What are Some Recent Advances in Artificial Intelligence?

Artificial Intelligence (Artificial Intelligence), the English abbreviation is AI. It is a new technological science that researches and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence.

Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this area includes robotics, language recognition, image recognition, Natural language processing and expert systems. Since the birth of artificial intelligence, the theory and technology have become more and more mature, and the application field has also expanded. It can be imagined that the technology products brought by artificial intelligence in the future will be the "containers" of human intelligence. Artificial intelligence can simulate the information process of human consciousness and thinking. Artificial intelligence is not human intelligence, but it can think like humans and may exceed human intelligence.

Artificial intelligence is a very challenging science, and those who do it must know computer knowledge, psychology, and philosophy. Artificial intelligence is a very wide range of sciences. It is composed of different fields, such as machine learning, computer vision, etc. In general, a major goal of artificial intelligence research is to make machines capable of performing tasks that normally require human intelligence. Complex work. But different ages and different people have different understandings of this "complex work." ^[1] In December 2017, artificial intelligence was selected as the "Top Ten Buzzwords of Chinese Media in 2017". ^[2]

Chinese name: artificial intelligence
Foreign name: ARTIFICIAL INTELLIGENCE
Short name: AI

Presentation time: 1956
Place of presentation: DARTMOUTH Society
Name origin: Hugo de Garris

Detailed definition of artificial intelligence

Artificial intelligence robot

The definition of artificial intelligence can be divided into two parts, namely "artificial" and "intelligent". "Manual" is easier to understand and less controversial. Sometimes we have to consider what can be made by manpower, or whether the level of intelligence of the person is high enough to create artificial intelligence, and so on. But in general, "artificial systems" are artificial systems in the usual sense.

There are many more questions about what is "smart". This involves other issues such as consciousness (CONSCIOUSNESS), self (SELF), thinking (MIND) (including unconscious thinking (UNCONSCIOUS_MIND)) and so on. The only intelligence that people understand is their own intelligence, which is a generally accepted view. But our understanding of our own intelligence is very limited, and we also have a limited understanding of the necessary elements that constitute human intelligence, so it is difficult to define what is called "intelligence" made by "artificial". Therefore, the research of artificial intelligence often involves the study of human intelligence itself. Other intelligence on animals or other man-made systems is also generally considered to be a research topic related to artificial intelligence.

Artificial intelligence has received more and more attention in the computer field. And it is used in robots, economic and political decision-making, control systems, and simulation systems.

Professor Nelson defines artificial intelligence as such: "Artificial intelligence is the discipline of knowledge-the science of how to represent knowledge and how to acquire and use knowledge." And another professor at Massachusetts Institute of Technology, Winston I think: "Artificial intelligence is the study of how to make computers do intelligent tasks that only humans could do in the past." These statements reflect the basic ideas and basic content of the subject of artificial intelligence. That is, artificial intelligence is the study of the laws of human intelligent activity, constructing artificial systems with a certain intelligence, studying how to make computers perform tasks that previously required human intelligence to be competent, that is, studying how to use computer software and hardware to simulate some human intelligence Basic theories, methods, and techniques of behavior.

Artificial intelligence is a branch of computer science. Since the 1970s, it has been called one of the world's three cutting-edge technologies (space technology, energy technology, artificial intelligence). It is also considered to be one of the three cutting-edge technologies (genetic engineering, nanoscience, artificial intelligence) in the 21st century. This is because it has developed rapidly in the past thirty years, has been widely used in many disciplines, and has achieved fruitful results.Artificial intelligence has gradually become an independent branch, both in theory and practice. Into a system.

Artificial intelligence is a discipline that studies the use of computers to simulate certain thinking processes and intelligent behaviors of people (such as learning, reasoning, thinking, planning, etc.). It mainly includes the principles of computer-implemented intelligence, manufactures computers similar to human brain intelligence, and makes computers Can achieve higher-level applications. Artificial intelligence will involve disciplines such as computer science, psychology, philosophy and linguistics. It can be said that it is almost all disciplines of natural sciences and social sciences, and its scope is far beyond the scope of computer science. The relationship between artificial intelligence and thinking science is the relationship between practice and theory. Artificial intelligence is at the level of technological application of thinking science. Is an application branch of it. From a thinking point of view, artificial intelligence is not limited to logical thinking. It is necessary to consider image thinking and inspirational thinking to promote the breakthrough development of artificial intelligence. Mathematics is often considered as a basic science in many disciplines. Mathematics has also entered the field of language and thinking. Intelligent disciplines must also borrow mathematical tools. Mathematics not only plays a role in standard logic, fuzzy mathematics, etc. Mathematics enters the discipline of artificial intelligence, they will promote each other and develop faster. ^[3]

Artificial intelligence research value

Robot with artificial intelligence For example, heavy scientific and engineering calculations were originally undertaken by the human brain. Today, not only can computers perform such calculations, but they can also do them faster and more accurately than the human brain. Therefore, contemporary people no longer regard such calculations as It is a "complex task that requires human intelligence to complete." It can be seen that the definition of complex work changes with the development of the times and technological progress, and the specific goal of the science of artificial intelligence naturally develops with the changes of the times. On the one hand, it keeps making new progress, and on the other, it turns to more meaningful and difficult goals.

Generally, the mathematical foundations of "machine learning" are "statistics", "information theory", and "control theory". Also includes other non-mathematics subjects. This type of "machine learning" is very dependent on "experience." Computers need to continuously acquire knowledge from the experience of solving a class of problems and learn strategies. When encountering similar problems, they use empirical knowledge to solve problems and accumulate new experiences, just like ordinary people. We can call this type of learning "continuous learning." But in addition to learning from experience, human beings also create "leap learning." This is called "inspiration" or "epiphany" in some cases. The hardest thing for a computer to learn is "epiphany." Or more strictly speaking, it is difficult for a computer to learn "qualitative change that does not depend on quantitative change" in terms of learning and "practice", it is difficult to go directly from one "qualitative" to another "qualitative", or from a "concept" To another "concept." Because of this, the "practice" here is not the same as human practice. Human practice includes both experience and creativity.

