The purpose of this article is to show us a broad and precise description of what Expert Systems (SE) are, which are also known as Knowledge-Based Systems. Likewise, we briefly show the origins, concepts, applications, advantages, limitations, and its basic architecture of this area and / or field of Artificial Intelligence. On the other hand, today the market becomes more competitive, where the administration and good management of information is fundamental for all companies and / or organizations, which if they want to survive must stay at the forefront in each and every one of the areas.
Introduction
Someone is considered an expert on a problem when this individual has specialized knowledge about the problem. In the area of (SE) this type of knowledge is called domain knowledge. The word domain is used to emphasize that knowledge pertains to a specific problem.
Before the advent of the computer, man wondered if the privilege of reasoning and thinking would be taken away from him. Currently there is a field within artificial intelligence to which this power is attributed: that of expert systems (SE). These systems are also known as Knowledge-Based Systems, which allow the creation of machines that reason like man, restricting themselves to a limited space of knowledge. In theory they can reason by following the steps that a human expert (doctor, analyst, businessman, etc.) would follow to solve a specific problem. This type of computer knowledge model offers a wide range of possibilities in problem solving and learning. Its use will be widely spread in the future,due to its significant impact on business and industry.
History of the (SE)
Its beginnings date back to the mid-sixties. During this decade, researchers Alan Newell and Herbert Simon developed a program called GPS (General Problem Solver; general problem solver). He could work with crypto arithmetic, with the Hanoi towers, and other similar problems. What GPS couldn't do was solve real-world problems, such as a medical diagnosis.
Some researchers then decided to completely change the focus of the problem by restricting their ambition to a specific domain and trying to simulate the reasoning of a human expert. Instead of dedicating themselves to computerizing general intelligence, they focused on very specific domains of knowledge. In this way the SE were born.
Beginning in 1965, a team led by Edward Feigenbaum began to develop SE using carefully defined knowledge bases. Two years later DENDRAL is built, which is considered the first SE. The fiction of said SE was to identify molecular chemical structures from its spectrographic analysis.
In the 1970s MYCIN was developed for consultation and diagnosis of blood infections. This system introduced new features: use of imprecise knowledge to reason and the ability to explain the reasoning process. The most important thing is that it worked correctly, giving conclusions analogous to what a human being would give after long years of experience. In MYCIN there are clearly differentiated inference engine and knowledge base. By separating these two parts, the inference engine can be considered in isolation. This results in an empty system or shell. This is how EMYCIN (MYCIN Esencial) arose with which SACON was built, used for engineering structures, PUFF to study lung function and GUIDON to choose therapeutic treatments.
At that time there were also developed: HERSAY, which tried to identify the spoken word, and PROSPECTOR, used to find mineral deposits.
The KAS (Knowledge Acquisition System) shell was derived from the latter.
In the eighties, SE became fashionable, numerous high-tech companies investigated in this area of artificial intelligence, developing SE for commercialization. It is concluded that the success of an SE depends almost exclusively on the quality of its knowledge base. The downside is that coding the expertise of a human expert can be difficult, time-consuming, and time-consuming.
An example of a modern SE is CASHVALUE, which evaluates investment projects and VATIA, which advises on value added tax or VAT.
Definitions of the (SE)
It is software that imitates the behavior of a human expert in solving a problem. They can store expert knowledge for a particular field and solve a problem by logical deduction of conclusions.
SE are those programs that are made by making the knowledge explicit in them, that have specific information about a specific domain and that perform a task related to this domain.
Programs that manipulate encoded knowledge to solve problems in a specialized domain in a domain that generally requires human expertise.
Programs that contain both declarative knowledge (facts about objects, events and / or situations) and control knowledge (information about the courses of an action), to emulate the reasoning process of human experts in a particular domain and / or area of expertise.
Software that incorporates expert knowledge on a given application domain, so that it is capable of solving problems of relative difficulty and supporting intelligent decision-making based on a symbolic reasoning process.
Applications
Its main applications are in business management because;
a) Almost all companies have a computer that performs the basic information processing functions: general accounting, financial decisions, treasury management, planning, etc.
b) This job involves handling large volumes of information and performing numerical operations to later make decisions. This creates an ideal terrain for the implementation of SEs.
