人工智能的起源，研究范围及发展前景
湖南农业大学课程论文
学 院：信息科学与技术学院 班 级：1班
姓 名：肖婷婷 学 号：200841842112
课程论文题目：人工智能的起源，研究范围及发展前景
课程名称：人工智能
评阅成绩：
成绩评定教师签名：
日期：2010-3-27
The origin of AI, research scope and prospects
Xiao Tingting
Information Science and Technology College of Computer and Science 1 class
Abstract: This article is the origin of artificial intelligence brief, simple introduction to the scope of the study of artificial intelligence and the development prospects of the speculation and analysis.
Keywords: The origin of AI, the research scope of AI, the prospects of development.
1.About Artificial Intelligence
AI (Artificial Intelligence, referred to as AI) is a branch of computer science, he's real intelligent inquiry and to create a human intelligence can react in a way similar to the purpose of intelligent machines. The emergence and development of artificial intelligence is a thinking first scientific revolution, the emergence and development of it to some extent dependent on the thinking and scientific revolution, but it also on the human way of thinking and methods of a profound transformation. Artificial intelligence is most closely with the philosophy of science topics, it brings together from cognitive psychology, linguistics, neuroscience, logic, mathematics, computer science, robotics, economics, sociology, etc. within the research . Over the past half-century of artificial intelligence in human knowledge and transform the world their own way played an important role. Artificial Intelligence usually refers to computer implemented through normal intelligence. Of artificial intelligence research can be divided into two basic schools, that school of thought and symbols Connectionism school. Symbol school of thought to logic, cognitive psychology, research methods, with knowledge of the form of symbols, and computing as a basis for reasoning that new knowledge. Connectionism and animals from the intelligence produced in the starting mechanism, as long as the human brain, through the establishment of a similar neural network structure, can produce intelligent, primarily a means for studying neural networks. There are two attitudes of artificial intelligence research: strong and weak artificial intelligence, artificial intelligence, the former thinking that the AI can be achieved with the degree of understanding, can have a real intelligent; the latter that of AI is only through it to explore human cognition, its intelligence but imitation is not entirely intelligent. In any view, they all rise to a series of important questions: What is intelligence? What is thinking? How do we get "smart" level?
2.The origins of AI
The legend of artificial intelligence can be traced back to ancient Egypt, But with the computer since 1941, the development of technology has finally able to create intelligent machines. "Artificial Intelligence" in 1956, the term was first put forward at Dartmouth Institute. Since then, researchers have developed many theories and principles which will expand the concept of artificial intelligence, in its not a long history, the development of artificial intelligence, slower than expected, but has been moving forward, from 40 years pre-emergence to the present, there have been a number of AI programs, and they also affect the development of other technologies.
The 1941 invention of a storage and processing of information in all aspects of the revolution took place. This also occurs in the United States and Germany, the invention is a computer. The first computer to be occupied by several large air-conditioned rooms, on the programmer it is a nightmare: just to run a program is necessary to set the line .1949 thousands of years of improved energy stored-program computer makes it easy to enter the program more, but also the development of theoretical computer generated computer science, and ultimately prompted the emergence of artificial intelligence. the computer to process data through electronic means in this invention may be achieved for the AI provides a medium.
Although the computer AI provides the necessary technical foundation, but not until the early 50s that people began to notice the human link between intelligence and machines. Norbert Wiener is the first study of feedback theory, one of the Americans. The most familiar example of feedback control automatic thermostat. it collected from the room temperature and desired temperature is, and respond to large or off to open a small heater to control the ambient temperature. The study of the feedback loop is important because: Wiener theory pointed out that all of the intelligence activities are the result of feedback mechanisms. while the feedback mechanism it is possible to use the machine simulation. The finding of early development of a great influence AI.
At the end of 1955, Newell and Simon made a program named Logic Theorist . This program considered by many to be the first AI program. It is said that every question into a tree model, and then select the most likely to be the correct conclusion that to solve a problem. The Logic Theorist to the public and the impact of AI research areas make it an important milestone in the development of AI.1956 years, is considered the father of artificial intelligence, John McCarthy organized a once learned, will be of interest to many pairs of Machine Intelligence experts and scholars together for a month discussion. He invited them to Vermont to participate in Dartmouth Summer Research Association of Artificial Intelligence. Since that time, this area has been named as Artificial Intelligence. Although the Dartmouth Institute is not very successful, but it does focus on the AI's founding fathers, and later laid the foundation for AI research.
Dartmouth meeting of the 7 years, AI research began to develop rapidly. Although not yet a clear definition of this area, the meeting some of the ideas have been re-considered and used. Carnegie Mellon University and MIT began to build AI research center. Research is facing new Challenges: The next step need to be more effective problem-solving systems, such as reducing search in the Logic Theorist; there is the establishment of self-learning system.
In 1957, a new program named General Problem Solving Machine whose first version was tested. This program is produced by the same group as the Logic Theorist . GPS expanded Wiener feedback principle, can solve many of common sense. Two years later, IBM set up an AI research group. Herbert Gelerneter spent three years has produced a solution procedure for geometric theorem.