This is what intelligent researchers dream of.

In 2013, SC Wang, a data researcher at Dijin Data Center, developed a new data analysis method, which led to a new method for studying the properties of functions. The authors found that the new data analysis method provided a way for the computer society to "create". In essence, this approach provides a fairly effective way to model human "creativity". This approach is given by mathematics and is a "capability" that ordinary people can't have, but computers can have. From then on, computers are not only good at computing, but also good at creating because of good computing. Computer scientists should decisively deprive "creative" computers of being too comprehensive to operate, or computers will one day "counter-trap" humans.

When looking back at the deduction process and mathematics of the new method, the author expanded his understanding of thinking and mathematics. The mathematics is concise, clear, reliable, and strong in modeling. In the history of mathematics, the brilliance of the creativity of mathematics masters shines everywhere. These creativity are presented in the form of various mathematical theorems or conclusions, and the biggest feature of mathematical theorems is that they are based on some basic concepts and axioms and are expressed in a patterned language with a logical structure containing rich information. It should be said that mathematics is the discipline which reflects (at least one type) the model of creativity in the simplest and most straightforward manner.

Artificial intelligence development stage

In the summer of 1956, a group of visionary young scientists led by McCarthy, Minsky, Rochester, and Shennong met together to jointly study and discuss a series of related issues using machine simulation intelligence, and first proposed The term "artificial intelligence" marks the official birth of the emerging subject "artificial intelligence". IBM's "dark blue" computer defeated human world chess champion is a perfect expression of artificial intelligence technology.

Since the formal introduction of the subject of artificial intelligence in 1956, it has made great progress in more than 50 years and has become an extensive cross-cutting and cutting-edge science. In general, the purpose of artificial intelligence is to make a computer, a machine, think like a human. If you want to make a machine that can think, then you must know what is thinking, and furthermore, what is wisdom. What kind of machine is intelligent? Scientists have made cars, trains, airplanes, radios, etc. They mimic the functions of our body organs, but can they mimic the functions of the human brain? So far, we only know that the thing contained in our Tianling Cap is an organ composed of billions of nerve cells. We know very little about this thing, and imitating it is probably the most difficult thing in the world.

After the advent of computers, humans began to really have a tool that could simulate human thinking. In the years that followed, countless scientists worked hard for this goal. Today, artificial intelligence is no longer the patent of several scientists. Computer departments in almost all universities around the world are studying this subject. University students studying computer must also take such a course. With the unremitting efforts of everyone, today Computers seem to have become very smart. For example, in May 1997, the Deep Blue computer developed by IBM defeated chess master Kasparov. You may not notice that in some places computers help people to perform other tasks that belonged to humans. Computers play their role for humans with its high speed and accuracy. Artificial intelligence has always been at the forefront of computer science. Computer programming languages and other computer software have existed because of advances in artificial intelligence.

On March 4, 2019, the second session of the Thirteenth National People's Congress held a press conference. Spokesperson Zhang Yesui said that legislative projects closely related to artificial intelligence have been included in the legislative plan ^[4] .

Introduction to artificial intelligence science

Practical application

Machine vision, fingerprint recognition, face recognition, retinal recognition, iris recognition, palm print recognition, expert system, automatic planning, intelligent search, theorem proof, game, automatic programming, intelligent control, robotics, language and image understanding, genetics Programming, etc.

Subject category

Artificial intelligence is a fringe discipline, which belongs to the intersection of natural science and social science.

Involved Disciplines

Philosophy and Cognitive Science, Mathematics, Neurophysiology, Psychology, Computer Science, Information Theory, Cybernetics, Uncertainty

Research category

Natural language processing, knowledge representation, intelligent search, reasoning, planning, machine learning, knowledge acquisition, combined scheduling problems, perceptual problems, pattern recognition, logical programming soft computing, imprecise and uncertain management, artificial life, neural networks, Complex systems, genetic algorithms

Awareness and artificial intelligence

Artificial intelligence, by its very nature, is a simulation of the information process of human thinking.

The simulation of human thinking can be carried out from two paths. One is structural simulation, which imitates the structural mechanism of the human brain to create a "human brain-like" machine. The second is functional simulation, which temporarily ignores the internal structure of the human brain, Functional processes are simulated. The emergence of modern electronic computers is the simulation of the thinking function of the human brain, and the simulation of the information process of the human brain.

Weak artificial intelligence is now developing rapidly, especially after the 2008 economic crisis, the United States, Japan and Europe hope to achieve re-industrialization with robots, etc. Industrial robots are developing at a faster rate than ever before, which has further promoted weak artificial intelligence and related industries. With the continuous breakthrough, many tasks that must be done by humans can now be achieved by robots.

Strong artificial intelligence is temporarily at the bottleneck and requires the efforts of scientists and humans.

Research on artificial intelligence technology

The machine used to study the main material basis of artificial intelligence and the platform that can realize the artificial intelligence technology is the computer. The development history of artificial intelligence is connected with the development history of computer science and technology. In addition to computer science, artificial intelligence also involves multiple disciplines such as information theory, cybernetics, automation, bionics, biology, psychology, mathematical logic, linguistics, medicine, and philosophy. The main contents of artificial intelligence subject research include: knowledge representation, automatic reasoning and search methods, machine learning and knowledge acquisition, knowledge processing systems, natural language understanding, computer vision, intelligent robots, and automatic program design.

Artificial intelligence research methods

There are no unified principles or paradigms today that guide artificial intelligence research. Researchers have debates on many issues. Several of the questions that have not been concluded for a long time are: Should artificial intelligence be simulated psychologically or neurally? Or is human biology irrelevant to artificial intelligence research like bird biology to aviation engineering? Can intelligent behavior be described by simple principles such as logic or optimization? Or do you have to solve a lot of completely unrelated issues?

