JP2015133138A - autonomous intelligence generator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To advance generation of intelligence and construction of an intelligence system by analyzing information autonomously.SOLUTION: In an initial stage, a connection relationship between patterns is generated from a problem pattern toward a pattern of a field, a problem type, and a solution. For several problems, the connection relationship between patterns is allowed to learn to strengthen the combination between patterns. When a problem pattern is inputted, a pattern corresponding to a field and a problem type and a pattern for achieving a target of a problem solution is excited to generate a connection relationship. Then, the pattern of a field and a problem type is allowed to operate on the problem pattern to excite a pattern indicating a problem point. A pattern for correlating solution to the pattern for indicating a problem point is operated and a pattern for indicating solution is excited. For the pattern for indicating a problem point, the pattern for indicating the solution is operated, thus exciting a pattern for indicating a state after execution of the solution to the problem point. A pattern is shifted by operation between patterns to bring the pattern close to a pattern for a problem solution.

Description

  The present invention relates to artificial intelligence and software that sequentially learns problem solving techniques from input information and constructs intelligence that autonomously solves problems.

  Conventional artificial intelligence performs actions according to pre-programmed procedures. Implementations other than the processing set by the program are difficult, and it is difficult to learn a problem solving method from the input information and solve similar problems autonomously.

Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. In addition, it is difficult to learn a problem solving method from input information and solve similar problems autonomously.
In the present invention, processing for analyzing and recording information and the structure of information is performed by human instruction and learning, and a problem solving method is learned for the relationship between information by connection between patterns and processing between patterns. In addition, we will analyze information autonomously, and proceed with the generation of intelligence and the construction of an intelligent system.

(Corresponding to claim 1)

When the machine performs processing of input information, 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 such that when a preset condition is detected, a corresponding operation is executed. If the detection of the condition and the corresponding action are not appropriate, the program installed on the computer is modified.
Conventionally, there is no artificial intelligence and software that converts input information into patterns and records them sequentially, learns problem solving methods by human instructions and autonomous learning, and autonomously generates answers to similar problems.

(Corresponding to claim 2)

When the machine performs processing of input information, 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 such that when a preset condition is detected, a corresponding operation is executed. If the detection of the condition and the corresponding action are not appropriate, the program installed on the computer is modified.
Converts input information into patterns and records them sequentially, and learns mathematical and physical formula processing and chemical formula processing by human instruction and autonomous learning, and generates answers autonomously for mathematical, physical and chemical problems There is no artificial intelligence and software.

(Corresponding to claim 3)

When the machine performs processing of input information, 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 such that when a preset condition is detected, a corresponding operation is executed. If the detection of the condition and the corresponding action are not appropriate, the program installed on the computer is modified.
Conventionally, there is no artificial intelligence and software that converts input information into a pattern and sequentially records it, learns translation into another language by human instruction and autonomous learning, and autonomously generates translation into another language.

(Corresponding to claim 4)

When the machine performs processing of input information, 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 such that when a preset condition is detected, a corresponding operation is executed. If the detection of the condition and the corresponding action are not appropriate, the program installed on the computer is modified.
The input information is converted into patterns and recorded sequentially, and problem solving methods are learned by human instruction and autonomous learning, situation analysis, problem extraction, problem solution drafting, solution drafting Traditionally, there is no artificial intelligence and software that performs evaluation and solution selection and autonomously generates solutions to similar problems.

Problems to be solved by the invention

(Corresponding to claim 1)

  Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. In addition, it has been difficult in the past to record input information sequentially, learn problem solving methods by human instructions and autonomous learning, and generate answers autonomously for similar problems.

Problems to be solved by the invention

(Corresponding to claim 2)

  Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. In addition to recording input information sequentially, learning mathematical and physical formula processing and chemical formula processing by human instruction and autonomous learning, and generating answers autonomously for mathematical, physical and chemical problems has been conventional It was difficult.

Problems to be solved by the invention

(Corresponding to claim 3)

  Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. In addition, it has been difficult in the past to record input information sequentially, learn translation into other languages by human instructions and autonomous learning, and autonomously generate translations into other languages.

Problems to be solved by the invention

(Corresponding to claim 4)

  Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. In addition to sequentially recording the input information, learn the problem solving method by human instruction and autonomous learning, analyze the situation, extract the problem, generate the solution of the problem, evaluate and solve the solution It has been difficult in the past to select solutions and generate solutions autonomously for similar problems.

Means for solving the problem

(Corresponding to claim 1)