As more and more procedures surge, McCarthy is busy in the history of an AI breakthrough .1958 McCarthy years, announced his new results: LISP language. LISP still use today. LISP means LISt Processing , it was soon adopted by most AI developers.
3.The research scope of AI
At present, the artificial intelligence research is carried out in specific areas combined. Basically, the following areas; Expert System Expert systems rely on existing knowledge of human experts set up knowledge systems, the current research in artificial intelligence expert system is carried out earlier, the most active areas of the most effective, widely used in medical diagnosis, geological exploration, petrochemical, military, education and so on. It is in the specific field of knowledge and experience with the corresponding system of procedures, which apply artificial intelligence techniques to simulate human experts to solve problems in the thinking process to solve various problems within the field, at or close to the expert level. Machine learning For a computer that has the knowledge that there are two methods; one is by the knowledge engineer will summarize the relevant knowledge, order, and that the computer can accept input into the computer processing methods. The other is the computer itself has the ability to acquire knowledge, it can learn from existing knowledge of mankind, and in practice does not sum up, improve, this approach is called machine learning. Machine learning research, mainly in the following three aspects: First, the mechanism of human learning, human brain thinking process; and machine learning methods; and the establishment of task-specific learning system. Machine learning research in information science, brain science, neural psychology, logic, fuzzy math and other subjects based on. Dependent on these subjects and common development. Great progress has been made, but failed to completely solve the problem.
Pattern Recognition
Pattern recognition is how to make machines with sensing capability, the main mode of visual and auditory pattern recognition. Such as object recognition, terrain, image, font (such as the signature) and so on. In the daily life and have a large military purposes. Developed rapidly in recent years, fuzzy mathematical model, artificial neural network models gradually replace the traditional method of using statistical models and structure pattern identification method. Special neural network method in pattern recognition made great progress. Understanding of natural language If the computer can "understand" human language (such as Chinese, English, etc.), they can operate the computer by speaking, this will give people with great convenience. Computer understanding of natural language has the following three objectives: First, the computer can correctly understand the nature of human language input information, and correctly answer (or response) the information entered. The second is the computer input of information to produce a corresponding summary of the contents of the input and repeat. Third, the computer can enter the natural language translated into another language requirements, such as the Chinese to English or English to Chinese and so on. At present, the study of language or computer language, automatic translation, it made a lot of try, have not yet found the best way, need to be further explored further.
Robotics
Robot is a simulated human behavior can be mechanical, it has experienced three generations of the development process:
The first generation (process control) robot: This robot is generally in the following two ways, "learn to" work; one is from the pre-press workflow designer written good program stored in the robot's internal memory, the program under the control of the work. The other is called "Demonstration - reproduction" mode, this method is the first mission in the robot before the operation by the technical staff for guiding the robot, the robot will step by step the operation record, every step of the operations are expressed as a command. Demonstration after the end of the robot in order to complete the work according to instructions (ie reproduction). If the task or the environment have changed, to re-programming. This robot can be conscientious in the machine tool, furnace, welding machine, production line work. Before commercialization, most practical robots fall into this category. The greatest disadvantage of this robot is that it can only be rigidly follow procedures to complete the work environment slightly changes (such as processing materials slightly inclined) will go wrong, even dangerous, this is because it has no sensation in Japan has occurred in the robot to the scene of workers arrested and stuffed into a tool the following situation.
The second generation (adaptive) Robot: This robot is equipped with a corresponding sense of sensors (such as visual, auditory, tactile sensors, etc.), can achieve operating environment, operation objects and other simple information, the robot body by computer analysis, processing, control the robot's movements. Although the second generation of intelligent robots has some primary, but it needs the coordination of technical staff. Already have some commercial products.
Third-generation (smart) robots: intelligent robots with similar human intelligence, it is equipped with highly sensitive sensors, which have more than most people's vision, hearing, smell, touch, capable of analyzing the information on the perception and control their behavior, dealing with environmental changes, entrusted to the complex and difficult task. But also self-study, draw conclusions, summarize, has to acquire knowledge to improve the capacity. Development of Intelligent Robot currently only have some intelligence, and genuine sense of intelligence robot, still a long way. Intelligent Decision Support System
Decision support system is a management science areas, with "knowledge - intelligence" has a very close relationship. Since the expert system in the 80's success in many ways, artificial intelligence, especially in intelligence and knowledge processing Jishu used in decision support systems, and expanded the application of decision support systems, improve the system ability to solve problems This, then, smart decision support system.
Artificial Neural Networks
Artificial neural network is in the mystery of the human brain be inspired to try to use a large number of processing units (artificial neurons, processing elements, electronic components, etc.) mimic the structure and working mechanism of neural systems engineering.
In artificial neural networks, information processing is the interaction between neurons is achieved, knowledge and information storage interconnect network elements showed physical contact between the distributed network of learning and recognition depend on the neuron connection weights of the dynamic evolution.
Over the years, artificial neural network research has made great progress, and become a unique style of information processing disciplines. Of course, the present study also only a few simple artificial neural network model. To establish a complete set of theories and technical systems, the need for greater efforts and explore. However, artificial intelligence, artificial neural networks have become an extremely important area of research.
