WO2016163039A1

WO2016163039A1 - Autonomous knowledge extraction machine

Info

Publication number: WO2016163039A1
Application number: PCT/JP2015/073288
Authority: WO
Inventors: 洋彰宮崎
Original assignee: 洋彰宮崎
Priority date: 2015-04-09
Filing date: 2015-08-10
Publication date: 2016-10-13
Also published as: JP2015164064A; JP5810476B2

Abstract

To make it possible to autonomously generate, from input information, a common sense, a general way of thinking, a solution method for a problem similar to the input information. Regarding input information, the source of the information, the field, the theme, the subject, the modification of the subject, the predicate, the modification of the predicate, the modification relationship, and when, where, who, what, how, and why the thing is done are analyzed, and the sentence structure is organized and recorded. Information determined to be useful is autonomously recorded and restructured/evaluated as a knowledge system, and a process (record of information, update/upgrade of knowledge system, execution of instruction content, answer to question) corresponding to an evaluation result is autonomously carried out. Further, a part of words included in the information is generalized, a significant relationship between pieces of information successively input is extracted by strengthening a relationship between sets of patterns corresponding to the information, and common senses and general ways of thinking are autonomously constructed from the series of input information. Furthermore, a part of words included in the input information relating to a series of problems and solution methods thereof is generalized, and a solution method of a similar problem is autonomously generated.

Description

Autonomous knowledge extractor

This invention analyzes the source of information, fields, themes, subjects, subject modifications, predicates, predicate modifications, modification relationships, when, where, who, what, how, and why It organizes into sentence structure and records information judged to be useful autonomously to construct as a knowledge system. Also, the input information is compared with the internally constructed knowledge system and evaluated, and the processing according to the evaluation result (recording of information, updating and improvement of the knowledge system, execution of instruction contents, answer to questions) is autonomously performed Do. Furthermore, by generalizing a part of the words contained in the information, and by strengthening the relationship between the set of patterns corresponding to the information and the relationship between the information input sequentially and the set of patterns, the significant patterns are provided between the patterns. It relates to artificial intelligence and software that extracts relationships and autonomously builds common sense, general thinking and problem solving methods from a series of input information.

In the case where the machine processes input information, etc., it is realized by incorporating and executing a program created in advance in a program language in a computer mounted on the machine. A program is created so that the corresponding operation is executed when a preset condition is detected. If the condition detection and the corresponding operation are not appropriate, modify the program installed on the computer.
Source of information, source, field, subject, subject, subject modification, predicate, predicate modification, modification relationship, when, where, where, who, what, how, why and how to analyze information There is no artificial intelligence and software which organizes and records and autonomously records information determined to be useful as a knowledge system. In addition, the input information can be compared with the internally constructed knowledge system and evaluated, the common sense and general thinking can be autonomously constructed from a series of input information, and the input information on a series of problems and solutions There is no artificial intelligence and software that can autonomously generate solutions to similar problems.

Conventionally, when making a machine process input information, it is necessary to set a program in advance in a computer. It was necessary to create a program to judge the situation from the input information and a program to operate the machine according to each condition, and install and execute it on a computer installed in the machine. The program needs to be created in a dedicated programming language, and has the disadvantage of requiring a lot of time for development. Moreover, it was difficult to compare input information with the internally constructed knowledge system and evaluate it, and to construct common sense and general thinking autonomously from a series of input information.
In the present invention, a process of analyzing and recording information and the structure of information by human instruction and learning is performed, and the relation between information is constructed as a knowledge system by the connection relation between patterns and the process between patterns. Source of information, source, field, subject, subject, subject modification, predicate, predicate modification, modification relationship, when, where, where, who, what, how, why and how to analyze information Organize and record information that is judged useful, autonomously record it, build it as a knowledge system, compare it with the knowledge system built inside and evaluate it, process according to the evaluation result (information recording , Update and improve the body of knowledge, execute instructions, answer questions, and so on. In addition, generalize a part of the words included in the information, and extract the significant relationship between the information input sequentially and the information by strengthening the relationship between the set of patterns corresponding to the information and the set of patterns, Common sense and general thinking are constructed autonomously from a series of input information. Furthermore, artificial intelligence and software that autonomously generate solutions to similar problems from input information on a series of problems and solutions are realized.

In the present invention, instead of individually processing the processing of input information, a method of analyzing the relationship between information and information in a machine, a method of constructing a knowledge system and a method of solving problems, a method of generalizing input information, etc. It is implemented by learning the method of processing the input information.
In addition, the source of information, the subject, subject, subject, subject modification, predicate, predicate modification, modification relationship, and when, where, where, who, what, how, why and how to analyze information about input information To organize and record truths, truths, facts, expertise, rules and common sense as a body of knowledge.
In the early days, humans teach the methods of processing, and machines record and learn the methods taught by humans. To some extent, as learning progresses, machines will execute processing autonomously. A human confirms the processing result of the machine, notifies the machine if the process is incorrect, and corrects the process as appropriate.
Human thinking is expressed by language, but the information expressed by this language is converted into what is called a pattern. When converted to patterns, individual thoughts of human beings can be expressed as individual patterns, and changes in human thinking can be regarded as changes from patterns to patterns.
The pattern is not only a representation of the language, but it is also possible to express concepts like sentences and sentences. Also, the pattern can sequentially excite the associated pattern and cause the excited pattern to perform a number of processes. Furthermore, it is possible to generate a signal for driving a driving device in order to process and operate image information and information, and so on, which is a concept that the range that can be handled is very wide.

