Lecture No. 13. Expert systems and knowledge production model

1. Appointment of expert systems

To get acquainted with such a new concept for us as expert systems we, for starters, will go through the history of the creation and development of the "expert systems" direction, and then we will define the very concept of expert systems.

In the early 80s. XNUMXth century in research on the creation of artificial intelligence, a new independent direction has been formed, called expert systems. The purpose of this new research on expert systems is to develop special programs designed to solve specific types of problems. What is this special kind of problem that required the creation of a whole new knowledge engineering? This special type of tasks can include tasks from absolutely any subject area. The main thing that distinguishes them from ordinary problems is that it seems to be a very difficult task for a human expert to solve them. Then the first so-called expert system (where the role of an expert was no longer a person, but a machine), and the expert system receives results that are not inferior in quality and efficiency to the solutions obtained by an ordinary person - an expert. The results of expert systems can be explained to the user at a very high level. This quality of expert systems is ensured by their ability to reason about their own knowledge and conclusions. Expert systems may well replenish their own knowledge in the process of interaction with an expert. Thus, they can be put with full confidence on a par with a fully formed artificial intelligence.

Researchers in the field of expert systems for the name of their discipline often also use the previously mentioned term "knowledge engineering", introduced by the German scientist E. Feigenbaum as "bringing the principles and tools of research from the field of artificial intelligence into solving difficult applied problems that require expert knowledge."

However, commercial success to the development firms did not come immediately. For a quarter of a century from 1960 to 1985. The successes of artificial intelligence have been mainly related to research developments. However, starting around 1985, and on a massive scale from 1987 to 1990. expert systems have been actively used in commercial applications.

The merits of expert systems are quite large and are as follows:

1) expert systems technology significantly expands the range of practically significant tasks solved on personal computers, the solution of which brings significant economic benefits and greatly simplifies all related processes;

2) expert systems technology is one of the most important tools in solving the global problems of traditional programming, such as the duration, quality and, consequently, the high cost of developing complex applications, as a result of which the economic effect was significantly reduced;

3) there is a high cost of operation and maintenance of complex systems, which often exceeds the cost of the development itself by several times, as well as a low level of reusability of programs, etc.;

4) the combination of expert systems technology with traditional programming technology adds new qualities to software products due, firstly, to the provision of dynamic modification of applications by an ordinary user, and not by a programmer; secondly, greater "transparency" of the application, better graphics, interface and interaction of expert systems.

According to ordinary users and leading experts, in the near future, expert systems will find the following applications:

1) expert systems will play a leading role at all stages of design, development, production, distribution, debugging, control and service delivery;

2) expert systems technology, which has received wide commercial distribution, will provide a revolutionary breakthrough in the integration of applications from ready-made intelligent-interacting modules.

In general, expert systems are designed for the so-called informal tasks, i.e., expert systems do not reject and do not replace the traditional approach to program development focused on solving formalized problems, but complement them, thereby significantly expanding the possibilities. This is exactly what a mere human expert cannot do.

Such complex non-formalized tasks are characterized by:

1) fallacy, inaccuracy, ambiguity, as well as incompleteness and inconsistency of the source data;

2) fallacy, ambiguity, inaccuracy, incompleteness and inconsistency of knowledge about the problem area and the problem being solved;

3) large dimension of the space of solutions of a specific problem;

4) dynamic variability of data and knowledge directly in the process of solving such an informal problem.

Expert systems are mainly based on the heuristic search for a solution, and not on the execution of a known algorithm. This is one of the main advantages of expert systems technology over the traditional approach to software development. This is what allows them to cope so well with the tasks assigned to them.

Expert systems technology is used to solve a variety of problems. We list the main of these tasks.

1. Interpretation.

Expert systems that perform interpretation most often use the readings of various instruments to describe the state of affairs.

Interpretive expert systems are capable of processing a variety of types of information. An example is the use of spectral analysis data and changes in the characteristics of substances to determine their composition and properties. Also an example is the interpretation of the readings of measuring instruments in the boiler room to describe the state of the boilers and the water in them.

Interpretive systems most often deal directly with indications. In this regard, difficulties arise that other types of systems do not have. What are these difficulties? These difficulties arise due to the fact that expert systems have to interpret clogged superfluous, incomplete, unreliable or incorrect information. Hence, either errors or a significant increase in data processing are inevitable.

2. Prediction.

Expert systems that make a prediction of something determine the probabilistic conditions of given situations. Examples are the forecast of damage caused to the grain harvest by adverse weather conditions, the assessment of demand for gas in the world market, weather forecasting according to meteorological stations. Forecasting systems sometimes use modeling, i.e., programs that display some relationships in the real world in order to recreate them in a programming environment, and then design situations that may arise with certain initial data.