Human thought is expressed in language, but the information expressed in this language is converted into what is called a pattern. When converted into a pattern, human thoughts can be expressed as individual patterns, and human thought transitions can be viewed as changes from pattern to pattern. Further, the pattern is not only a language expression, but can express image information, for example, and can process information. Further, it is a concept that has a very wide range that can be handled, for example, it is possible to generate a signal for driving the driving device to perform the operation.
If information is converted into such a pattern and human thought is regarded as a pattern, human thought can be expressed as a change from pattern to pattern. When we convert the language that expresses human thoughts into patterns and analyze the patterns and the relationships between patterns, the types of sentences (normal sentences, question sentences, imperative sentences, etc.) and features (truths, truths, facts, definitions, rules, Common sense, explanation, hypothesis, prediction, opinion, impression).
Sentence types can be trained by machines so that humans can identify the types and characteristics of sentences. The sentence type can be identified as a normal sentence, a question sentence, a command sentence, etc. by analyzing words included in the sentence. Among the features of the sentence, truth, truth, facts, definitions, rules, and common sense are taught by humans for each individual and learned by the machine. This can be done by identifying and recording each type of information (truth, truth, fact, definition, rule, common sense) as each information is entered into the machine. Identification of other types of information (explanation, hypothesis, prediction, opinion, impression) is that the sentence type is not recorded as (truth, truth, fact, definition, rule, common sense) and the word ( I think, think, think, etc.) can be implemented by analyzing.
If the language is converted into a pattern and expressed as a set of patterns corresponding to the words that make up the sentence, and the semantic relationship between the words is expressed by the connection relationship between the patterns, the semantic relationship of the sentence It can be expressed by the connection relation of the pattern.
In addition, this connection relationship can express various relationships such as logical relationship, definition, attribution relationship, similar relationship, relationship between action and result, inference development, inheritance of attributes, common features, etc. Inheritance and identification of individual features can be expressed flexibly.
The newly entered information can be checked against the already recorded knowledge system and evaluated for consistency and novelty. Information that is identified as truth, truth, facts, definitions, rules and common sense and is not consistent with the information recorded is likely to be incorrect. On the other hand, if it is not possible to judge whether the information is consistent or not, it is considered that knowledge sufficient to determine the information has not yet been stored, and therefore human judgment is required. If humans can judge that the information is correct, the result will be notified to the machine. The machine records the information as an explanation or hypothesis, and obtains other information in the future, so that it can be logically expanded with a combination of truth, truth, facts, definitions, rules and common sense as evaluated information. It shall be identified and recorded.
In this way, information is converted into patterns, information types and characteristics are identified, classified and recorded, and information-information relationships (logical relationships, similar relationships, reciprocal relationships, analogies, etc.) Information can be constructed as a knowledge system by expressing the connection relationship between patterns.
When information is constructed as a knowledge system in this way, problem solving methods can be learned and knowledge can be improved. First, analyze the problem, clarify the problem, and set a goal for problem solving. Next, a solution plan for solving the problem is extracted and applied. 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, the above process is applied again to a new state, and the problem-solving goal is approached.
Problem analysis and problem clarification are implemented by learning how to detect and identify differences between the target state and the current situation. Moreover, the solution proposal corresponding to each problem is implemented by learning the corresponding processing.
The above is explained using patterns as follows.
Analyze problem patterns and identify what areas and what types of problems. In the early stages, humans are taught to let the machine learn about domain and problem type identification, and solutions corresponding to problem types. This is because when each problem pattern is input to the machine, it excites patterns related to words, word meanings, features, attributes, etc. that humans consider for problem identification, and then the fields and problem types identified by humans. And, by sequentially exciting patterns representing solutions corresponding to the problem types, connection relationships between patterns are generated from the problem patterns to the field, problem types and solution patterns. For a number of problems, we will learn such connection relationships between patterns and strengthen the connection between patterns. As learning progresses to some extent, when a problem pattern is input, the machine autonomously excites the corresponding field, problem type, and solution pattern. If the result is incorrect, the machine is notified that the person is wrong, and the connection with the incorrect pattern is weakened and the connection with the correct pattern is strengthened. Such learning allows machines to identify problem areas, problem types and solutions,
When the pattern learning is completed, the process after the problem is input is learned. When a problem pattern is input, a pattern corresponding to a field, a problem type, and a pattern expressing a problem solving target are excited to generate a connection relationship. Next, the pattern of the field and problem type is applied to the problem pattern to excite the pattern representing the problem. (By defining the relationship between the patterns, it is possible to define the inter-pattern coupling so that the pattern representing the result of the action is excited by exciting a certain input pattern and the pattern representing the action.) Then, a pattern that associates the solution with the pattern that represents the problem is caused to act, and the pattern that represents the solution is excited. By applying the pattern representing the solution to the pattern representing the problem, the pattern representing the state after the solution is applied to the problem is excited. Patterns are transitioned by such an action between patterns, and are brought closer to a problem solving pattern. The state in which the solution is applied to the problem is generated by identifying the pattern of each element constituting the pattern of the problem and changing the pattern of each element in response to applying the solution.
There are two types of thought patterns when humans are trying to solve problems. One is a thought pattern corresponding to the problem itself, and the other is a thought pattern related to a procedure for solving the problem. When something goes wrong, humans examine the situation, analyze the situation, identify the cause of the problem, and think about ways to solve the cause of the problem. The thinking pattern of the procedure for solving the problem acts on the thinking pattern of the problem, and the thinking pattern is changed toward the problem solving. This process can be expressed using patterns as follows.
1. Define the problem Express the problem as a pattern (P (i))
2. Analyzing the problem Exciting the relevant pattern from the problem pattern and the pattern of the elements that make up the problem (if no difference from the state that should be the problem pattern alone is detected, an equivalent pattern, a logically derived pattern, (Exciting related patterns and revealing problematic patterns)
(Pdi)
3. Cause analysis Compare the pattern of the desired state and component with the above problem and the component of the problem, and identify the pattern corresponding to the different element (Ppi)
4). Selection of treatment Excites a pattern corresponding to the policy to eliminate the above-mentioned differences in question. (The policy to solve the problem is to change the pattern of the problem based on the procedure and rule of thumb for solving the problem. Generate the pattern to be used.)
(Pai)
5. Eliminate the cause of the problem. Apply the treatment pattern to the problem pattern. (By applying the treatment pattern, the pattern corresponding to the element in question is eliminated.)

Means for solving the problem

(Corresponding to claim 2)

A description will be given of means for sequentially recording input information and learning mathematical and physical mathematical formula processing and chemical formula processing by human instruction and autonomous learning to autonomously generate answers to mathematical, physical and chemical problems.
Analyze patterns of mathematics, physics and chemistry problems and identify what areas and what types of problems. In the early stages, humans are taught to let the machine learn about domain and problem type identification, and solutions corresponding to problem types. This is because when each problem pattern is input to the machine, it excites patterns related to the words, word meanings, features, attributes, etc. included in the problem, and then into the fields, problem types and problem types identified by humans. By sequentially exciting the patterns representing the corresponding solution and processing method, a connection relationship between the patterns is generated from the problem pattern to the field, the problem type and the solution pattern. For a number of problems, we will learn such connection relationships between patterns and strengthen the connection between patterns. As learning progresses to some extent, when a problem pattern is input, the machine autonomously excites the corresponding field, problem type and solution and processing pattern. If the result is incorrect, the machine is notified that the person is wrong, and the connection with the incorrect pattern is weakened and the connection with the correct pattern is strengthened. Such learning allows the machine to identify problem areas, problem types and solutions and processing methods.
When the pattern learning is completed, the process after the problem is input is learned. When a problem pattern is input, a pattern corresponding to a field, a pattern corresponding to the problem type, and a pattern expressing the answer type are excited to generate a connection relationship. Next, the pattern of the field and question type is applied to the question pattern, and the pattern representing the difference from the answer type is excited. By simultaneously exciting the pattern expressing the state and the pattern expressing the action, the inter-pattern coupling is defined so that the pattern expressing the state that produces the result of the action is excited. In this way, an action pattern corresponding to the state pattern is applied to change to a new state. By detecting the difference from the answer type for the new state again and applying the action pattern to eliminate the difference one after another, the pattern expressing the state approaches the pattern expressing the answer. To go. A pattern expressing a difference is excited from a difference between a pattern indicating an initial state and a pattern indicating an answer, and further a pattern indicating a solution and a processing method for eliminating the difference is excited. By applying the pattern indicating the solution and processing method to the pattern corresponding to the state, the state is changed to a state in which the difference is reduced. It is possible to generate an answer to be obtained by repeating such a pattern and the action of the pattern until it matches the answer type.