When we convert a language that expresses human thinking into patterns, and analyze patterns and relationships between patterns, we can identify sentence types (normal sentences, question sentences, imperative sentences, etc.), features (truths, truths, facts, definitions, rules, etc. Common sense, explanations, hypotheses, predictions, opinions, impressions) and relationships (causes and consequences, events and reasons, explanations and conclusions, outlines and details etc) can be identified. It is possible to make the machine learn the identification so that the type and feature of the sentence can be identified by humans. By analyzing the words contained in a sentence, the type of sentence can be identified as a plain sentence, an interrogative sentence, an imperative sentence, and the like. Among the features of sentences, human beings teach machines to learn about each of truth, truth, facts, definitions, rules and common sense. This can be implemented by identifying and recording each type of information (truth, truth, fact, definition, rule, common sense) as each information is input to the machine. The identification of the type of other information (explanation, hypothesis, prediction, opinion, impression) is that the type of sentence is not recorded as (truth, truth, fact, definition, rule, common sense) and the words included in the sentence ( It can be implemented by analyzing, thinking, thinking, etc.).
When the language is converted into a pattern and expressed as a set of patterns corresponding to the words constituting the sentence, and the semantic relationship between the words is expressed by the pattern-pattern connection relation, the meaning of the sentence is the pattern and the pattern It can be expressed in connection relation. Information input from word strings to word identification, word attributes (part-of-speech, meaning), analysis of sentence elements, sentence structure analysis (subject, predicate, subject modification, predicate modification, modification relationship), sentence elements and Relationship analysis of sentence elements (same meaning, definition, opposite meaning, etc.) is carried out, and the relation between information and information meaning is pattern and pattern by correlating the relation between sentence elements and sentence elements as pattern-pattern connection relation Convert to the connection relationship of The relationship between patterns and patterns can be set not only for sentence elements and sentence elements, but also for sentences and sentences, sentences and sentences. This can be implemented by defining a group of patterns as a new pattern. When a characteristic word indicating a pattern-pattern relationship (cause and result, event and reason, explanation and conclusion, outline and detail, etc.) is detected, the corresponding relationship between corresponding patterns is recorded autonomously. The relationship between the recorded patterns can be utilized in controlling the transition between the patterns.
A pattern-pattern connection relation can express various relations such as logical relation, definition, relation of attribution, similar relation, relation between action and result, development of inference, etc. Inheritance of attribute, common Inheritance of features and identification of individual features can also be expressed flexibly.
By defining the relationship between information and information as a connection relationship between patterns in this way, information is not recorded in a single state, but is recorded as a knowledge system having a relationship with other information. It becomes possible.
Although the relationship between information and information has various relationships, the relationship between information (the same meaning, definition, opposite meaning, similarity, logic, cause, result, detail, summary, summary, related information, etc.) By selecting an appropriate one in each processing phase (corresponding to the stage of the thinking process), it is possible to properly control the pattern transition from the pattern.

When information is input, the meaning is analyzed from the words contained in the information, and a pattern indicating the meaning equivalent to the information is generated. This pattern is generated while maintaining the subject, predicate, subject modification, predicate modification, and modification relationship of the sentence. The generated pattern is used to compare with the recorded pattern, and analysis is performed to determine whether or not the related pattern exists. Thus, since information can be searched at the semantic level, it is possible to search for content related information even if the words do not match.
The newly input information can be checked against the already recorded knowledge system to evaluate consistency and novelty. If it is not consistent with the information identified and recorded as truth, truth, facts, definitions, rules and common sense, the information is likely to be false. On the other hand, when it is not possible to judge whether the information is consistent or not, it is considered that the knowledge sufficient to judge the information is not stored yet, and the human is notified to make a decision. If human beings can judge that the information is correct, they shall notify the machine of the result. The machine records the information in the position of explanation or hypothesis, and identifies it as the information for which evaluation was conducted when it could be logically expanded by the combination of truth, truth, fact, definition, rule and common sense by obtaining other information in the future Will be recorded.
In this way, information is converted into patterns, information types, features are identified, classified and recorded, and information-information relationships (logical relationships, similar relationships, reciprocal relationships, analogies, various relationships, etc.) Information can be constructed as a knowledge system by expressing it as a pattern-pattern connection relation.

By constructing information as a knowledge system in this way, it becomes possible to solve problems using the knowledge system. First, analyze the problem using the knowledge system, clarify the problem and set the goal of the problem solution. Next, extract and apply a solution for solving the problem. The situation after application is evaluated, and if the goal of problem solving is reached, the process is completed. If the problem solving goal has not been reached, apply the above process to the new state again and bring it closer to the problem solving goal.
Analysis of problems and clarification of problems are implemented by learning how to detect and identify the difference between the target state and the current state. Moreover, the solution plan corresponding to each problem is also implemented by learning the corresponding processing. Learning is performed by recording the history of pattern transitions. The content taught by humans is recorded as a pattern-to-pattern transition.
As learning progresses, the taught process is autonomously executed as a pattern-to-pattern transition, and the taught content is executed.
Transition of human thinking pattern can be generally expressed as conditional processing from the viewpoint that the transition destination changes depending on conditions. In the present invention, the meaning of language is interpreted and converted autonomously as conditional processing. The condition of the conditional process is generated by generating a search pattern from the corresponding language and searching autonomously. It is determined whether the retrieved information satisfies the condition, and if it is satisfied, the corresponding processing is executed. Even if the problem solution used by human being as knowledge is directly input as a language, the meaning is sequentially interpreted, and the conditional processing is autonomously advanced to solve the problem. If new information is required during the process, the information request is notified, and when the corresponding information is acquired, the process according to the content is performed.
In the learning phase, when a pattern is excited, the history of the excited pattern is referred to in a set period before the pattern is excited, and the connection relationship with the excited pattern is strengthened. In the practical phase, the history of the excitation pattern in the set period is collated with the recorded connection relationship with other patterns recorded in each recording module, and when the correlation is large, the corresponding pattern is excited. When the pattern is excited, the history of the excitation pattern is updated, and a pattern having a large correlation with the history of the excitation pattern in a set period in a new state is sequentially excited.

As described above, in the learning phase, a series of patterns corresponding to processing are sequentially excited according to a human instruction, and the history is recorded, whereby the corresponding pattern is excited according to the instructed procedure. The behavior of the pattern is not static but shows dynamic behavior. The internal pattern recorded in the pattern can be used to retrieve the required information and store the search results in the required location. In addition, processing such as conversion of an internal pattern recorded in a pattern into a designated arrangement is also possible. A number of processes (formula processing, chemical formula processing, translation, etc.) can be performed by combining such patterns capable of dynamic behavior.
The teaching to the present autonomous knowledge extractor can be implemented by sequentially inputting linguistic information without programming. The input linguistic information is analyzed in relation to syntax, meaning, and information already recorded, and in accordance with the analysis result, the corresponding pattern is excited and processing is performed. A number of processes can be performed, such as evaluation and recording of input information, execution of instructed instructions, and generation of solutions to problems and problems.
The overall operation of the present autonomous knowledge extractor is managed by a pattern controller. In the transition cycle of each pattern, information input, information analysis (type of sentence, syntax, meaning etc.), information evaluation (newness, reliability, validity, usefulness etc.), information processing (problem / problem solution) Perform generation, recording, information output, etc.)

