Lecture
Это окончание невероятной информации про словарь по искусственному интеллекту.
...
operating in parallel, each of which executes the same command flow on its data. SIMD architecture is often used in matrix processors.
1. Ways of combining words into word combinations and simple sentences and simple sentences into complex ones.
2. The section of linguistics, which studies ways of combining words and sentences, together with morphology, constituting a grammar.
3. The syntax of the text is the same as the discourse scheme or the non-productive (narrative) scheme; text structure.
Automatic construction of programs on the initial conditions of the problem. Formally, the problem of S.P.A. can be put as follows: for a given value of x, satisfying the predicate P (x), calculate the value of y, satisfying the predicate R (x, y). Here x, y are finite sets of program input and output variables. The mapping Q =
must contain enough information for the synthesis of the desired program. There are three approaches: inductive program synthesis, deductive program synthesis, or transformational program synthesis.
Automatic program synthesis using the formal method of program construction; output of a program from a given specification is considered as a problem of constructive proof of the existence of the desired solution. During the implementation of C.P.D. The methods of automatic proof of theorems in the calculus of first-order predicates are often used.
Automatic synthesis of programs based on finding common for a given set of examples of procedures leading to the solution of the problem. S.P.I. can be carried out at the syntactic level, when external features of programs are used, or at the semantic level, when the semantics of the program model is based on.
Automatic synthesis of programs, in which the design of a program is carried out by successively transforming the original specification into an equivalent resulting program. S.P.T. usually divided into the following stages: transition from a formal specification to a recursive version of a general type program; transformation of the general recursive version into a recursive version of a special type; transition from recursive version to iterative version.
See Text Generation.
A set of tools designed to assist the design engineer in creating new technical objects. CAD includes a large number of calculation modules, a knowledge base of documentation and technical drawing.
The system of reasoning, which is based on axioms. If axioms are logical (that is, identically true, regardless of the subject area), then S.A. matches the formal system. If among axioms there are axioms that are valid only in this subject area, then S.A. applicable only in this area. Often, SA, which has subject axioms, is called a quasiaxiomatic system.
A set of programs that allows you to implement a question-answer relationship for a pair of "user - computer." C.V.O. provides the search for relevant data and knowledge on the question entered into the system, which form the answer to the user. C.V.O. can use questions and answers formulated in natural language, limited natural language or in a special formal language of communication.
An axiomatic system in which the deduction theorem takes place, providing the derivation of the entire set of expressions, which is allowed by axioms and inference rules (and only by them). In this case, it is assumed that the deducible expressions possess an a priori property (for example, the identical truth or truth with a likelihood estimate of no less than specified).
The set of procedures used in the systems of expert systems explanation. S.D. used in order to increase the level of user confidence in the results obtained by the expert system. For this sd gives the user a number of additional arguments in favor of the decision that was received by the system, provided that these arguments were not directly used in the process of obtaining the decision.
An intellectual system capable of perceiving texts or speech in a limited natural language, understanding them and functioning in accordance with the information contained in the text. If necessary, S.YE.Ya. can generate response messages associated with the text entered into them. (See also interface natural language.)
A model that uses inductive logic or inductive generalization (or both).
Technical or software system capable of solving problems, traditionally considered creative, belonging to a particular subject area, knowledge of which is stored in the memory of S.I. Structure S.I. It includes three main blocks - a knowledge base, a solver and an intelligent interface.
The intellectual system for teaching a person to any kind of activity or used in the process of teaching schoolchildren or students. S.AND.O. includes a natural-language interface, graphical communication tools, a knowledge base and a special unit that controls the process of presenting the material, analyzes the learner and makes decisions about assessing the degree of learning.