Can intelligence use advanced symbolic expressions such as words and ideas? Or do we need "sub-symbols"? JOHN HAUGELAND proposed the concept of GOFAI (Excellent Old-Fashioned Artificial Intelligence), and also proposed that artificial intelligence should be classified as SYNTHETIC INTELLIGENCE, [29] this concept was later adopted by some non-GOFAI researchers.

Brain simulation

Main article: Cybernetics and computational neuroscience

In the 1940s and 1950s, many researchers explored the links between neurology, information theory, and cybernetics. It also created some preliminary intelligence using electronic network construction, such as TURTLES and JOHNS HOPKINS BEAST by W. GREY WALTER. These researchers also frequently hold technical association meetings at Princeton University and RATIO CLUB in the UK. Until 1960, most people had abandoned this method, although the principles were again proposed in the 1980s.

Symbol processing

Main article: GOFAI

When the digital computer was successfully developed in the 1950s, researchers began to explore whether human intelligence could be simplified into symbol processing. Research has focused on Carnegie Mellon University, Stanford University and MIT, each with its own independent research style. JOHN HAUGELAND calls these methods GOFAI (Excellent Old-Fashioned Artificial Intelligence). [33] In the 1960s, the symbolic method had great achievements in simulating advanced thinking on small proof programs. Methods based on cybernetics or neural networks are secondary. [34] Researchers in the 1960s and 1970s were convinced that symbolic methods could eventually succeed in creating strong artificial intelligence machines, and that was their goal.

Cognitive simulation economists Herbert Simon and Alan Newell study human problem-solving capabilities and try to formalize them, while laying the foundation for the basic principles of artificial intelligence, such as cognitive science, operations research, and operations science. Their research team used the results of psychological experiments to develop programs that mimic human problem solving. This method has been inherited at Carnegie Mellon University, and reached its peak in SOAR in the 1980s. Based on logic unlike Allen Newell and Herbert Simon, JOHN MCCARTHY believes that machines do not need to simulate human thought, but should try to find the essence of abstract reasoning and problem solving, whether or not people use the same algorithm. His laboratory at Stanford University is dedicated to using formal logic to solve a variety of problems, including knowledge representation, intelligent planning, and machine learning. The University of Edinburgh is also working on logical methods, which has led to the development of the programming language PROLOG and logic elsewhere in Europe The science of programming. "Anti-Logic" Stanford researchers (such as Marvin Minsky and Simole Piper) have found that special solutions are needed to solve the difficult problems of computer vision and natural language processing-they argue that simple Common principles (such as logic) enable all intelligent behaviors. ROGER SCHANK describes their "anti-logic" method as "SCRUFFY". Knowledge bases (such as CYC of DOUG LENAT) are examples of "SCRUFFY" AI, because they have to manually write a complex concept at a time. Based on the emergence of large-capacity memory computers around 1970, researchers began to construct knowledge into application software in three ways. This "knowledge revolution" led to the development and planning of expert systems, which was the first successful form of artificial intelligence software. The "knowledge revolution" also made people realize that many simple artificial intelligence software may require a lot of knowledge.

Subsymbol method

In the 1980s, symbolic artificial intelligence stagnated, and many people thought that symbolic systems could never mimic all human cognitive processes, especially perception, robotics, machine learning, and pattern recognition. Many researchers have begun to focus on the subsymbol approach to solve specific artificial intelligence problems.

Bottom-up, interface AGENT, researchers in the field of embedded environment (robot), behavioralism, and new AI robots, such as RODNEY BROOKS, deny symbolic artificial intelligence and focus on basic engineering issues such as robot movement and survival. Their work once again focused on the opinions of early cybernetic researchers, while proposing the use of control theory in artificial intelligence. This is consistent with the representational perception thesis in the field of cognitive science: higher intelligence requires individual representations (such as movement, perception, and image). In the 1980s, DAVID RUMELHART and others proposed neural networks and connectionism again. This and other sub-symbol methods, such as fuzzy control and evolutionary computing, belong to the field of computational intelligence.

Statistical method

In the 1990s, artificial intelligence research developed sophisticated mathematical tools to solve specific branch problems. These tools are true scientific methods, meaning that the results of these methods are measurable and verifiable, and they are also the reason for the success of artificial intelligence. A shared mathematical language also allows collaboration in existing disciplines (such as mathematics, economics, or operations research). STUART J. RUSSELL and PETER NORVIG point out that these advances are no less than "revolutions" and "success of NEATS". Some people have criticized these technologies for being too focused on specific issues without considering long-term strong artificial intelligence goals.

Integrated approach

Intelligent AGENT Paradigm Intelligent AGENT is a system that senses the environment and acts to achieve its goals. The simplest smart agents are programs that solve specific problems. More complex AGENTs include humans and human organizations (such as companies). These paradigms allow researchers to study individual problems and find useful and verifiable solutions without considering a single method. An AGENT that solves a particular problem can use any feasible method-some AGENT uses symbolic and logical methods, some are subsymbol neural networks or other new methods. The paradigm also provides researchers with a common language for communicating with other fields-such as decision theory and economics (also using the concept of ABSTRACT AGENTS). The intelligent AGENT paradigm was widely accepted in the 1990s. Researchers of AGENT architecture and cognitive architecture have designed some systems to deal with the interaction between intelligent AGENTs in a multi-ANGENT system. A system that contains symbols and sub-symbols in a system is called a hybrid intelligent system, and the research on such systems is the integration of artificial intelligence systems. The hierarchical control system provides a bridge between the reaction level of the sub-symbol AI and the highest-level traditional symbol AI, while relaxing the planning and world modeling time. RODNEY BROOKS 'SUBSUMPTION ARCHITECTURE is an early grading system plan.

Artificial intelligence intelligent simulation

Simulation of machine vision, hearing, touch, feeling and thinking mode: fingerprint recognition, face recognition, retinal recognition, iris recognition, palm print recognition, expert system, intelligent search, theorem proof, logical reasoning, game, information sensing and dialectical processing .