Next, generalization of input information will be described.
Perform feature extraction and generalization of part of the words included in the input sentence. For example, proper nouns appearing in a sentence are generalized as person A, person B, thing C, thing D, and so on. Sequentially, the sentence is converted into a pattern, and the converted pattern is excited. At this time, the generalized pattern is also excited sequentially. By inputting a large amount of information, the connection between specific patterns and patterns constituting a sentence is strengthened. The input sentence has a relation with the neighboring sentence. Since this relation is expressed by the relation between words constituting the sentence, the same relation is often expressed by the combination of identical words or similar words. Each time the sentence pattern is excited, the words contained in the sentence are excited, but by generalizing the proper noun etc., the generalized relationship between the person and the person or object is strengthened, and so on. Sentences appear more often than when expressed as proper nouns. The effect is to extract relationships between objects that do not depend on proper nouns. Moreover, since the relationship between the features of the words is extracted instead of the relationship between the individual words by the feature extraction of the words, generalization from the feature aspect becomes possible. This relationship can be detected as the connection between a particular pattern and the pattern is strengthened. In particular, the relationships strengthened from the many sentence examples have generality, and it is considered that those corresponding to common sense or general thinking are extracted. This common sense or general thinking depends on the sentences to be input. That is, when a sentence group corresponding to another culture is input, the common sense or general idea corresponding to the culture is extracted. When the text group in the same culture is inputted, since the thinking is considered to be equivalent, the thinking strengthened by many in the same culture will be extracted.
Similarly, by performing feature extraction and generalization of some words on a number of problems and their solutions, it is possible to extract generalized relations about the problems and their solutions. As learning by information input progresses, this machine will be able to generate solutions autonomously for similar problems.

In the following, an operation of recording information in a pattern recorder in a state where information is identified in a condition part and a processing part and arranged as a sentence structure will be described. By analyzing the words of the information, the conditional part and the processing part of the information can be identified. For example, in the sentences "A", "ga", "B", "at", "C" and "perform", ["A", "ga", "B" and "at the time]] are condition parts, and" [C "Do it"] is a processing unit. In the sentence "A""GA""B""NA""C""HA""D""is" [[A] "GA""B""NA""is the condition part, [“C” “is” “D” “is”] is a processing unit.
As described above, the sentence is divided into the condition part and the processing part, and converted into a structure having connection between patterns of conditional processing. In addition, when the pattern of the condition part is excited, processing for confirming whether or not the condition part is established autonomously is excited. As described above, when the sentence is converted into a pattern having the structure of the conditional process, it is possible to realize an operation in which the sentence that can be expressed by the conditional process sequentially proceeds the process while confirming the satisfaction of the condition. In general, human problem solving and action decision can be expressed by conditional processing. This machine autonomously converts it into conditional processing by inputting human problem solutions and action decision measures in a language (statement), and then proceeds to perform processing while checking the validity of the condition. It becomes possible. By entering knowledge expressed in language (sentence) (procedure and way of thinking about problem solving and action determination) without programming actions corresponding to human problem solving and action determination, human beings solve problems by thinking It is possible to autonomously carry out problem solving or behavior determination so as to determine behavior.

FIG. 1 is a diagram showing the configuration of an autonomous knowledge extractor according to an embodiment of the present invention.
In FIG. 1, reference numeral 1 denotes a pattern converter for converting information into a pattern. Reference numeral 2 denotes a pattern recorder which records patterns, connection relationships between patterns, and relationships between patterns. Reference numeral 3 denotes a pattern registration unit which registers and changes patterns and connection relationships between patterns by human instruction or autonomously. A pattern controller 4 controls processing of the pattern. Reference numeral 5 is a pattern reverse converter which converts a pattern into information. 6 is a pattern analyzer that analyzes patterns and relationships between patterns. 7 records the history of excitation of the pattern, and when the pattern is excited, data of the connection relationship with the pattern is generated from the data of the history of the pattern excited before that, and the connection relationship recording part of the recording module of the pattern It is a pattern irradiator that excites recording modules with large correlation by collating the history data of excitation patterns from the present time set up to the past and the connection relationship data recorded in the connection relationship recording unit of each recording module. .

Next, the operation will be described.
In FIG. 1, a pattern converter 1 converts information into a pattern. The converted patterns are analyzed in the six pattern analyzers and processing according to the analysis results is performed. The second pattern recorder records patterns, connection relationships between patterns, and relationships between patterns. The pattern register 3 performs registration and change of patterns. The input pattern is collated with the recording module of the pattern recorder to check whether the associated pattern is recorded. If the same or equivalent pattern as the input pattern is recorded, the corresponding pattern is excited, and if not recorded, it is registered and excited as a new pattern. The history of the excited pattern is recorded on the pattern illuminator 7. When a certain recording module is excited, data of the connection relationship with the pattern is generated from data of the history of the pattern excited before that, and is recorded in the connection relationship recording part of the recording module of the pattern. Further, the data of the history of excitation patterns set from the present to the past and the data of the connection relationship recorded in the connection relationship recording unit of each recording module are collated to excite the recording module having a large correlation. In the initial stage, pattern-to-pattern connection generation is performed according to human instruction.