Means for solving the problem

(Corresponding to claim 3)

Record the input information sequentially, learn translation to other languages by human instruction and autonomous learning, and analyze the words that make up the sentence on the machine explaining the means to autonomously generate translations to other languages , To learn how to identify the type of sentence (normal sentence, question sentence, command sentence, exclamation sentence). Students will also learn how to organize them into the following formats:
Main part (what + subject) + predicate (how (when, where, what, why, how) + predicate (how))
Using symbols, it can be expressed as follows.
Main part (M + S) + Predicate ((When: Where: What: Why: How) + V)
M: What S: What, Who Who: When Where: Where What: What What: Why How: How V: What V: What, How How to make the machine learn.
In the initial stage, a human being teaches the type of sentence and the above-described method of organizing and makes the machine learn. The pattern of the word which comprises a sentence and the pattern which shows the attribute (a noun, a verb, an adjective, an adverb, and a particle) corresponding to a word are correlated and excited. In addition, after the excitation of each word and the corresponding attribute pattern, the position when the pattern is arranged in the above form (subject modification, subject, predicate modification, predicate), and further details on predicate modification ( Exciting patterns identified about when, where, what, why, how) and sentence types. In this way, the connection relationship between the pattern corresponding to the word appearing in the sentence and the attribute of the word and the position corresponding to each word and the pattern corresponding to the type of sentence is learned. As learning progresses to some extent, when a sentence pattern is input, the machine autonomously excites the pattern corresponding to the position of each corresponding sentence type and word. If the result is incorrect, the machine is notified that the person is wrong, and the connection with the incorrect pattern is weakened and the connection with the correct pattern is strengthened. The pattern processing is performed so that the pattern of the corresponding word is stored at the designated position using the excitation of the pattern corresponding to the position. Such learning allows the machine to identify sentence types and associate them with organized forms.
Next, a method for translation into another language will be described.
First, a pattern of a translation destination word corresponding to each word is searched. This is searched from a recording area storing a pattern indicating the pattern, attribute, nature, and usage of the word to be translated for each word. (The above-mentioned recording area can be constructed by generating a pattern-to-pattern connection from the word corresponding to the word dictionary to the information of the translated word.)
By the method explained above, the sentence type, main part (subject modification part + subject) + predicate (predicate modification part + predicate) can be identified, so the language pattern of the translation destination is the sentence type. In response to the above, pattern conversion to the arrangement of the translation destination language is performed.
Since the type of sentence and the constituent elements of the sentence are identified, a typical pattern can be converted into a pattern according to the grammar of the language to be translated. When typical conversion is difficult, a conversion example is prepared separately, and conversion between patterns from the current language to the translation destination language is performed.
The method of translation into other languages is also implemented by having the machine learn.
In the initial stage, a human teaches the machine to learn about the types of sentences and the arrangement of the translation target languages. When the pattern of the arrangement of the translation source language is detected, the pattern corresponding to the arrangement of the translation destination language is excited. Next, this pattern performs a process of arranging the translated word pattern at a predetermined position, and generates a pattern corresponding to the translated word sentence.
Depending on the word, it may be difficult to associate the words of the translated word with a sentence alone, but this is handled by considering the context (context) rather than processing the sentence alone. Learning to the machine is performed by presenting input patterns and specifying the corresponding output patterns to strengthen the connection between the patterns. By learning between patterns including the past of the history of patterns to be referred to, it is possible to learn connection between patterns in consideration of context, not associating single sentences.
Information on the context (context) is recorded in the connection relation recording unit of the recording module in FIG. Each time each recording module is excited, the excitation pattern history recorded in the pattern irradiator 7 is transferred from the excitation time of this recording module to the past set period. Alternatively, the connection relationship with the pattern related to excitation of the recording module is strengthened by using the recording module. This corresponds to considering the past context in the excitation of the recording module.

Means for solving the problem

(Corresponding to claim 4)

In addition to sequentially recording the input information, learn the problem solving method by human instruction and autonomous learning, analyze the situation, extract the problem, generate the solution of the problem, evaluate the solution and evaluate the solution A means for making selections and generating solutions autonomously for similar problems is described.
Have the machine learn how to analyze problems and extract problems, and how to set problem-solving goals. Further, a problem solving method for solving or improving the problem and a method for evaluating the effect when the solution is applied are learned. When there are a plurality of problem solving methods and it is necessary to select a solution from among them, the effect when applied is compared, and a method for selecting the best one is learned. In order to solve the problem, if information is necessary, the information is requested from the outside world, and the process is further advanced by obtaining the information.
The above is explained using patterns as follows.
Analyze problem patterns and identify what areas and what types of problems. In the early stages, humans are taught to let the machine learn about domain and problem type identification, and solutions corresponding to problem types. This is because when each problem pattern is input to the machine, it excites patterns related to the words, word meanings, features, attributes, etc. included in the problem, and then into the fields, problem types and problem types identified by humans. By sequentially exciting the patterns representing the corresponding solutions, connection relationships between the patterns are generated from the problem pattern to the field, the problem type and the solution pattern. For a number of problems, we will learn such connection relationships between patterns and strengthen the connection between patterns. As learning progresses to some extent, when a problem pattern is input, the machine autonomously excites the corresponding field, problem type, and solution pattern. If the result is incorrect, the machine is notified that the person is wrong, and the connection with the incorrect pattern is weakened and the connection with the correct pattern is strengthened. Such learning allows machines to identify problem areas, problem types, and solutions.
When the pattern learning is completed, the process after the problem is input is learned. When a problem pattern is input, a pattern corresponding to a field, a problem type, and a pattern expressing a problem solving target are excited to generate a connection relationship. Next, the pattern of the field and problem type is applied to the problem pattern to excite the pattern representing the problem. (By defining the relationship between the patterns, it is possible to define the inter-pattern coupling so that the pattern representing the result of the action is excited by exciting a certain input pattern and the pattern representing the action.) Then, a pattern that associates the solution with the pattern that represents the problem is caused to act, and the pattern that represents the solution is excited. By applying the pattern representing the solution to the pattern representing the problem, the pattern representing the state after the solution is applied to the problem is excited. Patterns are transitioned by such an action between patterns, and are brought closer to a problem solving pattern. The state in which the solution is applied to the problem is generated by identifying the pattern of each element constituting the pattern of the problem and changing the pattern of each element in response to applying the solution.
In solving the problem, if it is difficult to cope with only the information inside the machine, an additional required information pattern is excited. This can be done by learning what kind of information (internal information or external information?) Is necessary according to the process of solving each problem.
In this way, if information from the outside is required during the process, an additional information request is issued, and if necessary information is additionally input, the process can be performed using that information. Become. When there are multiple proposals for each solution proposal and it is necessary to select one of them, the impact (advantage, disadvantage) of implementing each solution is evaluated and selected. If more than one option is branched for each proposed solution, select the solution considering the impact of implementing the next step. Considering until the influence of the branching options can be judged, select an appropriate plan as a whole. In addition, there are proposals regarding solutions, and humans are notified of the evaluation results of each proposal and the reason for the selected proposal.