Category of Artificial Intelligence

Artificial intelligence is a frontier discipline that belongs to a three-way cross discipline of natural science, social science, and technical science.

Subjects related to artificial intelligence

Philosophy and Cognitive Science, Mathematics, Neurophysiology, Psychology, Computer Science, Information Theory, Cybernetics, Uncertainty, Bionics, Social Structure and Scientific Outlook on Development.

Artificial intelligence research category

Language learning and processing, knowledge expression, intelligent search, reasoning, planning, machine learning, knowledge acquisition, combined scheduling problems, perception problems, pattern recognition, logical programming, soft computing, imprecise and uncertain management, artificial life, The most critical problem of neural networks, complex systems, and genetic algorithms for human thinking is the shaping and improvement of the machine's independent creative thinking ability.

Artificial intelligence security issues

Artificial intelligence is still being researched, but some scholars believe that it is dangerous for computers to have IQ, and it may resist humans. This kind of hidden danger has also happened in many movies. The main key is to allow the machine to have the consciousness of self-generation and continuation. If the machine has autonomy, it means that the machine has the same or similar creativity and self Protect consciousness, emotions, and spontaneity.

Artificial intelligence implementation method

There are two different ways in which artificial intelligence can be implemented on computers. One is to use traditional programming techniques to make the system appear intelligent, regardless of whether the methods used are the same as those used by human or animal organisms. This method is called ENGINEERING APPROACH, and it has made achievements in some fields, such as text recognition, computer chess, etc. The other is MODELING APPROACH, which not only depends on the effect, but also requires that the implementation method is the same as or similar to that used by humans or biological organisms. Genetic algorithm (GENERIC ALGORITHM, referred to as GA) and artificial neural network (ARTIFICIAL NEURAL NETWORK, referred to as ANN) are both of the latter type. Genetic algorithms mimic the genetic-evolutionary mechanisms of humans or organisms, and artificial neural networks mimic the way neurons in human or animal brains move. In order to get the same smart effect, both methods can usually be used. In the former method, the program logic needs to be specified manually. If the game is simple, it is convenient. If the game is complicated, the number of characters and the space for activities increase, the corresponding logic will be complicated (exponential growth), manual programming is very cumbersome and error-prone. Once an error occurs, it is necessary to modify the original program, recompile, debug, and finally provide a new version or a new patch for the user, which is very troublesome. In the latter method, the programmer must design an intelligent system (a module) for each character to control. This intelligent system (module) does not understand anything at first, just like a newborn baby, but it can learn, can Gradually adapt to the environment and deal with complex situations. This kind of system often makes mistakes at the beginning, but it can learn a lesson, and it may be corrected the next time it is run, at least not forever, and it wo nt be necessary to release a new version or patch. Using this method to implement artificial intelligence requires programmers to have a biological way of thinking, and getting started is a bit more difficult. But once you enter the door, it can be widely used. Because this method does not need to specify the rules of the character's activities when programming, it is usually more labor-saving than the former method when applied to complex problems.

Artificial Intelligence Professional Agency

Artificial Intelligence America

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

STAN STANFORD UNIVERSITY (CA)

NE CARNEGIE MELLON UNIVERSITY (PA)

UNIVERSITY OF CALIFORNIA-BERKELEY

UNIVERSITY OF WASHINGTON

UNIVERSITY OF TEXAS-AUSTIN

UNIVERSITY OF PENNSYLVANIA

UNIVERSITY OF ILLINOIS-URBANA-CHAMPAIGN University of Illinois Urbana-Champaign

OF UNIVERSITY OF MARYLAND-COLLEGE PARK

RN CORNELL UNIVERSITY (NY)

UNIVERSITY OF MASSACHUSETTS-AMHERST

OR GEORGIA INSTITUTE OF TECHNOLOGY

UNIVERSITY OF MICHIGAN-ANN ARBOR University of MICHIGAN-ANN ARBOR

OF UNIVERSITY OF SOUTHERN CALIFORNIA

COLUMBIA UNIVERSITY (NY)

UNIVERSITY OF CALIFORNIA-LOS ANGELES

ROW BROWN UNIVERSITY (RI)

ALE YALE UNIVERSITY (CT)

UNIVERSITY OF CALIFORNIA-SAN DIEGO

UNIVERSITY OF WISCONSIN-MADISON

Artificial Intelligence China

1. Institute of Automation, Chinese Academy of Sciences

2. Tsinghua University

3. Peking University

4.Nanjing University of Science and Technology

5.University of Science and Technology Beijing

6. University of Science and Technology of China

7. Jilin University

8.Harbin Institute of Technology

9. Beijing University of Posts and Telecommunications

10.Beijing Institute of Technology

11. Institute of Artificial Intelligence, Xiamen University

12.Xi'an Jiaotong University Institute of Intelligent Vehicles

13. Institute of Intelligent System and Intelligent Software, Central South University

14. Institute of Intelligence, Xidian University

15. Institute of Image and Artificial Intelligence, Huazhong University of Science and Technology

16.Chongqing University of Posts and Telecommunications

17.Wuhan Engineering University

Main achievements of artificial intelligence

Artificial intelligence man-machine game

On February 10-17, 1996, GARRY KASPAROV defeated DEEP BLUE 4: 2.

From May 3 to 11, 1997, GARRY KASPAROV lost 2.5: 3.5 to the improved "dark blue".

In February 2003, GARRY KASPAROV 3: 3 draw "DEEP JUNIOR".

In November 2003, GARRY KASPAROV tied for the "X3D German" (X3D-FRITZ) 2: 2.