FIG. 2 shows an example of a pattern. Organize the input sentences into a main part (subject and subject modifications) and a predicate (predicates and predicate modifications). Predicate modification further organizes what, when, where, why and how it was implemented. Organizing and storing in this way is very useful for retrieving information. The input word pattern is stored as an intra-sentence pattern. Word-to-word relationships are analyzed, and sentence elements (subjects, predicates, modifiers) and modification relationships are analyzed.
FIG. 3 shows an operation in which words, part-of-speech / meanings, sentence elements, inter-statement relations and modification relations are sequentially identified from word strings. When a word is input, a recording module corresponding to the input word is excited in the word detection area. As the recording module corresponding to the word is sequentially excited, the word corresponding to the word string is detected, and the recording module corresponding to the word is excited. Since the connection relationship with the word string corresponding to each word is generated in the connection relationship recording part of the recording module corresponding to each word, the correlation is large when the word string history is irradiated in the pattern irradiator Is detected and the recording module is excited. When the word pattern is excited, the part of speech / meaning pattern of the word corresponding to the word is excited. Next, according to the appearance pattern of the part of speech of the word, the sentence element and the modification relation between the sentence elements are detected. If there is more than one word to modify, it is located apart from the words to be modified, and it is difficult to identify which word is modifying which word by the order of part of speech Do. (It can be identified by detecting the validity of the combination of the word to be modified and the word to be modified.) Once the relation between the sentence elements (subject, predicate, modification relation) is identified, the relation between the sentence elements Is recorded. Sentences input from this relationship are organized into a main part (subject and subject modifications) and a predicate (predicates and predicate modifications), and the modification of predicates further, what, when, where, why, how It is organized into what it has been implemented and stored as structured in the recording module as an intra-sentence pattern.

Next, it will be described how the input sentence is analyzed and recorded as an intra-sentence pattern in the recording module.
FIG. 4 shows an example of a newsletter. It shows about the procedure which analyzes the example of a text by processing of FIG. When a word is detected, the part of speech and meaning of the word are analyzed at the same time. As part of speech, nouns, verbs, adjectives, adjective verbs, adverbs, particles and the like are identified. Subject identification and modifier identification are performed from noun and particle types. If there is more than one word to modify, it is located apart from the words to be modified, and it is difficult to identify which word is modifying which word by the order of part of speech Do. It can be identified by detecting the validity of the combination of the word to be modified and the word to be modified. By tracing the modification relationship sequentially, it is possible to identify the subject modification and the modification part of the predicate. As a modifier of the predicate, identify what (O), (H1), when (H2), why (H3), how (H4) and how it was performed. This can be implemented by identifying the particle of each sentence element and identifying whether it is in the middle of a modification relationship or the delimitation of each modification sentence element. By the analysis described above, it is possible to identify the predicate, subject modification, subject, predicate modification part (what, when, where, why, how). The analysis result is defined as the connection relation of the intra-sentence pattern, and is recorded as the intra-sentence (word) pattern connection information in FIG. By carrying out such analysis, a sentence can accurately interpret a compound sentence (a sentence in which a sentence is present and a sentence in which a word modification or the like is performed). Usually, there are often multiple subjects, predicates and modifiers in a sentence, but it is necessary to strictly identify which subject and predicate are main and what is otherwise qualified. it can. By performing this analysis on the input sentence, it is possible to exactly identify the sentence and the relation between sentences. A variety of comparisons (formal comparison, semantic comparison, designation of a comparison place, etc.) can be made by comparing in the subject, predicate, and modifier correspondences of the sentence. Also, when searching for information from sentences recorded in the past, specify which information (who, who, what, when, where, why, how, what, what, etc.) you want to search for Because you can, you can search directly for the information you want.

Next, a method of searching for desired information will be described.
The question sentence is converted into the form of subject, subject modification, predicate, predicate modification (what, when, where, why, how) as described above. At this time, since information is missing for the part corresponding to the question, a pattern of [ ^* ] is temporarily arranged. From this, it is possible to generate a search pattern corresponding to the question. In addition, the search pattern is set as to what sentence elements are expected as answers to the questions (subject, subject modification part of the subject, predicate, modification part of the predicate). This can be identified from the position of the corresponding sentence element of [ ^* ]. For example, if you want to search for [when], it will be [ ^* ] H1. Similarly, [where] [[ ^* ] H2, [why] [[ ^* ] H3, [how] [[ ^* ] H4, [who] [[ ^* ] S, [what] [[ ^* ] O, Let's express it as [how] ⇒ [ ^* ] V. The search pattern generated from the question is irradiated to the pattern recorder to search for a correlated pattern. Among the searched patterns, if there is an internal pattern at the above-mentioned [ ^* ] corresponding position, it is a candidate for an answer. In the case of multiple sentences, the hierarchy of the predicate corresponding to the question is important. That is, it is important to extract a response from a hierarchy in which a predicate corresponding to a query is excited when a search pattern is irradiated and a related pattern is searched. If there are multiple predicates in a multiple sentence, it is necessary to identify which predicate the response corresponding to is expected.

When we convert a language that expresses human thinking into patterns, and analyze patterns and relationships between patterns, we can identify sentence types (normal sentences, question sentences, imperative sentences, etc.), features (truths, truths, facts, definitions, rules, etc. Common sense, explanations, hypotheses, predictions, opinions, impressions) and relationships (causes and consequences, events and reasons, explanations and conclusions, outlines and details etc) can be identified. It is possible to make the machine learn the identification so that the type and feature of the sentence can be identified by humans. By analyzing the words contained in a sentence, the type of sentence can be identified as a plain sentence, an interrogative sentence, an imperative sentence, and the like. Among the features of sentences, human beings teach machines to learn about each of truth, truth, facts, definitions, rules and common sense. This can be implemented by identifying and recording each type of information (truth, truth, fact, definition, rule, common sense) as each information is input to the machine. The identification of the type of other information (explanation, hypothesis, prediction, opinion, impression) is that the type of sentence is not recorded as (truth, truth, fact, definition, rule, common sense) and the words included in the sentence ( It can be implemented by analyzing, thinking, thinking, etc.).

