11-1   Artificial Intelligence Techniques

Artificial Intelligence (AI) is a term that in its broadest sense would indicate the ability of an artifact to perform the same kinds of functions that characterize human thought. The possibility of developing some such artifact has intrigued human beings since ancient times. With the growth of modern science, the search for AI has taken two major directions: psychological and physiological research into the nature of human thought, and the technological development of increasingly sophisticated computing systems. [1]

In the latter sense, the term AI has been applied to computer systems and programs capable of performing tasks more complex than straightforward programming, although still far from the realm of actual thought. The most important fields of research in this area are information processing, pattern recognition, game-playing computers, and applied fields such as medical diagnosis. Current research in information processing deals with programs that enable a computer to understand written or spoken information and to produce summaries, answer specific questions, or redistribute information to users interested in specific areas of this information. Essential to such programs is the ability of the system to generate grammatically correct sentences and to establish linkages between words, ideas, and associations with other ideas. [2] Research has shown that whereas the logic of language structure — its syntax — submits to programming, the problem of meaning, or semantics, lies far deeper, in the direction of true AI. [3]

In medicine, programs have been developed that analyze the disease symptoms, medical history, and laboratory test results of a patient, and then suggest a diagnosis to the physician. The diagnostic program is an example of so-called expert systems— programs designed to perform tasks in specialized areas as a human would. Expert systems take computers a step beyond straightforward programming, being based on a technique called rule-based inference, in which pre-established rule systems are used to process the data. [4] Despite their sophistication, systems still do not approach the complexity of true intelligent thought.

Many scientists remain doubtful that true AI can ever be developed. The operation of the human mind is still little understood, and computer design may remain essentially incapable of analogously duplicating those unknown, complex processes. Various routes are being used in the effort to reach the goal of true AI. One approach is to apply the concept of parallel processing — interlinked and concurrent computer operations. Another is to create networks of experimental computer chips, called silicon neurons that mimic data-processing functions of brain cells. Using analog technology, the transistors in these chips emulate nerve-cell membranes in order to operate at the speed of neurons.

In recent years there has been intense interest in developing artificial intelligence techniques for a wide variety of scientific and engineering applications. A comprehensive survey paper provides a thorough review of intelligent systems in process engineering. The process control research in this area has largely been concerned with three AI methods: knowledge-based systems, neural networks and fuzzy logic.

Knowledge-based systems

Knowledge-based systems (ABS), also referred to as expert systems, use a set of ‘rules’ to perform logical inferences about the state of a process operation or some other activity of interest. Industrial applications of KBS systems have largely been concerned with either supervisory control or diagnostic and monitoring activities. Supervisory control applications have included the following problems: complex control schemes, recovery from extreme conditions and emergency shutdowns.

Future applications of expert systems will be facilitated by real-time KBS which enable the user to integrate plant data and process models in an expert system shell which has a sophisticated graphical interface. This combination provides a powerful vehicle for on-line process monitoring, especially diagnostics and fault detection.

The early enthusiasm for KBS has been tempered by the realization that a considerable effort is required to codify the available expertise. Furthermore, if each potential application has a significant number of unique features, it is less feasible to spread the development costs over a large number of projects. Despite this inherent problem, the industrial employment of KBS for applications such as process diagnosis and supervisory control is significant and growing at an impressive rate.

Neural networks

Neural networks provide a powerful approach for developing empirical nonlinear models for a wide variety of physical phenomena. In the area of process control, they have been used for a variety of traditional activities, such as developing nonlinear dynamic models and control system design. Neural networks also provide a promising approach for pattern recognition problems such as sensor data analysis and fault detection where traditional modeling techniques are not easily applied.

The commercial availability of neural network software for use by non-specialists should continue the current widespread interest in neural network applications for process control. However, at the present time it is difficult to assess the extent to which process control applications of neural networks are being used in industry.

Fuzzy control systems

Fuzzy logic provides a conceptual framework for practical problems where some process variables are represented as “linguistic variables” which have only a few possible values (e.g. very large, large, normal, small etc.). The linguistic variables can then be processed a set of rules. Thus applications of fuzzy logic and fuzzy control can be viewed as special cases of KBS, which have fuzzy boundaries for the rules.

Unlike more general KBS and neural nets, fuzzy control strategies have appeared in the control literature for over 20 years. Early process control applications consisted of demonstrations that fuzzy control could be used to control simple laboratory apparatus. In recent years, the success of fuzzy control in Japan, especially in consumer products such as washing machines and camcorders, has generated a new wave of interest. Industrial applications of fuzzy control to process control problems have begun to appear more frequently in Japan and Europe than in the U. S. But even in Japan, a survey has indicated that MFC has been more widely used in the process industries than any of the three AI techniques considered in this section.

WORDS AND PHRASES

camcorder       便携式摄像机

codify           编撰成册，

collaborative      协作的，合作的

demonstration    示范，展示

diagnosis         诊断

emulate        仿效

enthusiasm       热心，狂热

expertise       专门技术，专家

fuzzy         模糊的

game-playing     博弈，对策

intrigue         吸引，激起……的兴趣

membrane      膜，薄膜

physician      医师

physiological     生理的

psychological     心理的

semantics     语义学

silicon neuron   硅神经元

sympto症状，征兆

NOTES

[1] With the growth of modern science, the search for AI has taken two major directions: psychological and physiological research into the nature of human thought, and the technological development of increasingly sophisticated computing systems.

随着现代科学的发展，对于AI两个主要方面展开探索：对于人类思维本质的心理学和生理学研究，以及对日益高级的计算机系统的开发。

[2] Essential to such programs is the ability of the system to generate grammatically correct sentences and to establish linkages between words, ideas, and associations with other ideas.

这种程序的本质就是要求系统能够生成语法正确的句子，能够在单词、意思，以及与其它意思的联想之间建立联系。

[3] Research has shown that whereas the logic of language structure — its syntax — submits to programming, the problem of meaning, or semantics, lies far deeper, in the direction of true AI.

研究表明，虽然语言结构的逻辑（句法）服从于编程，但解释的问题（语义学）与真正人工智能的方向还相距甚远。

[4] Expert systems take computers a step beyond straightforward programming, being based on a technique called rule-based inference, in which pre-established rule systems are used to process the data.

专家系统把计算机从直接编程提高了一步，根据所谓的基于规则的推理方法，利用预先建立的规则系统来处理数据。

How we're teaching computers to understand pictures