Artificial intelligence pattern recognition

Using $ pattern recognition engine, branches have 2D recognition engine, 3D recognition engine, standing wave recognition engine and multi-dimensional recognition engine

2D recognition engine has launched fingerprint recognition, portrait recognition, text recognition, image recognition, license plate recognition; standing wave recognition engine has launched voice recognition; 3D recognition engine has launched fingerprint recognition Yudai Linhanghang (playing smart version 1.25)

Artificial Intelligence Automatic Engineering

Autonomous driving (OSO system)

Banknote printing factory ( assembly line)

Falcon System (YOD Drawing)

Artificial Intelligence Knowledge Engineering

With knowledge itself as the object of processing, research how to use artificial intelligence and software technology to design, construct and maintain knowledge systems

expert system

Intelligent search engine

Computer vision and image processing

Machine translation and natural language understanding

Data mining and knowledge discovery

AI- related works

"Reading Artificial Intelligence": Can Machines Really Think? Is human thinking just a complex computer program? This book looks at one of the most difficult scientific issues in artificial intelligence ever, and focuses on some of the main topics behind it. Artificial intelligence is more than a fictional concept. Human half-century research on the structure of intelligent bodies has shown that machines can defeat humanity's greatest chess player, humanoid robots can walk and interact with humans. Despite the long-standing declaration that smart machines are just around the corner, progress in this area has been slow and difficult. Consciousness and environment are two major problems that plague research. How should we make intelligent machines? Should it function like a brain? Does it need a body? From the profound foundational research of Turing's influence to the leap of robots and new artificial intelligence, this book clearly shows the development of artificial intelligence in the past half century in front of readers.

The Future of Artificial Intelligence: Explains the connotation of intelligence, explains the working principle of the brain, and tells us how to create a truly intelligent machine-such an intelligent machine will no longer be just a simple imitation of the human brain , Their intelligence will far exceed the human brain in many ways. Hawkins believes that from artificial intelligence to neural networks, earlier attempts to replicate human intelligence have been unsuccessful. The reason is that people have not really understood the meaning of intelligence and the human brain. The so-called intelligence is the ability of the human brain to compare the past and predict the future. The brain is not a computer, and it does not follow the steps and generate output based on input. The brain is a huge memory system. It stores experiences that reflect the true structure of the world to some extent. It can remember the sequence of events and their relationships, and make predictions based on memory. The foundation of intelligence, feeling, creativity, and perception is the memory-prediction system of the brain ...

"Philosophy of Artificial Intelligence": The philosophy of artificial intelligence is a branch of philosophy that accompanies the development of modern information theory and computer technology. This book collects fifteen representative papers by scholars in the field of artificial intelligence research, which have made pioneering contributions to the development of computer science and the establishment of artificial intelligence philosophy. These articles summarize the development of artificial intelligence, the development trend of the discipline, and important topics in artificial intelligence. Among these landmark works are: "Computers and Intelligence" by Allen Turing, the father of modern computer theory; "Mind, Brain, and Program" by American philosopher Searle; J. E. Hinton and others "Distributed Expressions" and "Escape from the Chinese House" by the book's editor and British artificial intelligence scholar M.A. Boden.

"Artificial Intelligence: A Modern Approach": This book is a comprehensive and comprehensive explanation of the core content of the field of artificial intelligence from the perspective of rational agents. A rare comprehensive textbook. The book is divided into eight parts: the first part is "artificial intelligence", the second part is "problem solving", the third part is "knowledge and reasoning", the fourth part is "planning", the fifth part is "uncertain knowledge and reasoning", the sixth Part "learning", part 7 "communication, perception and action", part 8 "conclusion". This book not only introduces a large number of basic concepts, ideas, and algorithms, but also describes the most cutting-edge progress in each research direction. At the same time, it collects and organizes detailed historical literature and events. Therefore, this book is suitable for researchers and students of different levels and fields. It can be used as a textbook or instructional bibliography for undergraduates and graduates in colleges and universities in the field of information and related fields. It can also be used as a reference for scientific and engineering personnel in related fields. book.

A Brief History of Artificial Intelligence

The legend of artificial intelligence can be traced back to ancient Egypt, but with the development of electronic computers since 1941, technology can finally create machine intelligence. The word "artificial intelligence" (ARTIFICIAL INTELLIGENCE) was originally proposed at the DARTMOUTH Society in 1956 Since then, researchers have developed numerous theories and principles, and the concept of artificial intelligence has also expanded. In its not long history, the development of artificial intelligence has been slower than expected, but has been moving forward, from 40 Many years ago, many AI programs have appeared, and they have also affected the development of other technologies.

Artificial intelligence computer age

An invention in 1941 revolutionized all aspects of information storage and processing. This invention, which appeared in the United States and Germany at the same time, was an electronic computer. The first computer took up several large air-conditioned rooms and was a programmer's It was a nightmare: thousands of lines were set up just to run a program. The improved computer capable of storing programs in 1949 made it easier to enter programs, and the development of computer theory produced computer science and eventually prompted The emergence of artificial intelligence. The invention of the computer to process data electronically provided a medium for the possible realization of artificial intelligence.

Although computers provided the necessary technical foundation for AI, the connection between human intelligence and machines was not noticed until the early 1950s. NORBERT WIENER was one of the first Americans to study the theory of feedback. The most familiar example of feedback control is Thermostat. It compares the collected room temperature with the desired temperature, and responds by turning the heater on or off to control the ambient temperature. The importance of this research on feedback loops lies in: It was pointed out that all intelligent activities are the result of feedback mechanisms. It is possible to simulate the feedback mechanism with machines. This discovery has a great impact on the development of early AI.

In late 1955, NEWELL and SIMON made a program called "LOGIC THEORIST". This program was considered by many to be the first AI program. It represented each problem as a tree model, and then chose The branch that is most likely to get the correct conclusion to solve the problem. The impact of the "Logic Expert" on the public and the field of AI research has made it an important milestone in the development of AI. In 1956, the John McCarth organization, considered the father of artificial intelligence After a society, many experts and scholars interested in machine intelligence gathered for a month of discussion. He invited them to VERMONT to participate in the "DARTMOUTH Artificial Intelligence Summer Research Society". Since then, this field has been named " Artificial Intelligence ". Although the DARTMOUTH Society was not very successful, it did gather the founders of AI and laid the foundation for future AI research.