In addition, SE are also applied in accounting in sections such as: Auditing (it is the field in which more SE applications are being carried out) Taxation, planning, financial analysis and financial accounting.
Application areas
SEs apply to a wide variety of fields and / or areas. Some of the main ones are listed below:
- MilitaryComputer ScienceTelecommunicationsChemistryLawAeronauticsGeologyArcheologyAgricultureElectronicsTransportEducationMedicineIndustryFinance and Management
Advantage
These programs provide the ability to work with large amounts of information, which are one of the great problems faced by the human analyst that can negatively affect decision-making since the human analyst can purify data that they do not consider relevant, while an SE is due At its high processing speed, it analyzes all the information, including the non-useful ones, in order to provide a more solid decision.
Limitations
It is evident that to update these programs need to be reprogrammed (perhaps this is one of their most accentuated limitations) another of their limitations may be the high cost in money and time, in addition to that these programs are not very flexible to changes and difficult to access unstructured information.
Due to the shortage of human experts in certain areas, SEs can store their knowledge for when it is necessary to apply it. Likewise, SEs can be used by non-specialized people to solve problems. Also, if a person frequently uses an SE, they will learn from it.
On the other hand, artificial intelligence has not been able to develop systems that are capable of solving problems in a general way, of applying common sense to solve complex situations, or of controlling ambiguous situations.
The future of the SE revolves around the head of each person, provided that the chosen field has the need and / or presence of an expert to obtain any type of benefit.
Basic architecture of expert systems
- Knowledge base. It is the part of the expert system that contains the knowledge about the domain. Get the expert's knowledge and code it in the knowledge base. A classic way of representing knowledge in an expert system is the rules. A rule is a conditional structure that logically relates the information contained in the antecedent part with other information contained in the consequent part. Fact base (Working memory). Contains the facts about a problem that were discovered during a query. During a consultation with the expert system, the user enters the current problem information into the fact base. The system matches this information with the knowledge available in the knowledge base to deduce new facts. Inference engine. The expert system models the human reasoning process with a module known as the inference engine. This inference engine works with the information contained in the knowledge base and the fact base to deduce new facts. Contrast the particular facts in the fact base with the knowledge in the knowledge base to draw conclusions about the problem. Explain subsystem. A characteristic of expert systems is their ability to explain their reasoning. Using the explain subsystem module, an expert system can provide an explanation to the user of why they are asking a question and how they have reached a conclusion. This module provides benefits to both the system designer and the user. The designer can use it to detect errors and the user benefits from the transparency of the system. User interface.The interaction between an expert system and a user is done in natural language. It is also highly interactive and follows the pattern of conversation between human beings. To conduct this process in a user-acceptable manner, the design of the user interface is especially important. A basic requirement of the interface is the ability to ask questions. To obtain reliable information from the user, special care must be taken in the design of the questions. This may require designing the interface using menus or graphics.
Conclusions
Actualmente el duro, difícil y cambiante mercado competitivo se vuelve más complejo por la gran diversidad de información que se ven obligados a almacenar y analizar, razón por la cual las empresas se ven en la necesidad de recurrir a poderosas y/o robustas herramientas o sistemas que les sirvan de soporte a la hora de tomar decisiones. De esta forma estos inteligentes, precisos y eficientes sistemas son adoptados por más organizaciones, en las cuales se convierten y/o transforman en una importante estrategia de negocio.
Por otra parte es importante mencionar que estos seguirán siendo usados en los todos y cada una de las áreas y/o campos donde los expertos humanos sean escasos. Por consecuencia de lo anterior estos sistemas son utilizados por personas no especializadas, por lo cual el uso frecuente de los (SE) les produce y/o genera conocimiento a los usuarios.
Bibliografía
Viejo Hernando Diego (2003). Sistemas expertos.
Samper Márquez Juan José (2004). Introducción a los sistemas expertos.
Samper Juan (2003). Sistemas expertos. El conocimiento al poder
Criado Briz José Mario (2002). Introducción a los sistemas expertos
Wikipedia (2004). Sistema experto.
Castro Marcel (2002). Sistemas expertos.
Félix Justo (2004). Aplicaciones, ventajas y limitaciones de los sistemas expertos.
Montes Cerra Maria Clara (2003). Sistemas expertos.