The intellectual system in which there is a set of funds to supplement its knowledge base. There are two ways to learn: with a teacher and without a teacher. In the first case C.AND.O. должна иметь возможность пополнять свою базу знаний информацией, сообщаемой ей учителем, корректировать новую информацию с ранее имевшейся и задавать учителю вопросы, когда возникают конфликтные ситуации. При обучении без учителя С.И.О. обобщает информацию, которая содержится в ее базе данных или наблюдается во внешней среде. С помощью индуктивного вывода и, возможно, рассуждений по аналогии и рассуждений по ассоциации С.И.О. формирует гипотезы о новых закономерностях в той предметной области, в которой она работает. Этим новымзнаниям система может приписывать оценки правдоподобия.
Комплекс программных или аппаратно-программных средств, обеспечивающих взаимодействие пользователя с системой, решающей его задачу в процессе поиска решения.
Термин объясняется в статье Система аксиоматическая.
Вычислительная система, состоящая из большого числа процессорных модулей (элементов), в совокупности решающих сложную задачу. На организацию работы С.М. существенное влияние оказывает используемая соединительная сеть.
Система, способная к накоплению новых знаний в результате обращения к учителю или обобщения наблюдаемых фактов. (См. Система интеллектуальная обучающаяся.)
The part of the expert system designed to explain to the user, at his request, the way in which the decision was received. C.O. gives answers to questions of three types. "How-questions" make the system explain the solution. "Why-questions" cause S.O. the need for an explanation of why this particular decision was received, and not any other. “What-Questions” require from S.О. issuing all the information available in the expert system regarding the object or phenomenon to which the "What-Question" refers. For answers to such questions in S.O. There is a set of special procedures and a working memory in which, in the process of finding a solution, the information necessary to answer user questions is formed. From C.O. linked trust system, which increases the explanatory power of the expert system.
The intellectual system, the functioning of which is determined by the body of knowledge about the problem area in which it is used.
Intellectual system, in the knowledge base of which the knowledge representation in the form of production systems is accepted.
A set of tools for the automated filling of knowledge bases and maintaining them in working condition. Depending on the model of knowledge chosen in the intellectual system, they distinguish between CPS based on semantic networks, frames and production rules. S.P.Z. are part of all intelligent systems.
A collection of multiple products whose execution is specified using an output control strategy. The strategy is determined by the abstract machine M = 1 , M 2 , M 3 >, where M 1 forms the front of products for which the conditions of their applicability are satisfied; M 2 selects from the front of finished products the one that will be carried out at this stage of the process, M 3 makes the necessary changes to the products included in SP, on the basis of the information contained in the post-condition of the products made at this stage of the process.
See Production System.
The new generation of computers, which differs from those previously created by the presence of a natural language interface and means for the automatic synthesis of a program based on standard modules stored in the memory of the systems. Synthesis of the program is made according to the user, formulated in a professional natural language. Thus, in SPP.V. There is a built-in intellectual system, which makes it possible for users who do not have special programming knowledge to communicate with computers.
See Automated Control System.
The complex of tools that make up the database, by which search queries are made to the information stored in the database, the search procedures, service procedures related to preserving the integrity of the database, etc. are performed.
The set of software and hardware for the organization in the knowledge bases of procedures related to the search for knowledge, the replenishment of the knowledge base, its adjustment, etc.
See Automated Enterprise Management System.
See Automated Process Control System.
The model underlying many mathematical theories. S.F. is a four: a set of basic elements of the syntax rules of the axiom of the rule of inference.
Many frames connected by different relationships.
Intellectual system designed to provide advice to professionals working in a particular subject area. There are two types of C.E. Systems of the first type are designed for professionals whose professional level is not too high. The knowledge bases of such systems store knowledge obtained from top-class specialists. Systems of the second type are designed to help highly qualified specialists by performing for them a significant part of routine operations and viewing large amounts of information. The peculiarity of S.E. is the presence in them of a system of explanations that increases the consulting power of S.E.
An expert system whose knowledge base is empty. When using C.EP. in a certain subject area, the user must fill in the knowledge base either by himself with the help of special instructions for filling in the base, or by engaging a knowledge engineer. As experience has shown, S.E.P. can be effectively used only on the same subject areas, which are focused on adopted in S.E.P. ways of representing knowledge and reasoning. (See also Shell.)