In the 7 years after the DARTMOUTH conference, AI research has begun to develop rapidly. Although this field has not been clearly defined, some ideas in the conference have been reconsidered and used. CARNEGIE MELLON University and MIT have begun to set up AI research centers. Research faces new Challenge: The next step is to establish a system that can solve the problem more effectively, such as reducing the search among "logical experts"; and the system that can learn by itself.

A new program, the first version of the "Universal Problem Solver" (GPS), was tested in 1957. This program was developed by the same group that produced the "Logic Expert". GPS extends the feedback principle of WIENER and can solve Many common sense questions. Two years later, IBM set up an AI research group. HERBERT GELERNETER spent 3 years making a program to solve geometric theorems.

When more and more programs emerge, MCCARTHY is busy with a breakthrough in the history of AI. In 1958, MCCARTHY announced his new achievement: LISP language. LISP is still in use today. "LISP" means "table processing" ( LIST PROCESSING), which was quickly adopted by most AI developers.

In 1963, MIT received a $ 2.2 million grant from the US government for research on machine-assisted identification. This funding came from the Defense Advanced Research Projects Agency (ARPA), which has guaranteed that the United States is ahead of the Soviet Union in technological progress. This plan attracts Computer scientists from all over the world have accelerated the pace of development of AI research.

Artificial Intelligence Contest

LOEBNER (Artificial Intelligence)

The creation of a machine brain (artificial intelligence) that is parallel to the human brain with human wisdom is a very tempting field for human beings, and human beings have struggled for many years to realize this dream. From the perspective of a language researcher, it is quite difficult to let machines and people communicate freely, and it can even be said to be a question that has never been answered. Human language and human intelligence are so complicated that our research has not touched the edge of the extension of the guiding essence.

A large number of artificial intelligence programs

A large number of programs have appeared in the next few years. One of them is called "SHRDLU". "SHRDLU" is part of the "Mini World" project, including research and programming in the micro world (for example, only a limited number of geometric shapes). At MIT by MARVIN Researchers led by MINSKY have found that computer programs can solve spatial and logical problems in the face of small-scale objects. Others such as "STUDENT", which appeared in the late 1960s, can solve algebra problems, and "SIR" can understand simple English sentences. These The results of the program are helpful for handling language understanding and logic.

Another development in the 1970s was the expert system. The expert system can predict the probability of a certain solution under certain conditions. Due to the huge capacity of the computer at that time, the expert system may draw rules from the data. The market application of the expert system is very wide. For ten years, expert systems have been used for stock market predictions, helping doctors diagnose diseases, and instructing miners to determine the location of mineral deposits. All this is made possible by the ability of expert systems to store laws and information.

In the 1970s, many new methods were used for AI development, such as the construction theory of MINSKY. In addition, DAVID MARR proposed a new theory in machine vision, for example, how to identify the basic information such as the shadow, shape, color, border, and texture of an image Image. By analyzing this information, you can infer what the image may be. Another achievement of the same period is the PROLOGE language, which was proposed in 1972. During the 1980s, AI moved faster and entered the business field more. 1986 The sales of AI-related software and hardware in the United States amounted to $ 425 million. Expert systems are particularly demanded for their utility. Companies like Digital Electric use XCON expert systems to program VAX mainframes. DuPont, General Motors, and Boeing also rely heavily on experts System. In order to meet the needs of computer experts, some companies producing expert system-assisted production software, such as TEKNOWLEDGE and INTELLICORP, were established. In order to find and correct errors in existing expert systems, there are other expert systems designed.

Artificial Intelligence Everyday Life

People are beginning to feel the impact of computers and artificial intelligence technology. Computer technology is no longer just a small group of researchers in the laboratory. Personal computers and numerous technology magazines make computer technology appear to people. With the likes of the American Artificial Intelligence Association Foundation. Because of the need for AI development, there has also been a wave of researchers entering private companies. More than 150 companies like DEC (which employs more than 700 employees in AI research) spent a total of $ 1 billion on internal AI development teams.

Other AI fields also entered the market in the 1980s. One of them is machine vision. The results of MINSKY and MARR are now used in cameras and computers on the production line for quality control. Although still very crude, these systems have been able to distinguish between black and white. The shape of the object is different. By 1985, there were more than 100 companies in the United States producing machine vision systems, with sales totaling 80 million U.S. dollars.

But the 1980s were not all good times for the AI industry. The demand for AI systems fell from 86-87, and the industry lost nearly $ 500 million. Two companies, such as TEKNOWLEDGE and INTELLICORP, lost more than $ 6 million, about A huge loss of one-third of the profit has forced many research leaders to cut funding. Another disappointment is the so-called "smart truck" supported by the Defense Advanced Research Projects Agency. This project aims to develop a solution that can complete many battlefields Mission robot. Due to the project's flaws and unsuccessful success, PENTAGON stopped funding for the project.

Despite these frustrated events, AI is still slowly recovering. New technologies have been developed in Japan, such as the fuzzy logic pioneered in the United States, which can change from uncertain

Artificial intelligence robot (2 photos)

Conditional decision-making; and neural networks are considered as possible ways to achieve artificial intelligence. In short, AI was introduced to the market in the 1980s and showed practical value. It is certain that it will be the key to the 21st century. Technology Acceptance Test During the Desert Storm operation, the military s smart devices withstood the test of war. Artificial intelligence technology was used in missile systems and early warning displays and other advanced weapons. AI technology has also entered the home. The increase in intelligent computers has attracted Public interest; some applications for Apple and IBM compatible machines such as voice and text recognition are available; using fuzzy logic, AI technology simplifies camera equipment. Greater demand for artificial intelligence-related technologies is driving new advances Artificial intelligence has and will inevitably change our lives.

Comparison of AI strength

A popular definition of artificial intelligence, which is also an earlier definition of this field, was proposed by John McCarthy at the 1956 Dartmouth Conference: artificial intelligence is to make machines The behavior looks like the intelligent behavior that a person exhibits. But this definition seems to ignore the possibility of strong artificial intelligence (see below). Another definition refers to the intelligence displayed by artificial machines. Generally speaking, most definitions of artificial intelligence can be divided into four categories: machines "think like humans", "act like humans", "think rationally" and "act rationally". "Action" here should be understood broadly as taking action, or making decisions about action, rather than physical action.