Embodiment 1 of the Invention

(Corresponding to claim 1)

FIG. 1 is a diagram showing the configuration of an autonomous intelligence generator in one embodiment of the present invention.
In FIG. 1, reference numeral 1 denotes a pattern converter that converts information into a pattern. Reference numeral 2 denotes a pattern recorder that records patterns, connection relationships between patterns, and relationships between patterns. Reference numeral 3 denotes a pattern register that registers and changes a pattern and a connection relation between the patterns by a human instruction or autonomously. Reference numeral 4 denotes a pattern controller for controlling pattern processing. Reference numeral 5 denotes a pattern inverse converter for converting a pattern into information. Reference numeral 6 denotes a pattern analyzer that analyzes patterns and relationships between patterns. 7 records the excitation history of the pattern, and when the pattern is excited, the connection history between the patterns related to the excitation is generated by transferring / accumulating or using the history of the previously excited pattern in the recording module. The pattern irradiator excites a recording module having a high correlation by collating data recorded in a part with the history data set from the latest history of the excitation pattern to the past.

Next, the operation will be described.
In FIG. 1, the pattern converter converts information into a pattern. The converted pattern is analyzed and divided into a main part and a predicate. The main part is divided into a part that modifies the subject and the subject, and the predicate is divided into a part that modifies the predicate and the predicate. Further, the part that modifies the predicate is identified and arranged as information corresponding to 5W1H. This arrangement can be performed by analyzing the words included in the sentence and the attributes of the words. FIG. 2 shows an example of typical arrangement. FIG. 2 shows how the information included in the sentence is organized and recorded as a pattern in the pattern information recording unit.
Human thought is expressed in language, but the information expressed in this language is converted into what is called a pattern. When converted into a pattern, human thoughts can be expressed as individual patterns, and human thought transitions can be viewed as changes from pattern to pattern. Further, the pattern is not only a language expression, but can express image information, for example, and can process information. Further, it is a concept that has a very wide range that can be handled, for example, it is possible to generate a signal for driving the driving device to perform the operation.
If information is converted into such a pattern and human thought is regarded as a pattern, human thought can be expressed as a change from pattern to pattern. When we convert the language that expresses human thoughts into patterns and analyze the patterns and the relationships between patterns, the types of sentences (normal sentences, question sentences, imperative sentences, etc.) and features (truths, truths, facts, definitions, rules, Common sense, explanation, hypothesis, prediction, opinion, impression).
Sentence types can be trained by machines so that humans can identify the types and characteristics of sentences. The sentence type can be identified as a normal sentence, a question sentence, a command sentence, etc. by analyzing words included in the sentence. Among the features of the sentence, truth, truth, facts, definitions, rules, and common sense are taught by humans for each individual and learned by the machine. This can be done by identifying and recording each type of information (truth, truth, fact, definition, rule, common sense) as each information is entered into the machine. Identification of other types of information (explanation, hypothesis, prediction, opinion, impression) is that the sentence type is not recorded as (truth, truth, fact, definition, rule, common sense) and the word ( I think, think, think, etc.) can be implemented by analyzing.
If the language is converted into a pattern and expressed as a set of patterns corresponding to the words that make up the sentence, and the semantic relationship between the words is expressed by the connection relationship between the patterns, the semantic relationship of the sentence It can be expressed by the connection relation of the pattern.
In addition, this connection relationship can express various relationships such as logical relationship, definition, attribution relationship, similar relationship, relationship between action and result, inference development, inheritance of attributes, common features, etc. Inheritance and identification of individual features can be expressed flexibly.
The newly entered information can be checked against the already recorded knowledge system and evaluated for consistency and novelty. Information that is identified as truth, truth, facts, definitions, rules and common sense and is not consistent with the information recorded is likely to be incorrect. On the other hand, if it is not possible to judge whether the information is consistent or not, it is considered that knowledge sufficient to determine the information has not yet been stored, and therefore human judgment is required. If humans can judge that the information is correct, the result will be notified to the machine. The machine records the information as an explanation or hypothesis, and obtains other information in the future, so that it can be logically expanded with a combination of truth, truth, facts, definitions, rules and common sense as evaluated information. It shall be identified and recorded.
In this way, information is converted into patterns, information types and characteristics are identified, classified and recorded, and information-information relationships (logical relationships, similar relationships, reciprocal relationships, analogies, etc.) Information can be constructed as a knowledge system by expressing the connection relationship between patterns.
When information is constructed as a knowledge system in this way, problem solving methods can be learned and knowledge can be improved. First, analyze the problem, clarify the problem, and set a goal for problem solving. Next, a solution plan for solving the problem is extracted and applied. 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, the above process is applied again to a new state, and the problem-solving goal is approached.
Problem analysis and problem clarification are implemented by learning a method of detecting and identifying the difference between the intermediate target set according to the target state or problem type and the current state. Moreover, the solution proposal corresponding to each problem is implemented by learning the corresponding processing.
The above is explained using patterns as follows.
Analyze problem patterns and identify what areas and what types of problems. In the early stages, humans are taught to let the machine learn about domain and problem type identification, and solutions corresponding to problem types. This is because when each problem pattern is input to the machine, it excites patterns related to words, word meanings, features, attributes, etc. that humans consider for problem identification, and then the fields and problem types identified by humans. And, by sequentially exciting patterns representing solutions corresponding to the problem types, connection relationships between patterns are generated from the problem patterns to the field, problem types and solution patterns. For a number of problems, we will learn such connection relationships between patterns and strengthen the connection between patterns. As learning progresses to some extent, when a problem pattern is input, the machine autonomously excites the corresponding field, problem type, and solution pattern. If the result is incorrect, the machine is notified that the person is wrong, and the connection with the incorrect pattern is weakened and the connection with the correct pattern is strengthened. Such learning allows machines to identify problem areas, problem types, and solutions.
When the pattern learning is completed, the process after the problem is input is learned. When a problem pattern is input, a pattern corresponding to a field, a problem type, and a pattern expressing a problem solving goal or an intermediate goal are excited to generate a connection relation. Next, the pattern of the field and problem type is applied to the problem pattern to excite the pattern representing the problem. (By defining the relationship between the patterns, it is possible to define the inter-pattern coupling so that the pattern representing the result of the action is excited by exciting a certain input pattern and the pattern representing the action.) Then, a pattern that associates the solution with the pattern that represents the problem is caused to act, and the pattern that represents the solution is excited. By applying the pattern representing the solution to the pattern representing the problem, the pattern representing the state after the solution is applied to the problem is excited. The pattern is changed by such an action between the patterns, and is brought closer to the problem solving pattern that satisfies the intermediate target and the target.
The state in which the solution is applied to the problem is generated by identifying the pattern of each element constituting the pattern of the problem and changing the pattern of each element in response to applying the solution.
Machine learning is performed by automatically recording a machine instruction by a human. In FIG. 1, when a pattern taught by a human is recorded and excited, a record of the excitation history is recorded in the pattern irradiator 7. When a person excites a pattern, the connection relationship between the patterns related to excitation is transferred or used by transferring or using the excitation history of the pattern excited before exciting the pattern in the connection relationship recording unit of the two pattern recording units. Strengthened. Thereby, when the history of the excitation pattern of the next pattern irradiator becomes the same form as this time, the correlation with this pattern is increased and excitation is automatically performed. Thus, the present invention can record and learn the process itself taught by a human. FIG. 3 shows a problem pattern, a field, a pattern of a problem type, a pattern that expresses a problem-solving target, a pattern that represents a problem, a pattern that represents a solution, and a pattern that represents a state after the solution is applied to the problem. Shows how the transitions.
It can be seen that by applying a pattern corresponding to an action to a certain pattern, another pattern is generated, and a pattern corresponding to the action is further applied to the pattern to bring it closer to the target pattern.
FIG. 4 is a generalization of the operation of FIG. FIG. 5 represents the operation of FIG. 4 in a block diagram.
There are two types of thought patterns when humans are trying to solve problems. One is a thought pattern corresponding to the problem itself, and the other is a thought pattern related to a procedure for solving the problem. When something goes wrong, humans examine the situation, analyze the situation, identify the cause of the problem, and think about ways to solve the cause of the problem. The thinking pattern of the procedure for solving the problem acts on the thinking pattern of the problem, and the thinking pattern is changed toward the problem solving. This process can be expressed using patterns as follows.
6). Define the problem Express the problem as a pattern (P (i))
7). Analyzing the problem Exciting the relevant pattern from the problem pattern and the pattern of the elements that make up the problem (if no difference from the state that should be the problem pattern alone is detected, an equivalent pattern, a logically derived pattern, (Exciting related patterns and revealing problematic patterns)
(Pdi)
8). Cause analysis Compare the pattern of the desired state and component with the above problem and the component of the problem, and identify the pattern corresponding to the different element (Ppi)
9. Selection of treatment Excites a pattern corresponding to the policy to eliminate the above-mentioned differences in question. (The policy to solve the problem is to change the pattern of the problem based on the procedure and rule of thumb for solving the problem. Generate the pattern to be used.)
(Pai)
10. Eliminate the cause of the problem. Apply the treatment pattern to the problem pattern. (By applying the treatment pattern, the pattern corresponding to the element in question is eliminated.)