When the language is converted into a pattern and expressed as a set of patterns corresponding to the words constituting the sentence, and the semantic relationship between the words is expressed by the pattern-pattern connection relation, the meaning of the sentence is the pattern and the pattern It can be expressed in connection relation. Information input from word strings to word identification, word attributes (part-of-speech, meaning), analysis of sentence elements, sentence structure analysis (subject, predicate, subject modification, predicate modification, modification relationship), sentence elements and Relationship analysis of sentence elements (same meaning, definition, opposite meaning, etc.) is carried out, and the relation between information and information meaning is pattern and pattern by correlating the relation between sentence elements and sentence elements as pattern-pattern connection relation Convert to the connection relationship of The relationship between patterns and patterns can be set not only for sentence elements and sentence elements, but also for sentences and sentences, sentences and sentences. This can be implemented by defining a group of patterns as a new pattern. When a characteristic word indicating a pattern-pattern relationship (cause and result, event and reason, explanation and conclusion, outline and detail, etc.) is detected, the corresponding relationship between corresponding patterns is recorded autonomously. The relationship between the recorded patterns can be utilized in controlling the transition between the patterns.
A pattern-pattern connection relation can express various relations such as logical relation, definition, relation of attribution, similar relation, relation between action and result, development of inference, etc. Inheritance of attribute, common Inheritance of features and identification of individual features can also be expressed flexibly.
By defining the relationship between information and information as a connection relationship between patterns in this way, information is not recorded in a single state, but is recorded as a knowledge system having a relationship with other information. It becomes possible.

Although the relationship between information and information has various relationships, the relationship between information (the same meaning, definition, opposite meaning, similarity, logic, cause, result, detail, summary, summary, related information, etc.) By selecting an appropriate one in each processing phase (corresponding to the stage of the thinking process), it is possible to properly control the pattern transition from the pattern.
When information is input, the meaning is analyzed from the words contained in the information, and a pattern indicating the meaning equivalent to the information is generated. This pattern is generated while maintaining the subject, predicate, subject modification, predicate modification, and modification relationship of the sentence. The generated pattern is used to compare with the recorded pattern, and analysis is performed to determine whether or not the related pattern exists. Thus, since information can be searched at the semantic level, it is possible to search for content related information even if the words do not match.
The newly input information can be checked against the already recorded knowledge system to evaluate consistency and novelty. If it is not consistent with the information you identify and record as truth, truth, facts, definitions, rules and common sense, then the information is likely to be false. On the other hand, when it is not possible to judge whether the information is consistent or not, it is considered that knowledge sufficient to determine the information is not stored yet, and human judgment is requested. If human beings can judge that the information is correct, they shall notify the machine of the result. The machine records the information in the position of explanation or hypothesis, and by obtaining other information in the future, the information that was evaluated when it could be logically expanded with the combination of truth, truth, fact, definition, rule and common sense As identified and recorded.
In this way, information is converted into patterns, information types, features are identified, classified and recorded, and information-information relationships (logical relationships, similar relationships, reciprocal relationships, analogies, various relationships, etc.) Information can be constructed as a knowledge system by expressing it as a pattern-pattern connection relation.

FIG. 5 shows how to evaluate the input information when the information is input. In the first stage, we will conduct a reliability assessment of the information. This can be done by verifying the credibility of the information source (when, who, where did it come from?).
In the second stage, we will carry out field / theme analysis of information. This can be detected for the field / theme from the words contained in the information.
The third step is to identify the type of information. Of the types of input information, truth, truth, facts, definitions, rules, and common sense shall be in accordance with human designations. Descriptions, hypotheses, predictions, opinions, impressions, etc. can be identified from the words contained in the information (will, think, think ...). It is possible to carry out from the word included also about the plain text, the question sentence, the command sentence, and the identification of the exclamation sentence.
At the 4th stage, the degree of interest is evaluated. It is possible to record in advance the field / theme of the information of interest, and to evaluate by checking whether it matches with the field / theme analysis result of the information or not.
In the fifth stage, a novelty evaluation will be conducted. In this case, the input information and the pattern semantically equivalent to the input information are irradiated to the recording area, and the presence or absence of the related pattern is confirmed. When a related pattern is detected, the difference between the patterns can be compared and evaluated by evaluating whether or not there is a new pattern.
In the sixth stage, a validity assessment will be conducted. In this method, a pattern of input information and related information is applied to a recording area defined for inter-word relationships, and alignment and inconsistency are evaluated for each sentence element. If inconsistencies are detected, the reliability of the input information and related information (truth, truth, facts, definitions, rules, common sense, reliability of information sources) is evaluated, and reliable ones are preferentially recorded. I decided to.

6 to 10 show an operation example for evaluating input information and constructing it as a knowledge system.
FIG. 6 shows an operation example of analyzing the reliability of information, the field, the theme, and the degree of interest.
In FIG. 7, assuming that the pattern of the input information is [PA], the pattern of [PA] is irradiated to the recording area storing the meaning and the like of the word, and the meaning of the corresponding word is searched. By carrying out this search in association with the sentence structure (subject, predicate, subject modification, predicate modification), it is possible to generate a pattern that is semantically equivalent to the input information. This semantically equivalent pattern is expressed as [PA #]. When searching related information from input information, a number of search methods can be used. Information that exactly matches the input information, information that is semantically equivalent, information that partially matches, etc. can be selected according to the purpose of the search.
FIG. 8 shows that the semantically equivalent pattern [PA #] is irradiated, and [PB] is detected as related information. A semantic equivalent pattern [PB #] can be generated by similarly analyzing the meaning of [PB].
FIG. 9 shows an analysis of the relationship between the input information and the related information already recorded. Matching patterns and different patterns can be extracted from the difference between [PA #] and [PB #]. In addition, by irradiating [PA #] and [PB #] to a recording area that defines the relation between words (equivalent, similar, opposite, etc.), it is possible to detect a semantic difference.
By making the irradiation lines of the input information and the related information different, it is possible to identify whether the excitation of the pattern corresponding to each word is due to the input information or the related information.
In the recording area in which the relation between words is defined, the relation between each word is identified by the connection relation. For this reason, if the word in which the input information is excited and the word in which the related information is excited are semantically the same, a pattern having the same meaning is excited, and it can be identified that the meaning is the same. Conversely, if the word in which the input information is excited and the word in which the related information is excited are semantically opposite, a pattern indicating the opposite meaning is excited and it can be identified that the meaning is opposite. FIG. 10 shows the operation of evaluating the usefulness of the input information and, when it is evaluated as useful, adding identification results such as information type, field, theme, degree of interest, and recording in a predetermined recording area There is.