Strong artificial intelligence (BOTTOM-UP AI)

The strong artificial intelligence point of view believes that it is possible to create intelligent machines that can really reason and solve problems, and that such machines can be considered conscious and self-conscious. There are two types of strong artificial intelligence:

Human-like artificial intelligence, that is, the thinking and reasoning of machines, is just like human thinking.

Non-human-like artificial intelligence, that is, machines produce consciousness and consciousness that are completely different from humans, and use completely different reasoning methods than humans.

Weak artificial intelligence (TOP-DOWN AI)

The weak artificial intelligence point of view believes that it is impossible to create intelligent machines that can really reason (REASONING) and solve problems (PROBLEM_SOLVING). These machines just look like intelligent, but do not really have intelligence, and they will not have a sense of autonomy .

Mainstream scientific research focuses on weak artificial intelligence, and it is generally believed that this research field has achieved considerable achievements. Research on strong artificial intelligence is in a state of stagnation.

The philosophical debate on strong artificial intelligence

The term "strong artificial intelligence" was originally coined by John Rogers Hiller for computers and other information processing machines, and is defined as:

"The strong artificial intelligence point of view is that computers are not only a tool used to study human thinking; instead, as long as appropriate programs are run, the computer itself is thinking." , VOL. 3, 1980) This refers to activities that make computers intelligent. Here the meaning of intelligence is ambiguous and uncertain, as mentioned below are examples. When using a computer to solve a problem, you must know a clear procedure. However, even when people do not know the procedure, there are many cases where people have managed to solve the problem cleverly according to the HEU-RISTIC method. Such as recognition of written text, graphics, sound, etc., the so-called recognition model is an example. Furthermore, the improvement in learning ability and inductive reasoning, and reasoning based on analogy are also examples. In addition, although the procedure for solving the problem is clear, it takes a long time to implement. For such a problem, people can find a very good solution in a short period of time, such as competitive games. In addition, the computer cannot understand its meaning without giving sufficient and logically correct information, and people can only catch inadequate and incorrect information based on appropriate supplementary information and grasp Live it meaning. Natural language is an example. Processing natural language with a computer is called natural language processing.

The debate on strong artificial intelligence differs from the more general debate on monism and dualism. The main point of the argument is: if the only working principle of a machine is to convert encoded data, is this machine thinking? Hiller thinks this is impossible. He gave an example of a Chinese room to illustrate that if the machine is only converting data, and the data itself is a coded expression of some things, then the correspondence between this code and this actual thing is not understood It is impossible for the machine to have any understanding of the data it processes. Based on this argument, Hiller believes that even if the machine passes the Turing test, it does not necessarily mean that the machine really has the thinking and consciousness like humans.

There are also different views of philosophers. DANIEL C. DENNETT believes in his book CONSCIOUSNESS EXPLAINED that man is nothing more than a machine with a soul. Why do we think that man can have intelligence and ordinary machines cannot? He believes that data conversion machines like the above are likely to have thought and awareness.

Some philosophers believe that if weak artificial intelligence is achievable, then strong artificial intelligence is also achievable. For example, SIMON BLACKBURN said in his introduction to philosophy, THINK, that a person's actions that appear to be "smart" do not really indicate that the person is really intelligent. I can never know if another person is really as intelligent as me, or if she / he just looks intelligent. Based on this argument, since weak artificial intelligence believes that it can make a machine look intelligent, it cannot be completely denied that the machine is really intelligent. BLACKBURN considers this to be a subjective issue.

It should be pointed out that weak artificial intelligence is not completely opposed to strong artificial intelligence, that is, even if strong artificial intelligence is possible, weak artificial intelligence is still meaningful. At least, what computers can do today, such as arithmetic operations, was considered to require intelligence more than a hundred years ago.

Artificial intelligence policy measures

On June 17, 2019, the National New Generation Artificial Intelligence Governance Professional Committee released the "New Generation Artificial Intelligence Governance Principles-Development of Responsible Artificial Intelligence", which proposed a framework and action guide for artificial intelligence governance. This is an important result of China's promotion of the healthy development of next-generation artificial intelligence, the strengthening of artificial intelligence legal, ethical, and social issues, and the active promotion of global governance of artificial intelligence. ^[5]

Artificial Intelligence Research Project

The research direction of artificial intelligence has been divided into several sub-fields. Researchers hope that an artificial intelligence system should have certain specific capabilities. These capabilities are listed and explained below.

Artificial intelligence solves problems

Early artificial intelligence researchers directly imitated humans for step-by-step reasoning, just like the way humans think when playing board games or performing logical reasoning. By the 1980s and 1990s, using probability and economic concepts, artificial intelligence research had also developed very successful methods for dealing with uncertain or incomplete information.

For difficult problems, a large amount of computing resources may be required, that is, a "probable combination explosion" has occurred: when the problem exceeds a certain scale, the computer will require astronomical orders of memory or computing time. Finding more efficient algorithms is a priority AI research project.

The mode of human problem solving is usually the fastest and intuitive judgment, rather than conscious, step-by-step derivation. Early artificial intelligence research usually used step-by-step derivation. Artificial intelligence research has advanced with this "sub-representative" problem-solving approach: Substantiated AGENT research emphasizes the importance of perceptual motion. Neural network research attempts to reproduce this skill by simulating the structure of human and animal brains.

Artificial intelligence knowledge representation

AN ONTOLOGY REPRESENTS KNOWLEDGE AS A SET OF CONCEPTS WITHIN A DOMAIN AND THE RELATIONSHIPS BETWEEN THOSE CONCEPTS.