Embodiment 2 of the Invention

(Corresponding to claim 2)

Learn how to learn mathematical and physics mathematical processing and chemical formula processing by human instruction and autonomous learning and generate answers autonomously for mathematical, physical and chemical problems.
In FIG. 1, the pattern converter converts information into a pattern. The converted pattern is analyzed, and numbers, constants, variables, signs, equal signs, inequality signs, functions, operators (differentiation, integration), atomic symbols, reaction directions, molecules, molecular structures, etc. are identified. The input sentence is divided into a main part and a predicate. The main part is divided into a part that modifies the subject and the subject, and the predicate is divided into a part that modifies the predicate and the predicate. Further, the part that modifies the predicate is identified and arranged as information corresponding to 5W1H.
(Since sentence processing is described in [0017] and [0018], other processes will be described below.
First, the formula processing method will be described.
The input formula is identified from the beginning, numbers, constants, variables, signs (±), equal signs, inequality signs, functions, operators (differentiation, integration) are identified, and patterns are grouped immediately before the sign. Separate them. Such a group of patterns is generated and assigned to the pattern of the first term on the left side, the pattern of the second term on the left side, and so on. When an equal sign or inequality sign pattern is detected, a group of patterns is generated as the first group on the right side, the second term on the right side, and so on.
Patterns from the first term on the left side up to immediately before the equal sign or inequality sign are grouped into a left side pattern. Patterns of the first term on the right side, the second term on the right side,... By such assignment, the corresponding pattern can be called by designating the i-th term on the left side, the j-th term on the right side, the left side, and the right side.
FIG. 6 shows a state where mathematical expressions are recorded in the recording module of the pattern recorder.
Next, regarding the arrangement order of each item, it is rearranged and arranged in the order of symbols, numbers, variables, and functions. For the operators, the range of the object is identified, and the order of arrangement is arranged for the object and the object of the operator.
For each item, the elements that make up the item (numbers, constants (a, b, c,...), Variables (x, y, z,...),
Simplify by calculation.
These processes can be performed by processing the pattern.
The pattern processing is generally expressed as follows.
Input pattern: Pin
Action pattern: Pf
Output pattern: Pout
Pin + Pf ⇒ Pout
(It means that Pout is generated by acting Pf on Pin.)
Corresponding Pout can be obtained by variously combining the Pin pattern and the Pf pattern. Next, a pattern for analyzing a pattern and a pattern for extracting a problem as a result of the analysis are defined.
Analysis pattern: Pd
As a result of analysis, a pattern indicating the extracted problem: Pp
In addition, a solution pattern for the problem is defined.
Solution pattern for the problem: Pa
As described above, the pattern is defined and applied to the pattern corresponding to the mathematical expression.
If the pattern corresponding to the mathematical expression is Pmat (i). Processing is as follows.
The above processing is continued until the problem is solved.
As an example, FIG. 7 shows an example of processing for organizing mathematical expressions. FIG. 8 is a transition diagram showing pattern changes corresponding to the processing of FIG.
Similarly, FIG. 9 shows a process for solving a quadratic equation by this method as an example of mathematical expression processing.
FIG. 10 and FIG. 11 relate to processing of mathematical expressions, and show examples of functions, operators, and solutions of equations (differential equations). In these methods, by detecting a pattern of a mathematical formula and exciting a pattern (equivalent type, solution, etc.) that is strongly associated with the pattern, the knowledge of mathematical formula processing can be expressed by a change from pattern to pattern.
FIG. 12 shows an example of solving a physical (dynamic) problem as an example of mathematical expression processing.
As a problem, consider a system in which the mass M is fixed to a spring and is connected to an attenuator that provides an attenuation proportional to the relative velocity. In such a system, the motion of the mass M is solved.
Examples of solutions to such problems are shown in FIGS.
Associating a problem sentence as an input condition pattern, analyzing this problem sentence, detecting a difference from the target and intermediate target (solution type) as a problem, and applying the corresponding solution pattern to the problem To change the state. A pattern for extracting a problem from input conditions and an action pattern corresponding to a solution from the problem extraction pattern are taught in advance by a human as a method of solving a dynamic problem. By registering these patterns in advance, the problem extraction pattern and the action pattern are sequentially excited for each state, and the pattern to a desired state (solution) is changed.
The goal and intermediate goal for dynamical problems is to approximate the pattern of force relationships, equations of motion, solutions to equations of motion (x = f (t) type).
FIG. 18 shows a pattern transition corresponding to an example of a physics (dynamic problem) solution. FIG. 19 shows the transition of this pattern using a configuration diagram.
FIG. 20 shows an example in the case of dealing with a chemical related problem. The pattern corresponding to the atomic / molecular formula is recorded in the recording module.
FIG. 21 shows systematic separation of metal ions as an example of chemical formula treatment. It shows how metal ions are sequentially separated by applying an action pattern to an input pattern. FIG. 22 represents the systematic separation of FIG. 21 by pattern transition. FIG. 23 shows the production of phenol, which is an organic compound, as a chemical formula treatment example. The figure shows that phenol and acetone can be produced by treatment from benzene and propene. FIG. 24 shows this generation process by pattern transition.
