Expert systems technology
Creation and use of expert systems is one of the conceptual stages of information technology development. The intellectual solution of problems in a certain subject area is based on the principle of reproduction of knowledge of experienced specialists – experts.
Based on their own experience, the expert analyses the situation and recognizes the most useful information, optimizes decision-making, cutting off dead-end paths.
An expert system is a set of methods and means of organizing, accumulating and applying knowledge to solve complex problems in a certain subject area. The expert system achieves higher efficiency by searching a large number of alternatives when choosing a solution, relying on the high quality experience of a group of specialists, analyzes the impact of a large volume of new factors, evaluating them in the construction of strategies, adding the possibility of forecasting.
The basis of the expert system is a set of knowledge (knowledge bases) structured in order to formalize the decision making process.
Expert systems are designed with training in mind and are capable of justifying the logic of choice of a solution, i.e. they possess the properties of adaptability and its argumentation. Most expert systems have an explanation mechanism.
Advantages of expert systems in comparison with the use of experienced specialists are as follows:
- The competence achieved is not lost, can be documented, transferred, reproduced and built upon;
- there are more sustainable results, no emotional or other factors of human unreliability;
- the high cost of development is balanced by the low cost of operation, the ability to copy, and collectively, they are cheaper than highly skilled professionals.
The disadvantage of expert systems, which is characteristic of their current state, is less adaptability to learning new rules and concepts, to creativity and invention. The use of expert systems makes it possible in many cases to give up highly qualified specialists, but assumes leaving a place in the system for a lower qualified expert. Expert systems serve as a means to expand and strengthen the professional capabilities of the end user.
The expert system should demonstrate competence, i.e., reach the same level in a particular subject area as the expert. It is not enough to find good solutions, it has to be done quickly. The systems should have not only a deep, but also a broad enough understanding of the subject matter.
Methods to find solutions to problems are achieved on the basis of reasoning based on fundamental principles in the case of incorrect data or incomplete sets of rules. Such properties are least developed in computer expert systems, but they are peculiar to high level specialists.
Distinctions of expert systems from usual computer systems are:
- expert systems manipulate knowledge, while any other systems manipulate data;
- expert systems tend to provide effective optimal solutions and are sometimes able to make mistakes, but unlike traditional computer systems they have the potential to learn from their mistakes.
Some of the subject areas in which expert systems are used are listed. Of these, medicine is particularly popular.
Areas of application of expert systems Military Geology Engineering Informatics Computer systems Computer systems Space technology Mathematics Medicine.
Expert systems are the most vulnerable in recognizing the limits of their capabilities and demonstrate unreliable functioning near the limits of their applicability. Further progress in the field of artificial intelligence over time will offer ways to identify the limits of their capabilities.
Another disadvantage of expert systems is the significant labor required to supplement the knowledge base. Obtaining knowledge from experts and adding it to the knowledge base is a complex process, involving considerable time and cost. The design of expert systems also has certain difficulties and limitations that affect their development.
Foreign experience shows that expert systems are developed mainly in universities, research centres and commercial organizations, including the financial industry. In the field of financial services, these systems help insurance companies analyze and assess commercial risk, set the size of loans when lending to organizations, make estimates of projects, etc.
The scope of expert systems is expanding. In addition to covering different areas, one of the most important consequences of developing expert systems is the modification of knowledge. As developers build large, complex knowledge bases, a knowledge market emerges that is independent of computer systems.
Learning tools will be available for those studying a particular application area. A commercial product will be meta-appropriation, i.e. knowledge about optimal strategies and procedures for using subject knowledge.
The development of expert systems into intellectual ones consists in merging concepts of equipment, means of their creation (languages) and expert systems themselves. Unification of intellectual systems is particularly effective in complex infrastructures. Intellectual systems are already being developed and implemented abroad for commercial use.
Artificial competence of expert systems does not completely replace a person. An expert person is able to reorganize information and knowledge and use it to synthesize new knowledge. In the field of creative activity, people have greater abilities and capabilities than the smartest systems.
Experts cope with unexpected turns of events and, using new approaches, are able to draw analogies from other subject areas. Experts adapt to changing conditions and adapt their strategies to new circumstances in a wider range of problems and challenges.
Expert systems are less adapted to learning at the level of new concepts and new rules. They are less effective and less useful when the complexity of real-world problems has to be taken into account.
Experts can directly perceive the full range of input information: symbolic, visual, graphic, textual, sound, tactile, olfactory. The expert system has only symbols, with the help of which the knowledge bases embodying certain concepts are presented. The transformation of sensory information into symbolic information is accompanied by the loss of some information.
But the main thing is that the huge volume of knowledge possessed by experts-experts (professional knowledge and knowledge of the world and the laws acting in it) cannot be built into an intellectual system yet, all the more so specialized, which is any expert system.