Fact obtained in the process of logical inference in the formal system.
The basic structural unit of the frame. C. is a pair: (attribute (slot name) - value). Values may include constant facts, expressions containing variables, references to other C., etc. S. may have a structure, the elements of which are themselves slots. Often these smaller S. are called facets, cells, aspects, etc.
An information unit that is assigned a time interval in the knowledge base during which this unit exists.
The procedure for comparing a piece of knowledge or description with a sample reference. Part of the search procedure for the sample in the knowledge bases and is used in pattern recognition.
Knowledge of the subject area, which reflect the facts, patterns, properties and characteristics that are valid for any specific situations that may arise in this subject area.
Knowledge of the subject area, reflecting the facts, patterns, properties and characteristics typical for specific situations or classes of similar situations that may arise in this area.
A collection of attribute-value pairs.
Programming systems that simplify the creation of intelligent systems.
A way of linking two sentences in the text when the first sentence refers to a person, an object, a phenomenon, etc., and the second uses an anaphomical word (usually a pronoun), replacing the word from the first sentence. For example, in the text "The control system failed. It was not diagnosed during the" anaphoric word "she" in the second sentence indicates that it refers to the same control system, which was discussed in the first sentence. S.A. requires special procedures when analyzing texts in intelligent systems. S.A. are a special case of anaphor.
A set of rules by which the choice of rules of inference in the formal systems or the choice of products in the system of products is organized when searching for a solution. The most famous of S.U.V. in formal logical systems, this is direct inference and reverse inference. In production systems, much is known. (the principle of "blackboard", taking into account the longest condition of production, the principle of "agenda", etc.).
A structure that reflects the semantics of a natural language text. In S.G. there are no features of the morphological or syntactic structure of the natural language. S.G. the use of a translation from one language to another is used, and it also provides an understanding of texts in natural language, since SG The transition to knowledge representations that are used in the knowledge base is easily made.
A hypothetical structure through which knowledge about the world, possible actions in it and about the person is displayed and stored in a person’s memory. S.K. studied in cognitive psychology.
The architecture of new generations of computers formed by the same active elements or processors interconnected in the same type of way (for example, located on the surface of a torus with regular connection of each active element or processor with the four nearest neighbors). Theoretical models associated with the operation of S.O. studied in cellular automata theory.
The structure of the text that uses the deep cases (Fillmore cases). These cases do not depend on the grammatical cases of any natural language, but reflect the deep (role) interrelationships of the elements that are referred to in the text. For example, for the phrases "Ivan cuts wood" and "Wood is chopped by Ivan", the grammatical cases for the words "Ivan" and "tree" are different. But in both cases, the depth case for "Ivan" is the subject case, and for the "tree" - the object case. The number of depth cases for different authors ranges from a half dozen to thirty. S.P. is one of the varieties of the deep structure.
In artificial intelligence, a model in which the objects of reality are correlated with their internal representation in the memory of an intellectual system. S.P. does not depend on the peculiarities of the language in which the text is written, but is determined by the mechanisms that human consciousness has for perceiving the surrounding world and its comprehension.
An expression connecting the subject and the predicate with the help of the "is" bundle. C. is studied in logic.
Any object in the subject area.
See Integrity Restriction.
A semantic network in which causal relations or relations such as action-result, action-goal, instrument-action, etc. are used as relations
Direction in artificial intelligence, within the framework of which models and methods of generating on a computer analogs of the results of human creative activity are created. These include programs for composing music, poems, prose, computer graphics and painting, game programs, etc.
The statement that the "lambda" -calculus is equivalent to the intuitive understanding of the algorithm of T.Ch. similar to other theses (for example, the Turing thesis that the Turing machine is equivalent to an intuitive understanding of the algorithmic process), which allows to formalize the concept of an algorithm.
A logical theory in which all the initial axioms are identically true, and the rules of inference are authentic.
The theory, which is based on a formal system.