By constructing information as a knowledge system in this way, it becomes possible to solve problems using the knowledge system. First, analyze the problem using the knowledge system, clarify the problem and set the goal of the problem solution. Next, extract and apply a solution for solving the problem. The situation after application is evaluated, and if the goal of problem solving is reached, the process is completed. If the problem solving goal has not been reached, apply the above process to the new state again and bring it closer to the problem solving goal.
Analysis of problems and clarification of problems are implemented by learning how to detect and identify the difference between the target state and the current state. Moreover, the solution plan corresponding to each problem is also implemented by learning the corresponding processing. Learning is performed by recording the history of pattern transitions. The content taught by humans is recorded as a pattern-to-pattern transition.
As learning progresses, the taught process is autonomously executed as a pattern-to-pattern transition, and the taught content is executed.
Learning is performed by recording the history of pattern transitions. The content taught by humans is recorded as a pattern-to-pattern transition. As learning progresses, the taught process is autonomously executed as a pattern-to-pattern transition, and the taught content is executed.

Transition of human thinking pattern can be generally expressed as conditional processing from the viewpoint that the transition destination changes depending on conditions. In the present invention, the meaning of language is interpreted and converted autonomously as conditional processing. The condition of the conditional process is generated by generating a search pattern from the corresponding language and searching autonomously. It is determined whether the retrieved information satisfies the condition, and if it is satisfied, the corresponding processing is executed. Even if the problem solution used by human being as knowledge is directly input as a language, the meaning is sequentially interpreted, and the conditional processing is autonomously advanced to solve the problem. If new information is required during the process, the information request is notified, and when the corresponding information is acquired, the process according to the content is performed.
In the learning phase, when a pattern is excited, the history of the excited pattern is referred to in a set period before the pattern is excited, and the connection relationship with the excited pattern is strengthened. In the practical phase, the history of the excitation pattern in the set period is collated with the recorded connection relationship with other patterns recorded in each recording module, and when the correlation is large, the corresponding pattern is excited. When the pattern is excited, the history of the excitation pattern is updated, and a pattern having a large correlation with the history of the excitation pattern in a set period in a new state is sequentially excited.

As described above, in the learning phase, a series of patterns corresponding to processing are sequentially excited according to a human instruction, and the history is recorded, whereby the corresponding pattern is excited according to the instructed procedure. The behavior of the pattern is not static but shows dynamic behavior. The internal pattern recorded in the pattern can be used to retrieve the required information and store the search results in the required location. In addition, processing such as conversion of an internal pattern recorded in a pattern into a designated arrangement is also possible. A number of processes (formula processing, chemical formula processing, translation, etc.) can be performed by combining such patterns capable of dynamic behavior.
The teaching to the present autonomous knowledge extractor can be implemented by sequentially inputting linguistic information without programming. The input linguistic information is analyzed in relation to syntax, meaning, and information already recorded, and in accordance with the analysis result, the corresponding pattern is excited and processing is performed. A number of processes can be performed, such as evaluation and recording of input information, execution of instructed instructions, and generation of solutions to problems and problems.
The overall operation of the present autonomous knowledge extractor is managed by a pattern controller. In the transition cycle of each pattern, information input, information analysis (type of sentence, syntax, meaning etc.), information evaluation (newness, reliability, validity, usefulness etc.), information processing (problem / problem solution) Generation, recording, information output, etc.).

FIG. 11 shows an execution example of the instruction content. The conditional processing is identified from the input sentence (language), and it is shown about the operation | movement which performs a processing, confirming the establishment of a condition autonomously. By analyzing the input sentence, the in-sentence pattern corresponding to the condition is extracted from the pattern corresponding to the input sentence, and the search pattern [PQ] for the condition search is generated. By irradiating the search pattern [PQ] to the pattern recording device, information corresponding to the pattern [ ^* ] corresponding to the condition in the search pattern is searched, and the search result is stored. It is assumed that information on conditions is recorded in the pattern recorder. If the information regarding the condition is not recorded in the pattern recording device, the fact is notified and the additional input of the information is requested.
When information corresponding to the condition is retrieved, this information is compared with the condition described in the input sentence. The processing is determined according to whether or not the matching results match. If the collation matches, the pattern corresponding to the process described in the input sentence is excited to execute the process.

FIG. 12 shows an operation of analyzing the input language (question) and generating an answer to the question autonomously. When it is detected that the input sentence is a question, a search pattern corresponding to the question is generated from the intra-sentence pattern contained in the input sentence. Since the input sentence is already analyzed about the subject, the predicate, the subject modification, the predicate modification (what, when, where, why, how), the search pattern is a query for information on which part of the sentence element Can identify the subject of A search pattern is generated from this identification result. Next, the search pattern is irradiated to the pattern recorder to search for related information. Among the patterns showing correlation by irradiation of the search pattern, the pattern having the highest correlation with the search pattern is stored, and the pattern having information in the target portion of the question is stored as a search result. If necessary, it is possible to add a process for confirming whether the search result satisfies the expected response condition. A pattern corresponding to the answer to the question is generated and output from the search result and the search pattern. The pattern is converted into information (language) by the pattern reverse converter and output.