In FIG. 25, the transition process of this pattern is shown using a block diagram.
FIG. 26 expresses definitions of chemical reaction formulas, features, and the like using pattern transition diagrams. As can be seen from the figure, it is possible to express chemical reaction formula definitions, characteristics of each substance and group, groups and elements, inclusion relations, and the like. As the relationship between patterns, various relationships can be defined as shown in FIGS. FIG. 27 shows logical relationships, definitions, attribution relationships, and similar relationships. FIG. 28 shows the relationship between attribute inheritance, action, and result. FIG. 29 shows inference development, common feature inheritance, and individual feature identification.
FIG. 30 shows the relationship of FIG. 26 using a configuration diagram. As described above, the chemical formula, the chemical reaction formula, the feature of the substance, the relation between the substance and the group, and the relation between the group and the group can be expressed by the pattern and the connection relation between the patterns. Individual information is recorded in association with related information, not individually. Since information can be constructed as a knowledge system in this way, various analysis, analysis, inference, etc. are possible.

Embodiment 3 of the Invention

(Corresponding to claim 3)

A method for learning a translation method into another language by human instruction and autonomous learning and generating a translation into another language autonomously will be described.
FIG. 31 shows a state in which languages are organized into a main part, a predicate, a subject, a predicate, and a modifier and recorded in the recording module of the pattern recorder. Similarly, the language of the translation destination is organized and recorded in the main part and predicate. Arranging in this way clarifies the type of information contained in the sentence, and also clarifies the structure of the sentence, which is useful when translating between languages.
FIG. 32 and FIG. 33 show an outline of processing when translating to another language.
Using word attributes, features, etc. from the word sequence corresponding to the sentence, subject, predicate, modifier identification, relation between words (modification relation, etc.) and sentence type (normal sentence, question sentence, imperative sentence) , Exclamation, etc.) will be conducted sequentially.
For translation into other languages, identification is made as to whether the input sentence is an idiom or non-idiom. When translating, it is appropriate to use idioms in correspondence with specific translations rather than conversion according to grammar. For sentences other than idioms, processing proceeds by sequentially performing conversions according to sentence types, grammar, and the like.
34 to 37 show detailed examples of the translation processing into other languages. The sentence structure is clarified while analyzing the input sentence and identifying the subject, predicate, modifier and relationship between words. When the sentence structure is identified, conversion between words and word placement are performed. For conversion between words, search is performed from a recording area storing patterns indicating patterns, attributes, properties, and usage methods of words to be translated for each word. (The above-mentioned recording area can be constructed by generating a pattern-to-pattern connection from the word corresponding to the word dictionary to the information of the translated word.)
The word arrangement is processed using the sentence structure and grammar. Since the type of sentence and the constituent elements of the sentence are identified, a typical pattern can be converted into a pattern according to the grammar of the language to be translated. When typical conversion is difficult, a conversion example is prepared separately, and conversion between patterns from the current language to the translation destination language is performed.
Furthermore, analysis of the translated text (draft) after conversion is performed, and it is confirmed whether the translation text conforms to the grammar using the pattern indicating the word and the attribute of the word. Make corrections.
The method of translation into other languages is also implemented by having the machine learn.
In the initial stage, a human teaches the machine to learn about the types of sentences and the arrangement of the translation target languages. When the pattern of the arrangement of the translation source language is detected, the pattern corresponding to the arrangement of the translation destination language is excited. Next, this pattern performs a process of arranging the translated word pattern at a predetermined position, and generates a pattern corresponding to the translated word sentence.
Depending on the word, it may be difficult to associate the words of the translated word with a sentence alone, but this is handled by considering the context (context) rather than processing the sentence alone. Learning to the machine is performed by presenting input patterns and specifying the corresponding output patterns to strengthen the connection between the patterns. By learning between patterns including the past of the history of patterns to be referred to, it is possible to learn connection between patterns in consideration of context, not associating single sentences.
Information on the context (context) is recorded in the connection relation recording unit of the recording module in FIG. Each time each recording module is excited, the excitation pattern history recorded in the pattern irradiator 7 is transferred from the excitation time of this recording module to the past set period. Alternatively, the connection relationship with the pattern related to excitation of the recording module is strengthened by using the recording module. This corresponds to considering the past context in the excitation of the recording module.
The above processing can be performed by processing using the pattern transition shown in FIG.

