Development and application of artificial neural network
With the development of science and technology, artificial neural network technology has been developed unprecedentedly, and it has been widely used in many fields, providing a strong impetus for the development of artificial intelligence. The development of artificial neural network has gone through different stages, and it is an important part of artificial intelligence, and has formed its own unique characteristics in the process of development. This paper reviews the development of artificial neural network and discusses its application in various fields.
Artificial neural network; Development; App application
With the development of science and technology, various industries and fields are carrying out artificial intelligence research, which has become a hot spot for experts and scholars. Artificial neural network is an important branch developed on the basis of artificial intelligence, which plays an important role in promoting the development of artificial intelligence. Since its birth, artificial neural network has gone through different stages of development, and has been widely used in economic, biological, medical and other fields, and solved many technical problems.
Summary of 1 artificial neural network
Up to now, there is not a widely accepted unified definition of artificial neural network. Based on the opinions of experts and scholars, artificial neural network can be simply summarized as a computer information processing system that imitates the structure and function of human brain [1]. Artificial neural network has its own development characteristics, strong parallel structure and parallel processing ability, and can play a very good role in real-time and dynamic control; Artificial neural network has the characteristics of nonlinear mapping, which is helpful to deal with nonlinear control problems. Artificial neural network can master the ability of data induction and processing through training, so it can solve the problem that mathematical model is difficult to deal with. Artificial neural network has strong adaptability and integration, which can adapt to different scale information processing and large-scale integrated data processing and control; Artificial neural network is not only mature in software technology, but also has great development in hardware in recent years, which improves the information processing ability of artificial neural network system.
2. Development of artificial neural network
2. 1 budding stage
In the 1940s, biologist mcculloch and mathematician Pitts published an article, and put forward the M-P model of neurons for the first time. This theory laid a foundation for the research and development of neural network model, and on this basis, the research of artificial neural network was gradually launched. 195 1 year, psychologist Hebb put forward the numerical reinforcement rule of connection right, which paved the way for the development of neural network learning function. Later, biologist eccles confirmed the real shunt of synapses through experiments, which provided a real model basis and biological basis for neural network to study the simulation function of synapses [2]. Subsequently, a processor and an adaptive linear network model that can simulate behavior and conditioned reflex appeared, which improved the speed and accuracy of artificial neural network. The appearance of this series of research results provides the possibility for the formation and development of artificial neural networks.
2.2 trough period
In the early days of the formation of artificial neural network, people were only keen on its research and ignored its own limitations. Minskyh and Papert questioned the previous research results in 1969 through years of research on neural networks, and thought that the neural networks developed at present were only suitable for dealing with relatively simple linear problems, but could not solve nonlinear problems and multi-layer network problems. Because of their doubts, the development of neural networks has entered a trough, but during this period, experts and scholars have not stopped studying neural networks, and some corresponding research results have been achieved in response to their doubts.
2.3 Renaissance
Hopfield, an American physicist, put forward a new neural network model in 1982, and proved through experiments that the neural network can reach a stable state under certain conditions. Through his research and promotion, many experts and scholars resumed their research on artificial neural networks, which in turn promoted the development of neural networks [3]. With the continuous efforts of experts and scholars, various artificial neural network models have been put forward, the theoretical research of neural network has been deepened, and new theories and methods have emerged one after another, which has brought the research and application of neural network into a brand-new period.
2.4 stable development period
With the resurgence of artificial neural network research in the world, China has also ushered in the upsurge of related theoretical research, and made breakthrough progress in artificial neural network and computer technology. In the 1990s, the research in the field of neural networks in China has been further improved and developed. Neural networks can be used to solve the control problems of nonlinear systems, and the research results are remarkable. With the establishment of various publications related to artificial neural networks and the convening of relevant academic conferences, the research and application conditions of artificial neural networks in China have gradually improved and attracted international attention.
With the steady development of artificial neural network, the optical neural network system is gradually established, and the learning ability and adaptive ability of artificial neural network are improved by using the powerful function of optics. For the control problem of nonlinear dynamic system, effective measures are taken to improve the smoothness and accuracy of hyperplane. Later, some experts put forward the extraction algorithm of artificial neural network, which ensured the accuracy, but also increased the consumption and reduced the efficiency of neural network to a certain extent, so on this basis, an improved algorithm FERNN was proposed. The development of chaotic neural network has also made corresponding progress, which improves the generalization ability of neural network.
3 Application of artificial neural network
3. The application of1in information field
The application of artificial neural network in information field is mainly reflected in two aspects: information processing and pattern recognition. With the development of science and technology, contemporary information processing has become more and more complicated. Using artificial neural network system can imitate or even replace people's thinking, diagnose and solve problems automatically, and can easily solve many problems that traditional methods can't solve. It is widely used in military information processing. Pattern recognition is a process of sorting out and analyzing all kinds of information on the appearance of things and distinguishing and explaining things, so the process of processing information is very similar to the way of thinking of the human brain. Pattern recognition methods can be divided into two types, one is statistical pattern recognition, and the other is structural pattern recognition, which has been widely used in speech recognition and fingerprint recognition.
3.2 Application in medical field
Artificial neural network is very effective in dealing with nonlinear problems, but the composition of human body and the causes of disease formation are very complex and unpredictable, so it is difficult to grasp the manifestations and changing laws of biological signals. There are complex nonlinear links in information detection and analysis, and it is of special significance to apply artificial neural network to solve these nonlinear problems [5]. At present, the application in the medical field involves all aspects of theory and clinic, the most important of which is the detection and automatic analysis of biological signals and the application of expert systems.
3.3 Application in the economic field
The information composition of commodity prices, supply and demand, risk coefficient and so on in the economic field is also very complex and difficult to predict. Artificial neural network can deal with incomplete information and fuzzy uncertain information simply and clearly, which has incomparable advantages over traditional economic statistical methods and has stronger stability and reliability in data analysis.
3.4 Application in other fields
Artificial neural network is widely used in control field, traffic field, psychology field and so on. It can deal with difficult nonlinear problems, comprehensively manage transportation, solve many problems that traditional methods can't solve with its high adaptability and excellent simulation performance, and promote the rapid development of various fields.
4 abstract
With the development of science and technology, artificial intelligence system will enter a more advanced stage of development, and artificial neural network will also get faster development and wider application. Artificial neural network may not completely replace the human brain, but its unique nonlinear information processing ability has solved many problems that cannot be solved by human beings, and it has been successfully applied in various fields of intelligent systems, and its future development trend will be more intelligent and integrated.
refer to
[1] Xu Yongmao, Feng. Development of artificial neural network and its application in control [J]. Progress of chemical industry,1993 (5): 8-12,20.
Luo Yufeng Tang Suli. Development and application of artificial neural network technology [J]. Computer Development and Application, 2009 (10): 59-6 1.
Chai Ren. Development and prospect of artificial neural network and neural network control [J]. Journal of Xingtai Vocational and Technical College, 2009 (5): 44-46.
, Zhu,. Development of artificial neural network and its research status in geotechnical engineering [J]. Henan Water Conservancy, 2004 (1): 22-23.
Cui Yonghua. Study on forecasting model of river confluence based on artificial neural network and its application [D]. Zhengzhou University, 2006.
Sharing is better.