FIG. 13 shows an operation in which the history of the excited pattern is recorded in the pattern irradiator. Also, when a certain recording module is excited, data of connection relation with the pattern is generated from data of the history of the pattern excited before that and recorded in the connection relation recording part of the recording module of the pattern concerned and related to excitation It shows about the operation | movement which the connection relation with a pattern strengthens. The pattern irradiator collates the data of the history of excitation patterns from the present time to the past set up with the connection relationship data recorded in the connection relationship recording unit of each recording module, and excites the recording module having a large correlation.
FIG. 14 shows an operation in which the connection relationship between the corresponding pattern and the pattern is strengthened when the excitation pattern repeatedly appears.
Next, generalization of input information will be described.
Perform feature extraction and generalization of part of the words included in the input sentence. For example, proper nouns appearing in a sentence are generalized as person A, person B, thing C, thing D, and so on. Sequentially, the sentence is converted into a pattern, and the converted pattern is excited. At this time, the generalized pattern is also excited sequentially. By inputting a large amount of information, the connection between specific patterns and patterns constituting a sentence is strengthened. The input sentence has a relation with the neighboring sentence. Since this relation is expressed by the relation between words constituting the sentence, the same relation is often expressed by the combination of identical words or similar words. Every time the sentence pattern is excited, the words contained in the sentence are excited, but by generalizing the proper noun etc., the generalized relationship between the person and the person or thing is emphasized, and so on. Sentences appear more often than when expressed as proper nouns.
The effect is to extract relationships between objects that do not depend on proper nouns. Moreover, since the relationship between the features of the words is extracted instead of the relationship between the individual words by the feature extraction of the words, generalization from the feature aspect becomes possible. This relationship can be detected as the connection between a particular pattern and the pattern is strengthened. In particular, the relationships strengthened from the many sentence examples have generality, and it is considered that those corresponding to common sense or general thinking are extracted. This common sense or general thinking depends on the sentences to be input. That is, when a sentence group corresponding to another culture is input, the common sense or general idea corresponding to the culture is extracted. When the text group in the same culture is inputted, since the thinking is considered to be equivalent, the thinking strengthened by many in the same culture will be extracted.
FIG. 15 creates common sense or general idea by generating a pattern which generalizes or features extracted a part of words of the input sentence, and sequentially inputs the generalized pattern and strengthens the connection relation between the patterns. An operation example is shown.
Similarly, by performing feature extraction and generalization of some words on a number of problems and their solutions, it is possible to extract generalized relations about the problems and their solutions. As learning by information input progresses, this machine will be able to generate solutions autonomously for similar problems.
FIG. 16 shows an operation example for generalizing the problem and the solution by generalizing the problem and the solution pattern, and inputting and strengthening the connection relationship between the patterns.

FIGS. 17 to 20 show an operation example of executing processing directly from the input language without programming.
FIG. 17 shows an operation example of irradiating an input sentence to a recording area for detecting the kind of sentence, identifying the kind of sentence (a plain sentence, an inquiring sentence, an imperative sentence) and calling a processing sequence according to the kind of sentence. There is.
In FIG. 18, when an input sentence (language) corresponding to conditional processing is input, a pattern for confirming the satisfaction of the condition is excited, and information corresponding to the condition is autonomously searched. It shows about the operation | movement which performs corresponding processing, collating the searched information with the conditions described in the input sentence, and matching.
FIG. 19 shows an operation example of autonomously solving a problem by utilizing knowledge recorded in a language. Call up a pattern that shows the state of the object and detect the difference from the state of the target. Excise patterns of analyzing causes / problems from patterns of differences and patterns of corresponding countermeasures. When the pattern corresponding to the countermeasure is executed and the state changes, the above process is repeated until the target state is reached for the changed state.
FIG. 20 shows an operation example of recording information in a pattern recorder in a state where information is identified as a condition part and a processing part and arranged as a sentence structure. By analyzing the words of the information, the conditional part and the processing part of the information can be identified. For example, in the sentences "A", "ga", "B", "at", "C" and "perform", ["A", "ga", "B" and "at the time]] are condition parts, and" [C "Do it"] is a processing unit. In the sentence "A""GA""B""NA""C""HA""D""is" [[A] "GA""B""NA""is the condition part, [“C” “is” “D” “is”] is a processing unit.
As described above, the sentence is divided into the condition part and the processing part, and converted into a structure having connection between patterns of conditional processing. In addition, when the pattern of the condition part is excited, processing for confirming whether or not the condition part is established autonomously is excited. As described above, when the sentence is converted into a pattern having the structure of the conditional process, it is possible to realize an operation in which the sentence that can be expressed by the conditional process sequentially proceeds the process while confirming the satisfaction of the condition. In general, human problem solving and action decision can be expressed by conditional processing. This machine autonomously converts it into conditional processing by inputting human problem solutions and action decision measures in a language (statement), and then proceeds to perform processing while checking the validity of the condition. It becomes possible. By entering knowledge expressed in language (sentence) (procedure and way of thinking about problem solving and action determination) without programming actions corresponding to human problem solving and action determination, human beings solve problems by thinking It is possible to carry out problem solving or action determination so as to determine action.
[Effect of the invention]

According to the present invention, the source of the information, the subject, the subject, the subject, the subject modification, the predicate, the predicate modification, the modification relationship, when, where, who, what, how, why, and the like regarding the input information It is possible to analyze and organize into a sentence structure and autonomously record information judged to be useful as a knowledge system. Also, the input information is compared with the internally constructed knowledge system and evaluated, and the processing according to the evaluation result (recording of information, updating and improvement of the knowledge system, execution of instruction contents, answer to questions) is autonomously performed Do. Furthermore, by generalizing a part of the words contained in the information, and by strengthening the relationship between the set of patterns corresponding to the information and the relationship between the information input sequentially and the set of patterns, the significant patterns are provided between the patterns. It is possible to extract relationships and autonomously construct common sense, general thinking and problem solving methods from a series of input information. Conventionally, when making a machine process input information, it is necessary to set a program in advance in a computer. It was necessary to create a program for judging the situation from the input information and a program to operate the machine according to each condition, and install and execute it on a computer mounted on the machine. The program needs to be created in a dedicated programming language, and has the disadvantage of requiring a lot of time for development. If the condition detection and the corresponding operation are not appropriate, the program installed in the computer needs to be corrected by a human, and there is a disadvantage that the correction takes a lot of time.
According to the present invention, processing is performed by learning how to process information, so there is no need to program sequentially. By inputting information on the method of processing, the machine records and learns the method of processing, so that labor can be greatly reduced.
In addition, it is possible to cope with the change of processing by changing the pattern and the connection between the patterns without changing the program, so the system is very flexible and highly responsive.
In addition, with regard to problems that can be solved by thinking (language) by humans, the autonomous knowledge extractor of the present invention uses the method or procedure of the input problem solving by indicating in language the method or procedure for solving the problem. According to the problem can be solved autonomously.
If there is any ambiguity or uncertainty in the method or procedure of problem solving, each time it is notified, the problem can be solved while clarifying the method and procedure of the solution. Furthermore, the present invention autonomously converts into a conditional process by inputting human problem solutions and action decision measures in a language (statement), and then proceeds with the process while checking the validity of the condition. Operation is possible. By entering knowledge expressed in language (sentence) (procedure and way of thinking about problem solving and action determination) without programming actions corresponding to human problem solving and action determination, human beings solve problems by thinking It is possible to carry out problem solving or action determination so as to determine action.