Embodiment 4 of the Invention

(Corresponding to claim 4)

In addition to sequentially recording the input information, learn the problem solving method by human instruction and autonomous learning, analyze the situation, extract the problem, generate the solution of the problem, evaluate the solution and evaluate the solution A means for making selections and generating solutions autonomously for similar problems is described.
Have the machine learn how to analyze problems and extract problems, and how to set problem-solving goals. Further, a problem solving method for solving or improving the problem and a method for evaluating the effect when the solution is applied are learned. When there are a plurality of problem solving methods and it is necessary to select a solution from among them, the effect when applied is compared, and a method for selecting the best one is learned. In solving the problem, if information is necessary, the information is requested from the outside, and the process is further advanced by obtaining the information.
The above is explained using patterns as follows.
Analyze problem patterns and identify what areas and what types of problems. In the early stages, humans are taught to let the machine learn about domain and problem type identification, and solutions corresponding to problem types. This is because when each problem pattern is input to the machine, it excites patterns related to the words, word meanings, features, attributes, etc. included in the problem, and then into the fields, problem types and problem types identified by humans. By sequentially exciting the patterns representing the corresponding solutions, connection relationships between the patterns are generated from the problem pattern to the field, the problem type and the solution pattern. For a number of problems, we will learn such connection relationships between patterns and strengthen the connection between patterns. As learning progresses to some extent, when a problem pattern is input, the machine autonomously excites the corresponding field, problem type, and solution pattern. If the result is incorrect, the machine is notified that the person is wrong, and the connection with the incorrect pattern is weakened and the connection with the correct pattern is strengthened. Such learning allows machines to identify problem areas, problem types, and solutions.
When the pattern learning is completed, the process after the problem is input is learned. When a problem pattern is input, a pattern corresponding to a field, a problem type, and a pattern expressing a problem solving target are excited to generate a connection relationship. Next, the pattern of the field and problem type is applied to the problem pattern to excite the pattern representing the problem. (By defining the relationship between the patterns, it is possible to define the inter-pattern coupling so that the pattern representing the result of the action is excited by exciting a certain input pattern and the pattern representing the action.) Then, a pattern that associates the solution with the pattern that represents the problem is caused to act, and the pattern that represents the solution is excited. By applying the pattern representing the solution to the pattern representing the problem, the pattern representing the state after the solution is applied to the problem is excited. Patterns are transitioned by such an action between patterns, and are brought closer to a problem solving pattern. The state in which the solution is applied to the problem is generated by identifying the pattern of each element constituting the pattern of the problem and changing the pattern of each element in response to applying the solution.
In solving the problem, if it is difficult to cope with only the information inside the machine, an additional required information pattern is excited. This can be done by learning what kind of information (internal information or external information?) Is necessary according to the process of solving each problem.
In this way, if information from the outside is required during the process, an additional information request is issued, and if necessary information is additionally input, the process can be performed using that information. Become. When there are multiple proposals for each solution proposal and it is necessary to select one of them, the impact (advantage, disadvantage) of implementing each solution is evaluated and selected. If more than one option is branched for each proposed solution, select the solution considering the impact of implementing the next step. Considering until the influence of the branching options can be judged, select an appropriate plan as a whole. In addition, there are proposals regarding solutions, and humans are notified of the evaluation results of each proposal and the reason for the selected proposal.
FIG. 38 shows an example of processing when it is difficult to process only the internal information when processing the problem, and external information is required. When a solution pattern corresponding to the problem is excited in the figure, this pattern makes a necessary information request to the outside. When information is provided from the outside, this pattern determines the contents of the process, and the process is applied to the problem pattern as an action pattern. In this way, it is possible to proceed with processing while requesting external information provision as necessary.
FIG. 39 shows the pattern transition of FIG. 38 using a configuration diagram. In the figure, when the pattern A1 is excited, an information request is issued. The processing module of the pattern controller 4 detects the information request and generates an information request to the outside. The information request is converted from a pattern to information (language) by a pattern inverse converter 5. When the requested information is input from one pattern converter, the processing module stores the information and irradiates the second pattern recorder with the information pattern. The pattern A1 that issued the information request determines the processing based on the irradiated information and acts as an action pattern. (Pattern A1 is composed of a detailed pattern, and the internal pattern is excited in accordance with the irradiated information, and the process to be selected is thereby determined.)

Effect 1 of the Invention

(Corresponding to claim 1)

According to the first invention, it is possible to learn a problem solving method by human instruction and autonomous learning, and autonomously generate an answer to a similar problem.
Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. Also, it was difficult to learn how to solve problems and generate answers autonomously for similar problems.
According to the present invention, it is not necessary to design and implement a program or the like for information processing, so that development effort can be greatly reduced.

Effect 2 of the invention

(Corresponding to claim 2)

According to the second aspect of the present invention, mathematics and physics formula processing and chemical formula processing can be learned by human instruction and autonomous learning, and answers can be autonomously generated for mathematics, physics and chemistry problems.
Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. In addition, it has been difficult to learn mathematical and physical formula processing and chemical formula processing and autonomously generate answers to mathematical, physical and chemical problems.
According to the present invention, it is not necessary to design and implement a program or the like for information processing, so that development effort can be greatly reduced.

Effect 3 of the invention

(Corresponding to claim 3)

According to the third invention, it is possible to learn a translation method into another language and autonomously generate a translation into another language.
Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. In addition, it was difficult to learn how to translate into other languages and autonomously generate translations into other languages.
According to the present invention, it is not necessary to design and implement a program or the like for information processing, so that development effort can be greatly reduced.