Linguistic section that describes how the speaker’s motives and goals are reflected in natural language texts. IN T.R.A. the extra-linguistic components (pause duration, tone, loudness of pronouncing, facial expressions and gestures) play a big role, which makes TP R.A. used when creating a communication model in intelligent systems.
A constant, variable, or expression of the form f (t 1 , t 2 , ... t n ) where f is a function symbol; t i - terms.
Combined under the same name on any basis data. T.D. used in databases.
Description of data in programming languages, allowing to create in the computer memory those data structures that are convenient for solving the problem.
The region on which the predicate values are defined. From this area, values are taken to denote the variables included in the predicate. For multiplace predicates, U. is a Cartesian product of the domains for each predicate argument.
A special area on which the impracticability of multiple clauses (clauses) is verified when implementing procedures for deducing the type of resolution method.
The term is explained in the article Unification.
The unifier r, which has the property that for any other unifier q for the same set of expressions, there exists a substitution t such that q = zt, where zt is the composition of the permutations r and t.
The procedure for substituting terms into two logical expressions instead of variables. Terms are selected in such a way that when they replace the variables of the same name in two expressions, both expressions become identical. The substitution itself is called a unifier. W. is used for logical inference in the resolution method.
A method of managing complex technical and organizational systems in which a classifier is built with the help of expert information, which allows you to break all the observed situations into classes and assign a single-step solution to each class or the first step in a multi-step solution. To describe the situations and classify them in the US A knowledge model similar to that in the semantic network is used.
See Slot.
In artificial intelligence, a body of knowledge that is activated when a message is formed or interpreted in natural language.
An expression of the form H (t 1 , t 2 , ..., t m ), where H is the m-local predicate, t i are terms.
An expression in a formal system in which either there are no variables, or the variables are bound quantifiers.
A closed formula that preserves the identity of truth with all interpretations.
An expression in a formal system that includes at least one variable that is not bound by a quantifier.
A set with a non-integer dimension to describe nonsmooth curves and surfaces. F. is used in computer graphics systems for defining complex graphic images and when zooming images. With the help of F. it is possible, when changing the scale, to remove and introduce elements of the image, depending on the selected scale.
A special form of knowledge representation, which is defined recursively. F. consists of a finite number of slots, each of which has a name and a value.
1. The frame acting as a sample when searching on the sample in the knowledge bases.
2. Synonym for frame instance.
A frame in which all slots have names that are deep cases.
A frame that has no constant values in part of the slots (or in all slots). F.P. describes knowledge of the subject area. When marking all slots. F.P. with constant values, it turns into a frame instance.
A prototype frame whose values for all slots are filled with constative information. The combination of F.E. forms an extensional database.
The term is explained in the article Automaton finite.
A function whose value characterizes the subject's belief in the truth of an event or fact.
The term is explained in the article Automaton finite.
The characteristic function for a fuzzy set, varying in the interval (0,1).
Functional operation symbol for eliminating existence quantifiers in logical expressions reduced to prefix normal form.
A sequence of formulas, starting with axioms in which each subsequent formula is derived based on the totality of the preceding elements of the chain.
The scale on which the metric is set, allowing you to count the distances from the absolute beginning.
The scale on which the metric is set, allowing to estimate the distance between the elements displayed on the scale.
A scale that uses pairs of words-antonyms ( sharp - dull, good - bad , etc.) at the ends of the scale. The average position of the scale is neutral. In addition, there are several intermediate divisions, as a rule, not denoted by any words. Sh.O. used in psychological experiments related to the identification of features of the individual psychosemantic space of the subjects. (See also Osgood Space.)
The metric scale in which distances are counted from any relative marker, for example, from an object currently observed.
The ordinal scale, on which the values of the linguistic variable or intervals, obtained from membership functions by cutoffs of the corresponding level, are located.
A scale that displays only the order relation between the elements located on the scale. PC. possesses properties inherent in metric scales.