3. Diagnostics of various devices.

Expert systems perform such diagnostics by using descriptions of any situation, behavior, or data on the structure of various components in order to determine the possible causes of a malfunctioning diagnosable system. Examples are the establishment of the circumstances of the disease by the symptoms that are observed in patients (in medicine); identification of faults in electronic circuits and identification of faulty components in the mechanisms of various devices. Diagnostic systems are quite often assistants that not only make a diagnosis, but also help in troubleshooting. In such cases, these systems may well interact with the user to assist in troubleshooting and then provide a list of actions required to resolve them. Currently, many diagnostic systems are being developed as applications to engineering and computer systems.

4. Planning various events.

Expert systems designed for planning design various operations. Systems predetermine an almost complete sequence of actions before their implementation begins.

Examples of such planning of events are the creation of plans for military operations, both defensive and offensive, predetermined for a certain period in order to gain an advantage over enemy forces.

5. Design.

Expert systems that perform design develop various forms of objects, taking into account the prevailing circumstances and all related factors.

An example is genetic engineering.

6. Control.

Expert systems that exercise control compare the present behavior of the system with its expected behavior. Observing expert systems detect controlled behavior that confirms their expectations versus normal behavior or their assumption of potential deviations. Controlling expert systems, by their very nature, must work in real time and implement a time-dependent and context-dependent interpretation of the behavior of the controlled object.

Examples include monitoring the readings of measuring instruments in nuclear reactors in order to detect emergencies or evaluating diagnostic data from patients in the intensive care unit.

7. Management.

After all, it is widely known that expert systems that exercise control, very effectively manage the behavior of the system as a whole. An example is the management of various industries, as well as the distribution of computer systems. Control expert systems must include observing components in order to control the behavior of an object over a long period of time, but they may also need other components from the types of tasks already analyzed.

Expert systems are used in various fields: financial transactions, oil and gas industry. Expert systems technology can also be applied in energy, transportation, pharmaceutical industry, space development, metallurgical and mining industries, chemistry and many other areas.

2. Structure of expert systems

The development of expert systems has a number of significant differences from the development of a conventional software product. The experience of creating expert systems has shown that the use of the methodology adopted in traditional programming in their development either greatly increases the amount of time spent on creating expert systems, or even leads to a negative result.

Expert systems are generally divided into static и dynamic.

First, consider a static expert system.

Standard static expert system consists of the following main components:

1) working memory, also called the database;

2) knowledge bases;

3) a solver, also called an interpreter;

4) components of knowledge acquisition;

5) explanatory component;

6) dialog component.

Let's now consider each component in more detail.

working memory (by absolute analogy with the working, i.e., computer RAM) is designed to receive and store the initial and intermediate data of the task being solved at the current moment.

Knowledge base is intended for storing long-term data describing a specific subject area, and rules describing the rational transformation of data in this area of ​​the problem being solved.

Solveralso called interpreter, functions as follows: using the initial data from the working memory and long-term data from the knowledge base, it forms the rules, the application of which to the initial data leads to the solution of the problem. In a word, he really "solves" the problem set before him;

Knowledge Acquisition Component automates the process of filling the expert system with expert knowledge, i.e. this component provides the knowledge base with all the necessary information from this particular subject area.

Explain component explains how the system obtained a solution to this problem, or why it did not receive this solution, and what knowledge it used in doing so. In other words, the explain component generates a progress report.

This component is very important in the entire expert system, since it greatly facilitates the testing of the system by an expert, and also increases the user's confidence in the result obtained and, therefore, speeds up the development process.

Dialog Component serves to provide a friendly user interface both in the course of solving a problem and in the process of acquiring knowledge and declaring the results of work.

Now that we know what components a statistical expert system generally consists of, let's build a diagram that reflects the structure of such an expert system. It looks like this:

Static expert systems are most often used in technical applications where it is possible not to take into account changes in the environment that occur during the solution of a problem. It is curious to know that the first expert systems that received practical application were precisely static.

So, on this we will finish the consideration of the statistical expert system for now, let's move on to the analysis of the dynamic expert system.

Unfortunately, the program of our course does not include a detailed consideration of this expert system, so we will limit ourselves to analyzing only the most basic differences between a dynamic expert system and a static one.

Unlike a static expert system, the structure dynamic expert system In addition, the following two components are introduced:

1) a subsystem for modeling the outside world;

2) a subsystem of relations with the external environment.