Main article: Knowledge representation and knowledge base

Artificial intelligence planning

Smart AGENT must be able to set goals and achieve them. They need a way to build a predictable model of the world (representing the entire state of the world as a mathematical model and predicting how their behavior will change the world) so that they can choose the behavior that has the greatest effect. In traditional planning problems, intelligent AGENT is assumed to be the only influential in the world, so what behavior it will do is already determined. However, if this is not the case, it must regularly check whether the state of the world model is consistent with its own predictions. If it doesn't, it must change its plan. Therefore, the intelligent agent must have the ability to reason in the state of uncertain results. In multi-AGENT, multiple AGENT plans achieve certain goals in a cooperative and competitive manner. The use of evolutionary algorithms and swarm intelligence can achieve an overall emergent goal.

Artificial intelligence learning

Main article: Machine learning

The main purpose of mechanical learning is to obtain knowledge from users and input data, which can help solve more problems, reduce errors, and improve the efficiency of problem solving. For artificial intelligence, mechanical learning has been important from the beginning. In 1956, at the original Dartmouth Summer Conference, Raymond Solomonov wrote about probabilistic mechanical learning without surveillance: a machine for inductive reasoning.

Artificial intelligence natural language processing

Main article: Natural language processing

Artificial intelligence movement and control

Main article: Robotics

AI perception

Main articles: Machine perception, computer vision, and speech recognition

Machine perception refers to the ability to use sensor input (such as cameras, microphones, sonars, and other special sensors) to infer the state of the world. Computer vision can analyze image input. There are also speech recognition, face recognition and object recognition.

Artificial intelligence social

Main article: Affective computing

KISMET, a robot with social capabilities such as expressions

Emotional and social skills are important for a smart AGENT. First, by understanding their motivations and emotional states, agents can predict the actions of others (this involves factor game theory, decision theory, and the detection of emotions and emotional perception capabilities that can shape people). In addition, for good human-computer interaction, smart agents also need to show emotions. At least it must appear polite to deal with humans. At the very least, it should have normal emotions.

Artificial intelligence creativity

Main article: Computer creativity

A sub-area of artificial intelligence that represents creativity as defined by theory (from a philosophical and psychological perspective) and practical (the output of a system produced through a specific implementation is an idea that can be considered, or the system recognizes and evaluates creativity) . Research in related fields includes artificial intuition and artificial imagination.

Artificial Intelligence Multiple Intelligences

Most researchers hope that their research will eventually be incorporated into a multi-intelligence (called strong artificial intelligence) that combines all the above skills and surpasses most human capabilities. Some people think that achieving these goals may require anthropomorphic traits, such as artificial consciousness or artificial brains. Many of these issues are considered artificial intelligence integrity: to solve one of them, you have to solve all of them. Even a simple and specific task, such as machine translation, requires the machine to follow the author's argument (inference), know what is being talked about (knowledge), and faithfully reproduce the author's intention (emotional computing). Therefore, machine translation is considered to have artificial intelligence integrity: it may require strong artificial intelligence, just like humans.

Artificial intelligence impact

(1) Impact of artificial intelligence on natural sciences. In disciplines that require the use of mathematical computer tools to solve problems, the help of AI is self-evident. More importantly, AI in turn helps humans finally recognize the formation of their own intelligence.

(2) The impact of artificial intelligence on the economy. Expert systems go deeper into all walks of life and bring huge macro benefits. AI has also promoted the development of the computer industry and the network industry. But at the same time, it also brings about labor and employment issues. Due to the application of AI in technology and engineering, it can replace human beings in various technical work and mental labor, which will cause drastic changes in social structure.

(3) The impact of artificial intelligence on society. AI also provides a new model for human cultural life. Existing games will gradually develop into more intelligent interactive cultural and entertainment means. Today, the application of artificial intelligence in games has penetrated into the development of major game manufacturers.
With the development of artificial intelligence and intelligent robots, it has to be discussed that artificial intelligence itself is advanced research, and it is necessary to carry out modern scientific research with a future perspective, so it is likely to reach the ethical bottom line. As a sensitive issue that may be involved in scientific research, it is necessary to prevent possible conflicts and prevent them early, instead of waiting for solutions to resolve the conflicts.

Artificial intelligence applications

Machine translation, intelligent control, expert systems, robotics, language and image understanding, genetically programmed robot factories, automatic programming, aerospace applications, huge information processing, storage and management, execution of complex or large-scale organisms that cannot be performed by living organisms Tasks and much more.

It is worth mentioning that machine translation is an important branch and first application area of artificial intelligence. However, judging from the existing achievements in machine translation, the quality of the machine translation system is still far from the ultimate goal; and the quality of machine translation is the key to the success of the machine translation system. Professor Zhou Haizhong, a Chinese mathematician and linguist, once pointed out in his dissertation "Fifty Years of Machine Translation": To improve the quality of machine translation, the first problem is to solve the problem of the language itself, not the problem of program design; Machine translation system is definitely unable to improve the quality of machine translation. In addition, it is impossible for machine translation to reach the level of "faithfulness, elegance, and elegance" without human beings understanding how the brain performs fuzzy recognition and logical judgment of language. of. After smart home, artificial intelligence has become a new outlet for the home appliance industry, and Changhong is becoming the first home appliance giant to set off this wave. ^[6] Changhong released two new products of CHiQ smart TVs, which mainly focus on mobile phone remote control, take away, watch at any time, and view by category ^[7]

Artificial intelligence buzzwords

In December 2017, artificial intelligence was selected as the "Top Ten Buzzwords of Chinese Media in 2017".

Reason for selection: After years of evolution, the development of artificial intelligence has entered a new stage. In order to seize the major strategic opportunities for the development of artificial intelligence, build the first-mover advantage of the development of artificial intelligence in China, and speed up the construction of innovative countries and the world's science and technology powerhouse. On July 20, 2017, the State Council issued the "New Generation Artificial Intelligence Development Plan." The "Planning" put forward the guiding ideology, strategic goals, key tasks and guarantee measures for the development of China's new generation of artificial intelligence in 2030, laying an important foundation for the further acceleration of artificial intelligence in China. ^[2]