Advantage 4 of the Invention

(Corresponding to claim 4)

According to the fourth invention, a problem solving method is learned, a situation analysis, a problem extraction, a problem solution proposal generation, a solution proposal evaluation and a solution selection are performed, and similar problems are analyzed. Solution can be generated autonomously.
Conventionally, when a machine performs processing of input information, etc., it has been necessary to set a program in the computer in advance. It was necessary to create a program for judging the situation from the input information and a program for causing the machine to operate according to individual conditions, and to install and execute it on a computer mounted on the machine. The program had to be created in a dedicated programming language, and there was a demerit that it took a lot of time for development. If the detection of the condition and the corresponding operation are not appropriate, there is a demerit that it is necessary for a person to correct the program installed in the computer, and it takes a long time for the correction. In addition, it was difficult to learn how to translate into other languages and autonomously generate translations into other languages. It also learns how to solve problems, analyzes the situation, extracts problems, generates solutions for problems, evaluates solutions, and selects solutions, and solves similar problems autonomously. It was difficult to produce.
According to the present invention, it is not necessary to design and implement a program or the like for information processing, so that development effort can be greatly reduced.

Configuration example of autonomous intelligence generator Information is organized and recorded as a pattern Pattern transition for problem solving (expressed as pattern transition) Pattern transition toward problem solving (Pattern transition is generalized) Pattern transition toward problem solving (expressed in configuration diagram) Recording mathematical expressions as patterns Example of processing to organize mathematical formulas Example of processing to organize mathematical expressions (expressed by pattern transition) Example of processing to solve mathematical formula (secondary equation) Formula processing example (function, operator) Formula processing example (differential equation) Physics (dynamics) problems Solving examples of physics (dynamics) problems (1/5) Solving examples of physics (dynamics) problems (2/5) Solving examples of physics (dynamics) problems (2/5) Solving examples of physics (dynamics) problems (2/5) Solving examples of physics (dynamics) problems (2/5) Solving examples of physics (dynamics) problems (represented by pattern transitions) Example of solving a physics (dynamics) problem (represented in a block diagram) Recording chemical formula as a pattern Example of systematic separation of metal ions Example of systematic separation of metal ions (expressed by pattern transition) Example of organic compound (phenol) production Example of organic compound (phenol) production (expressed by pattern transition) Example of organic compound (phenol) production (expressed in configuration diagram) Definition of chemical reaction formulas and features (represented by pattern transitions) Relationships between patterns (logical relationships, definitions, attribution relationships, similar relationships) Relationship between patterns (attribute inheritance, relationship between action and result) Relationships between patterns (development of inference, inheritance of common features and identification of individual features) Definition of chemical reaction formulas, features, etc. (expressed in a configuration diagram) Information is organized and recorded as a pattern (after Japan, other languages) Example of translation into other languages (outline) (1/2) Example of translation into another language (outline) (2/2) Example of translation into other languages (details) (1/4) Example of translation into another language (details) (2/4) Example of translation into another language (details) (3/4) Example of translation into another language (details) (4/4) Example of problem solving that requires external information (expressed by pattern transition) Example of problem solving that requires external information (expressed in the configuration diagram)

1 Pattern Converter 2 Pattern Recorder 3 Pattern Register 4 Pattern Controller 5 Pattern Inverse Converter 6 Pattern Analyzer 7 Pattern Irradiator

Claims (4)

  A pattern converter that converts information into a pattern, a pattern recorder composed of a pattern, a connection between patterns, and a recording module that records the relationship between patterns, and a pattern or a connection between patterns. Pattern registerer to register and change automatically, pattern controller to control pattern processing, record pattern excitation history, and when pattern is excited, transfer / accumulate the history of previously excited pattern Use to create a connection relationship between patterns related to excitation and collate the data recorded in the part in the recording module with the latest history of the excitation pattern and the history data set up to the past, and record with high correlation A pattern irradiator that excites the module, a pattern inverse converter that converts the pattern into information, and a pattern And includes a pattern analyzer for analyzing the relationship between the patterns, the human instruction and autonomous learning learns problem solving method, artificial intelligence and software to generate autonomously respond to similar problems.   A pattern converter that converts information into a pattern, a pattern recorder composed of a pattern, a connection between patterns, and a recording module that records the relationship between patterns, and a pattern or a connection between patterns. Pattern registerer to register and change automatically, pattern controller to control pattern processing, record pattern excitation history, and when pattern is excited, transfer / accumulate the history of previously excited pattern Use to create a connection relationship between patterns related to excitation and collate the data recorded in the part in the recording module with the latest history of the excitation pattern and the history data set up to the past, and record with high correlation A pattern irradiator that excites the module, a pattern inverse converter that converts the pattern into information, and a pattern And a pattern analyzer that analyzes the relationship between patterns, learns mathematical and physical formula processing and chemical formula processing by human instruction and autonomous learning, and generates answers autonomously for mathematical, physical and chemical problems Artificial intelligence and software.   A pattern converter that converts information into a pattern, a pattern recorder composed of a pattern, a connection between patterns, and a recording module that records the relationship between patterns, and a pattern or a connection between patterns. Pattern registerer to register and change automatically, pattern controller to control pattern processing, record pattern excitation history, and when pattern is excited, transfer / accumulate the history of previously excited pattern Use to create a connection relationship between patterns related to excitation and collate the data recorded in the part in the recording module with the latest history of the excitation pattern and the history data set up to the past, and record with high correlation A pattern irradiator that excites the module, a pattern inverse converter that converts the pattern into information, and a pattern And it includes a pattern analyzer for analyzing the relationship between the patterns, human instruction and by autonomous learning to learn how to translate into other languages, autonomously generate artificial intelligence and software translation into another language.   A pattern converter that converts information into a pattern, a pattern recorder composed of a pattern, a connection between patterns, and a recording module that records the relationship between patterns, and a pattern or a connection between patterns. Pattern registerer to register and change automatically, pattern controller to control pattern processing, record pattern excitation history, and when pattern is excited, transfer / accumulate the history of previously excited pattern Use to create a connection relationship between patterns related to excitation and collate the data recorded in the part in the recording module with the latest history of the excitation pattern and the history data set up to the past, and record with high correlation A pattern irradiator that excites the module, a pattern inverse converter that converts the pattern into information, and a pattern And a pattern analyzer that analyzes the relationship between patterns, learns problem solving methods by human instruction and autonomous learning, analyzes the situation, extracts problems, generates problem solutions, Artificial intelligence and software that performs evaluation and solution selection and autonomously generates solutions to similar problems.