Subsystem of relations with the external environment it just makes connections with the outside world. She does this through a system of special sensors and controllers.

In addition, some traditional components of a static expert system undergo significant changes in order to reflect the temporal logic of events currently occurring in the environment.

This is the main difference between static and dynamic expert systems.

An example of a dynamic expert system is the management of the production of various medicines in the pharmaceutical industry.

3. Participants in the development of expert systems

Representatives of various specialties are involved in the development of expert systems. Most often, a specific expert system is developed by three specialists. This is usually:

1) expert;

2) knowledge engineer;

3) a programmer for the development of tools.

Let us explain the responsibilities of each of the specialists listed here.

Expert - this is a specialist in the subject area, the tasks of which will be solved with the help of this particular expert system being developed.

Knowledge Engineer is a specialist in the development of an expert system directly. The technologies and methods used by him are called knowledge engineering technologies and methods. A knowledge engineer helps an expert to identify from all the information in the subject area the information that is necessary to work with a particular expert system being developed, and then structure it.

It is curious that the absence of knowledge engineers among the participants in the development, that is, their replacement by programmers, either leads to the failure of the entire project of creating a specific expert system, or significantly increases the time for its development.

Finally, programmer develops tools (if tools are newly developed) designed to accelerate the development of expert systems. These tools contain, in the limit, all the main components of an expert system; the programmer also interfaces his tools with the environment in which it will be used.

4. Operating modes of expert systems

The expert system operates in two main modes:

1) in the mode of acquiring knowledge;

2) in the mode of solving the problem (also called the mode of consultations, or the mode of using the expert system).

This is logical and understandable, because at first it is necessary, as it were, to load the expert system with information from the subject area in which it is to work, this is the "training" mode of the expert system, the mode when it receives knowledge. And after loading all the information necessary for the work, the work itself follows. The expert system becomes ready for operation, and it can now be used for consultations or for solving any problems.

Let's consider in more detail knowledge acquisition mode.

In the mode of acquiring knowledge, work with the expert system is carried out by an expert with the mediation of a knowledge engineer. In this mode, the expert, using the knowledge acquisition component, fills the system with knowledge (data), which, in turn, allows the system to solve problems from this subject area in the solution mode without the participation of an expert.

It should be noted that the knowledge acquisition mode in the traditional approach to program development corresponds to the stages of algorithmization, programming and debugging performed directly by the programmer. It follows that, in contrast to the traditional approach, in the case of expert systems, the development of programs is carried out not by a programmer, but by an expert, of course, with the help of expert systems, i.e., by and large, a person who does not know programming.

And now let's consider the second mode of functioning of the expert system, i.e.

problem solving mode.

In the problem solving mode (or the so-called consultation mode), communication with expert systems is carried out directly by the end user, who is interested in the final result of the work and sometimes the method of obtaining it. It should be noted that depending on the purpose of the expert system, the user does not have to be an expert in this problem area. In this case, he turns to expert systems for the result, not having sufficient knowledge to obtain results. Or, the user may still have a level of knowledge sufficient to achieve the desired result on their own. In this case, the user can get the result himself, but turns to expert systems in order to either speed up the process of obtaining the result, or to assign monotonous work to the expert systems. In consultation mode, data about the user's task, after being processed by the dialog component, enters the working memory. Based on input data from working memory, general data about the problem area, and rules from the database, the solver generates a solution to the problem. When solving a problem, expert systems not only execute the prescribed sequence of a specific operation, but also preliminarily form it. This is done for the case when the reaction of the system is not entirely clear to the user. In this situation, the user may require an explanation of why this or that expert system asks this or that question or why this expert system cannot perform this operation, how this or that result supplied by this expert system is obtained.

5. Production model of knowledge

At its core, production models of knowledge close to logical models, which allows you to organize very effective procedures for logical data inference. This is on the one hand. However, on the other hand, if we consider production models of knowledge in comparison with logical models, then the former more clearly display knowledge, which is an indisputable advantage. Therefore, undoubtedly, the production model of knowledge is one of the main means of representing knowledge in artificial intelligence systems.

So, let's start a detailed consideration of the concept of a production model of knowledge.

The traditional production model of knowledge includes the following basic components:

1) a set of rules (or productions) representing the knowledge base of the production system;

2) working memory, which stores the original facts, as well as facts derived from the original facts using the inference mechanism;

3) the logical inference mechanism itself, which allows deriving new facts from the existing facts, according to the existing inference rules.

And, curiously, the number of such operations can be infinite.