Configuration example of an autonomous knowledge extractor Pattern configuration example Example of operation in which words, sentence elements and modification relationships are sequentially identified from word strings Examples of current events Evaluation example of input information Example of Construction of Knowledge System (Part 1) Example of Knowledge System Construction (Part 2) Example of construction of knowledge system (3) Example of construction of knowledge system (4) Example of construction of knowledge system (5) Execution example of instruction contents (example of conditional processing) Example of answer to question Operation example of connection enhancement of excitation pattern and related pattern (Part 1) Operation example of connection enhancement of excitation pattern and related pattern (part 2) Examples of construction of common sense and general thinking Operation example of problem and solution generalization How to execute processing directly from the input language without programming Example of sequential execution of input and language groups An example of generating the processing procedure autonomously and further generating the next processing procedure autonomously for the situation resulting from the processing Operation example of identifying input information in condition part and processing part and recording as sentence structure

1 pattern converter 2 pattern recorder 3 pattern register 4 pattern controller 5 pattern reverse converter 6 pattern analyzer 7 pattern irradiator

Claims

The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who An artificial intelligence that is equipped with a pattern analyzer that analyzes what is, how, and why, and organizes and records in sentence structure, and autonomously records information that is judged to be useful and builds it as a knowledge system.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , What, how, why, it has a pattern analyzer that organizes and records in sentence structure, records information judged to be useful autonomously, constructs as a knowledge system, and inputs the information Artificial intelligence that autonomously performs processing (Recording of information, updating and improvement of knowledge system, execution of instruction contents, answer to questions) according to the evaluation result by collating and evaluating the internally constructed knowledge system.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , Analysis of what has been done, how and why, organized in sentence structure and recorded, as well as feature extraction of some of the words contained in the information and a pattern analyzer to generalize, Information is extracted by strengthening the relationship between the set of patterns corresponding to the information and the set of patterns, and autonomously constructing common sense and general ideas from a series of input information. Artificial intelligence.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , Analysis of what has been done, how and why, organized in sentence structure and recorded, as well as feature extraction of some of the words contained in the information and a pattern analyzer to generalize, Information is extracted by strengthening the relationship between a set of patterns corresponding to information and a set of patterns corresponding to the information, and the solution of the similar problem from the input information on the series of problems and the solution is autonomously Artificial intelligence to generate.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who A processing procedure is provided with a pattern analyzer that analyzes what has been done, how and why, and organizes and records in sentence structure, and appropriately recalls related information from the input information and the knowledge system built inside An artificial intelligence that generates autonomously, and autonomously generates the next processing procedure for the situation resulting from the processing.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , Analysis of what was done, how, why, and further analysis to identify the input information into the conditional part and the processing part Conducted analysis, organized into sentence structure and recorded, provided with input Performs processing by appropriately calling relevant information from the generated information and the knowledge system built inside, and when the called information has a conditional part, artificially confirms the validity of the conditional part and advances the processing Intelligence.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who As artificial intelligence that is equipped with a pattern analyzer that analyzes what is, how, and why, and organizes and records in sentence structure, and records information judged to be useful autonomously as a knowledge system Software to make it work.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , What, how, why, it has a pattern analyzer that organizes and records in sentence structure, records information judged to be useful autonomously, constructs as a knowledge system, and inputs the information Function as artificial intelligence that autonomously executes processing (information recording, update and improvement of knowledge system, execution of instruction contents, answer to question) according to the evaluation result by collating and evaluating the knowledge system built inside Software to make it happen.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , Analysis of what has been done, how and why, organized in sentence structure and recorded, as well as feature extraction of some of the words contained in the information and a pattern analyzer to generalize, Information is extracted by strengthening the relationship between the set of patterns corresponding to the information and the set of patterns, and autonomously constructing common sense and general ideas from a series of input information. Software to function as artificial intelligence.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , Analysis of what has been done, how and why, organized in sentence structure and recorded, as well as feature extraction of some of the words contained in the information and a pattern analyzer to generalize, Information is extracted by strengthening the relationship between a set of patterns corresponding to information and a set of patterns corresponding to the information, and the solution of the similar problem from the input information on the series of problems and the solution is autonomously Software to function as artificial intelligence to generate on a regular basis.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who A processing procedure is provided with a pattern analyzer that analyzes what has been done, how and why, and organizes and records in sentence structure, and appropriately recalls related information from the input information and the knowledge system built inside Software for autonomously generating and processing artificial intelligence to generate the next processing procedure autonomously and proceed with processing for the situation that resulted from the processing.
The pattern converter consists of a pattern converter that converts information into patterns, a pattern recording unit that records patterns, the connection between patterns, and the relationship between patterns, and the pattern relationship between the patterns and patterns is indicated by humans or autonomous A pattern register that registers and changes the pattern, a pattern controller that controls processing of the pattern, and a history of pattern excitation, and when the pattern is excited, the history of the pattern excited before the pattern is excited Data of connection relationship with the pattern is generated from the data and recorded in the connection relationship recording unit of the recording module of the pattern and data of the history of excitation patterns up to the past set from the present time are connection relationship recording unit of each recording module Check the connection data recorded in the file, and excite the recording module with a large correlation A pattern irradiator, a pattern reverse converter for converting a pattern into information, a source of information regarding input information, field, theme, subject, subject modification, predicate, predicate modification, modification relationship, when, where, who , Analysis of what was done, how, why, and further analysis to identify the input information into the conditional part and the processing part Conducted analysis, organized into sentence structure and recorded, provided with input Performs processing by appropriately calling relevant information from the generated information and the knowledge system built inside, and when the called information has a conditional part, artificially confirms the validity of the conditional part and advances the processing Software to function as intelligence.