A special scale for projecting blurry scales, in which the location of quantifiers is influenced by the specific semantics of the situations for which they are used. At W.W. these quantifiers are translated into other ones reflecting the general measure of fuzziness. W.W. allows you to compare with each other the statements related to different blurred scales.
A new generation of computers, whose architecture is based on a formal neuron model. Depending on the problem being solved, formal neurons are connected to each other with the help of transit cells. The advantage of such computers is the possibility of parallelization of the processes. Из формальных нейронов образуется однородная структура или неоднородная, на которой асинхронно протекает несколько процессов.
Прием решения задачи, основанный не на строгих математических моделях и алгоритмах, а на соображениях, восходящих к "здравому смыслу". Как правило, Э. отражает особенности того, как такие задачи решает человек, когда от не пользуется строго формальными приемами. Если эти человеческие способы решения удается запрограммировать, то такие программы называются эвристическими. Э. часто используются при программировании игр, имитации творческих процессов и т. п. В экспертных системах при формализации профессиональных знаний человека, касающихся способов решения задач в той или иной проблемной области, широко используются те Э., которыми руководствуются профессионалы-эксперты.
Множество цепочек символов, порождаемых с помощью автоматной грамматики.
Язык для обращения в базы данных и базы знаний за необходимой информацией. Формально Я.З. это счетное множество цепочек из символов некоторого конечного алфавита. На этом множестве цепочек выделено подмножество правильных цепочек или правильных запросов. Каждый правильный запрос имеет процедурную интерпретацию а операциях, разрешенных для информационных единиц, находящихся в базе данных или базе знаний.
Множество цепочек символов, порождаемых контекстно-свободной грамматикой.
The set of strings of characters generated by context-related grammar.
Способ описания моделей знаний в системах представления знаний. Today known Ya.P.Z. for the knowledge model in the form of frames (LISP, FRL, KRL, etc.), as well as a number of production languages.
The language of knowledge representation based on the first-order predicate calculus. Expressions Ya.P.Z.L. are the syntactically correct formulas of this calculus. In the form of such formulas, all declarative and procedural knowledge stored in the system is recorded. Dignity Ya.P.Z.L. their full formalizability and the presence of formal procedures for them, which allow to perform the derivation and analysis of such characteristics of records as consistency, equivalence, etc., is the disadvantage of Ya.P.L. is a poor visibility for the user of information units, written in the form of formulas of logical calculus.
The language of knowledge representation, the basic unit of which is products.
A language focused on the description of discrete automatic devices, in particular finite automata with memory and without memory. It allows you to build compact descriptions, especially for highly underdetermined automata. Description by Ya.S. consists of sequences, each of which is a special case of the product, in the left and right parts of which are Boolean functions. The meaning of the sequence is that the right-hand function takes the value 1 if the left-hand function equals 1, and is not defined if the left-hand function equals 0.
The language of knowledge representation and knowledge manipulation using frame representations as a knowledge model. The most famous Ya.F. are the FRL and KRL languages.
The object of research, about the internal structure of which is either partially known, or there are some hypotheses. Unlike the black box, the model Y.S. in addition to the links between reactions and external influences, they also take into account the partial information that is known about its internal structure.
The name of the object of study, introduced by U. R. Ashby, whose internal structure is unknown or not taken into account. Model Ya.Ch. based on his behavior i. reactions to the influences coming to the input from the outside, and characterizes the relationship between the reactions and the effects that caused them. Models Ya.Ch. theoretically justified in the direction of experimental psychology, called behaviorism, and are usually called models of "stimulus - reaction".
Часть 1 Dictionary of Artificial Intelligence
Часть 2 DENOTAT - Dictionary of Artificial Intelligence
Часть 3 Nemonotonnost with the withdrawal - Dictionary of Artificial Intelligence
Часть 4 SYNTAX - Dictionary of Artificial Intelligence
Comments
To leave a comment
Artificial Intelligence. Basics and history. Goals.
Terms: Artificial Intelligence. Basics and history. Goals.