4.The prospects of AI
Since the emergence of artificial intelligence, along with the machine can do more and more people are "smart," more and more demanding requirements, from complex mathematical operations, logical calculus to deductive reasoning, to require the machine with "creative" , defeated world chess champion "Deep Blue" men (although some people think that the machine is to win by brute force rather than intelligence), machine to composition, painting, poetry is also not news. machine in the "creation" in accordance with the rules when running the same, this is one of the reasons against machine intelligence. However, the human creation is not a groundless, fabricated, and it must also have a set of rules on which basis. "Created" before the request is contrary to expected, intelligence is the condition differences according to the environment should change their ways, when the computer's rules complex enough to experience the steps are enough for a long time, people could not anticipate that its final act, it also have the ability to achieve intelligence.
The development of two types of artificial intelligence view, a class that was indexed according to the technical progress of the trend, the computer's calculation speed will be more than one person's brain by 2020, before the end of the 21st century, more than the sum of all brain, the machine will self- design their own future; another seemingly crazy idea from the "Father of the world's artificial brain," Hugo Degaris, late in this century, whether in the manufacturing of human intelligence like God will lead to a difference machine war. negative view that artificial intelligence created by man is bound to imitate the human brain, with people raising the level of self-awareness, imitation can only be infinitely close to but not beyond the human brain, and mind the complexity of the machine can not copy of, like people, even beyond the human imagination intelligent machines will remain at the stage of artificial intelligence will always be a tool used by man.
I think that discussion of "artificial intelligence than human intelligence" is meaningless, because from a variety of ways, they both are two different kinds of intelligence, biological intelligence and machine intelligence have their own characteristics, should be changed to " intelligent "understanding, but to attach importance to an existing machine intelligence. In today's development level, machine intelligence is under the control of the people running, as a tool is one purpose of extension, with the practice of raising the level of human, machine intelligence will continue to develop, but a reasonable person would find its cooperation, to man the methods used.
Some scientists have said so: "If a scientist pointed out that something is possible he is almost certainly correct; but when he says something is impossible, he is probably wrong." Someone 30 years of the 20th century predict a bomb that can destroy a city, when people regard him as a madman. Bill Gates predicted in 1981 has: "for anyone, 640K of memory should be good enough." Now it seems the reality of the rapid development of production and a strong contrast. Likely future of the world beyond our imagination, I think, on the development of artificial intelligence should also keep an open mind.
5.Conclusion
The study of artificial intelligence to enable computers to simulate the human thinking process, and some intelligent behavior (such as learning, reasoning, thinking, planning) of science, mainly including computer intelligent in principle, Zhi Zao Leisiyuren intelligent computer brain so that the computer to achieve a higher level of application. Artificial intelligence will involve computer science, psychology, philosophy and linguistics studies. Can be said that almost all the natural and social sciences disciplines, ranging far beyond the scope of computer science, artificial intelligence and thinking is the relationship between science and the relationship between practice and theory, artificial intelligence is the science and technology applications in the thinking level is a branch of one of its applications. Thinking point of view, artificial intelligence is not limited to logical thinking, to consider the image of the thinking, inspired breakthrough thinking can promote the development of artificial intelligence, mathematics is often considered to be the basis for a variety of disciplines in science, mathematics also entered the language, thinking in the field of artificial intelligence disciplines must borrow mathematical tools, not only in standard mathematical logic, the scope of the role of fuzzy mathematics, mathematics into artificial intelligence, they will promote each other and develop faster.
The development of artificial intelligence led to a series of deep philosophical questions, so that our consciousness, thinking further understand the nature. Since the artificial intelligence and cognitive science since 1956 there has been constantly exploring, it lengthened the hand, the liberation of the human brain, further understanding of its own people. However, when faced with awareness, people even more confused, and we as a kind of understanding, emotion, consciousness of the mystery of view, do not believe the machine can also produce these spiritual state. But the spirit of mystery, the human brain and artificial differences in the brain should not be opposed to artificial intelligence, reason, with the exploration of the deep mystery of consciousness will gradually disappear, the machine even if the person can not have the same sense, also With its unique "machine intelligence" to become a part of human society.
6.References
[1]Richard O.Duda,Peter E.Hart,David G.Stork(2001) Pattern recognition (second edition).
[2]Koza, J.R. (1994), Genetic Programming II: Automatic Discovery of Reusable Programs, MIT Press
[3]Langdon, W. B., Poli, R. (2001), Foundations of Genetic Programming, Springer-Verlag
[4] F.Zambonelli, N.Jennings, et al. Agent-oriented software engineering for intern applications
[5] S. Russel. P. Norvig. Artificial Intelligence: A Modern Approach, Prentice Hall/Pearson Education Internation:Englewood Cliffs(NJ),(2nd Edn),2003.
[6] Banzhaf, W., Nordin, P., Keller, R.E., Francone, F.D. (1997), Genetic Programming: An Introduction: On the Automatic Evolution of Computer Programs and Its Applications, Morgan Kaufmann
人工智能的起源，研究范围及发展前景.doc