Each rule representing the knowledge base of the production system contains a conditional and a final part. The conditional part of the rule contains either a single fact or several facts connected by a conjunction. The final part of the rule contains facts that need to be replenished with working memory if the conditional part of the rule is true.

If we try to schematically depict the production model of knowledge, then the production is understood as an expression of the following form:

(i) Q; P; A→B; N;

Here i is the name of the knowledge production model or its serial number, with the help of which this production is distinguished from the entire set of production models, receiving some kind of identification. Some lexical unit reflecting the essence of this product can act as a name. In fact, we name products for better perception by consciousness, in order to simplify the search for the desired product from the list.

Let's take a simple example: buying a notebook" or "a set of colored pencils. Obviously, each product is usually referred to by words suitable for the moment. In other words, call a spade a spade.

Move on. The Q element characterizes the scope of this particular knowledge production model. Such spheres are easily distinguished in the human mind, therefore, as a rule, there are no difficulties with the definition of this element. Let's take an example.

Consider the following situation: let's say that in one area of ​​our consciousness the knowledge of how to cook food is stored, in another, how to get to work, in the third, how to properly operate the washing machine. A similar division is also present in the memory of the production model of knowledge. This division of knowledge into separate areas allows you to significantly save time spent on searching for some specific production models of knowledge that are currently needed, and thereby greatly simplifies the process of working with them.

Of course, the main element of the product is its so-called core, which in our above formula was denoted as A → B. This formula can be interpreted as "if condition A is met, then action B should be performed."

If we are dealing with more complex kernel constructs, then the following alternative choice is allowed on the right side: "if condition A is satisfied, then action B should be performed₁, otherwise you should perform action B₂".

However, the interpretation of the core of the production model of knowledge can be different and depend on what will be on the left and right of the sequent sign "→". With one of the interpretations of the core of the production model of knowledge, the sequent can be interpreted in the usual logical sense, i.e. as a sign of the logical consequence of the action B from the true condition A.

Nevertheless, other interpretations of the core of the knowledge production model are also possible. So, for example, A can describe some condition, the fulfillment of which is necessary in order for some action B to be performed.

Next, we consider an element of the production model of knowledge R.

Element Р is defined as a condition for the applicability of the product core. If condition P is true, then the production core is activated. Otherwise, if the condition P is not satisfied, i.e. it is false, the core cannot be activated.

As an illustrative example, consider the following knowledge production model:

"Availability of money"; "If you want to buy thing A, then you should pay its cost to the cashier and present the check to the seller."

We look, if the condition P is true, that is, the purchase is paid and the check is presented, then the core is activated. Purchase completed. If in this production knowledge model the condition of applicability of the core is false, i.e. if there is no money, then it is impossible to apply the core of the knowledge production model, and it is not activated.

And finally go to the element N.

The element N is called the postcondition of the production data model. The postcondition defines the actions and procedures that must be performed after the implementation of the production core.

For a better perception, let's give a simple example: after buying a thing in a store, it is necessary to reduce the number of things of this type by one in the inventory of goods of this store, i.e. if the purchase is made (hence, the core is sold), then the store has one unit of this particular product less. Hence the postcondition "Cross out the unit of the purchased item".

Summing up, we can say that the representation of knowledge as a set of rules, i.e. through the use of a production model of knowledge, has the following advantages:

1) it is the ease of creating and understanding individual rules;

2) it is the simplicity of the logical choice mechanism.

However, in the representation of knowledge in the form of a set of rules, there are also disadvantages that still limit the scope and frequency of application of production knowledge models. The main such disadvantage is the ambiguity of the mutual relations between the rules that make up a specific production model of knowledge, as well as the rules of logical choice.

<< Back: Entity class relationships (Hierarchical recursive communication. Network recursive communication. Association. Generalizations. Composition. Aggregation. Attribute unification)

We recommend interesting articles Section Lecture notes, cheat sheets:

▪ General theory of statistics. Lecture summary

▪ Insurance. Lecture notes

▪ infectious diseases. Lecture notes

See other articles Section Lecture notes, cheat sheets.

Read and write useful comments on this article.

