructions"? Can you have perception、Consciousness and emotion?

Facing the "understanding" of artificial intelligence,We can regard "understanding" as a ternary relationship,It can be regarded as "computer" "information assignment" for "instructions and statements"。In simple terms,It is to give "semantics" to "word -shaped symbols"。"Symbol" is the form language of the data,The "understanding" of artificial intelligence on these language symbols is achieved by "semantic assignment"。In proposition logic and predicate logic,It is not difficult to realize the semantic assignment of the "word -shaped symbol",But according to the concept of arithmetic concepts of Alfred Tarski, the theorem,When the language is rich to minimal arithmetic,"True" concept is not defined by arithmetic,In a enough system,Unable to define the true of the language。This description,In a simple form language,Artificial intelligence can "understand" semantics "true",But when it expands to a rich enough arithmetic language,Artificial intelligence will not be able to understand the most basic semantics "True",Not to mention the instructions and sentences of its input and output in the real "understanding"。and,Human intelligence includes judgment、Inference、Learning、Memory and perception, etc.。But the computer runs according to the instructions designed by human beings,No perception、Consciousness、Emotional、Intuition and imagination functions。In other words,It is the "injection" computer for humans to "inject" a large amount of knowledge and information,Not a computer active sensation,It does not have autonomy。

Bet365 lotto review So,Artificial intelligence is just a special tool,is a machine that executes instruction set,No autonomy、Consciousness and emotion。The so -called machine has self -consciousness, etc.,Just make the machine "look" self -consciousness,Some selectivity and autonomy of human beings in the execution instructions are not the same thing,It's just a "false autonomy"。It is not difficult to see,Some aspects of artificial intelligence difficult to handle mental mind,Therefore, it is difficult to say that having "real" intelligence。

In terms of understanding and transformation of the world,Human beings as an individual,There is a certain limitations in its cognition,But for the entire human beings,It can break through this limitations。The rapid development of computer science and artificial intelligence has aroused the expansion and change of philosophical paradigms,A series of new paradigms emerged。For example,Artificial Intelligence Philosophy、Logic machine philosophy、Computer Philosophy、Information philosophy, etc.。To some extent,Human beings can have a broader than any computer、GM problem judgment and solution ability,Therefore, intelligent computers based on computing theory cannot achieve a complete replacement of humans。but but,There is still a lot of room for development of artificial intelligence。We should take a different approach、Pioneering Innovation,Can't be limited to the development and application of smart machines,It is also necessary to study deeper specific laws in human thinking activities,So put forward a new research program,Form new research field,New method、Concepts and mechanisms are fused with new theories or machine models,Plug -in one by bet365 live casino games one、Create a new type of smart machine。

  (This article is the phased achievement of the "Logical Frontier Issue of Artificial Intelligence" in the Youth Marxist theorist project of the Marxist College of Hebei Normal University)

(Author Unit: Marxist College of Hebei Normal University)

Editor in charge: Zhang Jing
QR code icon 2.jpg
Key recommendation
The latest article
Graphics
bet365 live casino games

Friendship link:

Website filing number: Jinggong.com Anmi 11010502030146 Ministry of Industry and Information Technology:

All rights reserved by China Social Sciences Magazine shall not be reprinted and used without permission

General Editor Email: zzszbj@126.com This website contact information: 010-85886809 Address: Building 11-12, Building 1, Building 1, No. 15, Guanghua Road, Chaoyang District, Beijing: 100026