<< Back

Latest news of science and technology, new electronics:

The existence of an entropy rule for quantum entanglement has been proven 09.05.2024

Quantum mechanics continues to amaze us with its mysterious phenomena and unexpected discoveries. Recently, Bartosz Regula from the RIKEN Center for Quantum Computing and Ludovico Lamy from the University of Amsterdam presented a new discovery that concerns quantum entanglement and its relation to entropy. Quantum entanglement plays an important role in modern quantum information science and technology. However, the complexity of its structure makes understanding and managing it challenging. Regulus and Lamy's discovery shows that quantum entanglement follows an entropy rule similar to that for classical systems. This discovery opens new perspectives in the field of quantum information science and technology, deepening our understanding of quantum entanglement and its connection to thermodynamics. The results of the study indicate the possibility of reversibility of entanglement transformations, which could greatly simplify their use in various quantum technologies. Opening a new rule ... >>

Mini air conditioner Sony Reon Pocket 5 09.05.2024

Summer is a time for relaxation and travel, but often the heat can turn this time into an unbearable torment. Meet a new product from Sony - the Reon Pocket 5 mini-air conditioner, which promises to make summer more comfortable for its users. Sony has introduced a unique device - the Reon Pocket 5 mini-conditioner, which provides body cooling on hot days. With it, users can enjoy coolness anytime, anywhere by simply wearing it around their neck. This mini air conditioner is equipped with automatic adjustment of operating modes, as well as temperature and humidity sensors. Thanks to innovative technologies, Reon Pocket 5 adjusts its operation depending on the user's activity and environmental conditions. Users can easily adjust the temperature using a dedicated mobile app connected via Bluetooth. Additionally, specially designed T-shirts and shorts are available for convenience, to which a mini air conditioner can be attached. The device can oh ... >>

Energy from space for Starship 08.05.2024

Producing solar energy in space is becoming more feasible with the advent of new technologies and the development of space programs. The head of the startup Virtus Solis shared his vision of using SpaceX's Starship to create orbital power plants capable of powering the Earth. Startup Virtus Solis has unveiled an ambitious project to create orbital power plants using SpaceX's Starship. This idea could significantly change the field of solar energy production, making it more accessible and cheaper. The core of the startup's plan is to reduce the cost of launching satellites into space using Starship. This technological breakthrough is expected to make solar energy production in space more competitive with traditional energy sources. Virtual Solis plans to build large photovoltaic panels in orbit, using Starship to deliver the necessary equipment. However, one of the key challenges ... >>

Random news from the Archive Samsung Wireless: 1,5TB wireless storage, hotspot, offline charging 14.05.2014 Seagate has announced that its Samsung HDD division has begun shipping a new Samsung Wireless device to the market. This drive is aimed primarily at mobile users. And its main feature is the Wi-Fi hotspot function, which allows you to connect up to five users. Samsung Wireless is positioned as a portable accessory for smartphones and tablets. The novelty has a compact form factor and low weight. You can access the data on the drive wirelessly. Moreover, such streaming file transfer is configured quickly and easily. The manufacturer explains the relevance of his device by the fact that the built-in memory of modern mobile devices is often limited and does not allow you to carry a collection of movies in HD quality. The capacity of Samsung Wireless reaches 1,5 TB. Supported operating systems include Anroid 2.3 and above, Windows 8, 7, Vista, XP SP2, Ma OS X 10.6 and above. Built-in battery provides up to seven hours of continuous operation. The device can connect to wireless networks and provide Internet distribution. The USB 3.0 port allows you to use your Samsung Wireless as a portable battery to recharge your mobile devices. In addition, it can be used to connect the HDD directly to the computer without using Wi-Fi. The recommended retail price of new items in the world market is $179.

Other interesting news:

▪ Western Digital iNAND 7350 256GB Smartphone Drive

▪ New ISOCELL sensors from Samsung

▪ Chemists against global warming

▪ Birds avoid radioactivity

▪ Mushrooms found to extract gold from the soil

News feed of science and technology, new electronics

Interesting materials of the Free Technical Library:

▪ section of the site Personal transport: land, water, air. Article selection

▪ article by Erich Seligmann Fromm. Famous aphorisms

▪ article Which club won the main trophy after removing the word mighty from the name? Detailed answer

▪ article Chemical analysis laboratory assistant. Standard instruction on labor protection

▪ article LED phase indicator. Encyclopedia of radio electronics and electrical engineering

▪ article Standards for testing electrical equipment and devices for electrical installations of consumers. Conduction currents of surge arresters at alternating voltage frequency 50 Hz. Encyclopedia of radio electronics and electrical engineering

Leave your comment on this article:

All languages ​​of this page

Arabic	Hebrew	Polish
Bengali	Hindi	Portuguese
Chinese	Italian	Spanish
English	Japanese	Turkish
French	Korean	Ukrainian
German	Malay	Vietnamese

Home page | Library | Articles | Website map | Site Reviews

www.diagram.com.ua
2000-2024