An interactive mind map knowledge base device

Abstract

The application discloses an interactive thought map knowledge base device, which can be used for constructing a fusion type knowledge base based on natural language and without limitation of language types, and can be used as a component module of other artificial intelligence systems to provide a knowledge representation or a function of answering questions based on reasoning. The application allows knowledge and thought methods to be represented based on natural language texts, has strong representation capability, and is easy to understand and use. The application supports progressive fusion of knowledge represented based on different languages. The application supports establishment of a large knowledge base in a progressive fusion manner. Knowledge and thought methods are cooperatively represented in a mutually related manner, reasoning calculation results can further drive new reasoning calculation processes, autonomous driving thinking is realized, new information can be actively acquired on demand during the reasoning calculation process, and the reasoning calculation process is based on facts when the thought map content in the knowledge base is consistent with facts, and the reasoning calculation results also necessarily reflect the facts.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence and natural language processing, and is a method for representing knowledge and thinking methods. The knowledge and thinking methods represented based on the method can be used for reasoning and answering questions. Specifically, it is an interactive thinking map knowledge base device. BACKGROUND

[0002] In order to establish an artificial intelligence system that can think like a human being, it is necessary to model and represent the knowledge and thinking methods of human beings. The commonly used knowledge modeling method is knowledge graph, which represents knowledge in the form of triples. Although the knowledge graph structure is normative and suitable for queries, this representation does not conform to the understanding habits of ordinary people, and it is difficult for ordinary people to quickly and accurately understand complex knowledge represented by a series of triples as a whole, making it very difficult to create, review and design computational methods for processing knowledge based on triples. On the other hand, the knowledge graph is a representation method of knowledge rather than thinking methods. The schema technology based on knowledge graph technology can represent some very limited thinking methods, such as the transitivity of relations, but it still lacks effective representation ability for complex thinking methods.

[0003] On the other hand, the commonly used modeling method for thinking methods is computer program language. Computer program language also does not conform to the understanding habits of ordinary people, and it needs to be mastered through special training. It is very difficult for different natural persons to communicate thinking methods through computer program language, and it is difficult for a person to understand a complex program written by another person without annotations or incomplete annotations, making it very difficult to carry out large-scale integration and dynamic adjustment of thinking methods based on computer program language.

[0004] At the same time, thinking methods are also a kind of knowledge. However, it is difficult for computer program language to be represented by knowledge graph, and it is also difficult to be created or modified based on knowledge graph. Therefore, we still lack a technology that allows us to represent knowledge and thinking methods collaboratively; at this time, the knowledge and thinking methods obtained by the artificial intelligence system based on known knowledge and thinking methods can be represented by the same technology; thus, the results of reasoning and calculation can further drive new reasoning and calculation processes, and the artificial intelligence system can achieve self-driven thinking.

[0005] The thinking method can generate some knowledge from some knowledge. The structure of knowledge and thinking method connected with each other can be abstracted as a thinking map, in which knowledge is regarded as a node and thinking method is regarded as an edge. In this sense, the natural language used by human is essentially a representation method of thinking map, and people fuse thinking map through language communication: the speaker encodes a limited range of thinking map into natural language, and the listener fuses the thinking map obtained by decoding with the thinking map in his own brain to obtain the information conveyed by the speaker.

[0006] The neural network technology can establish a structure-unexplainable thinking map based on the knowledge represented by natural language. A representative technology is ChatGPT. This technology establishes a thinking map model based on a large amount of training text, and the model can calculate a text sequence as a response based on the user's prompt word. ChatGPT fills in the most probable word at each position when generating a response text; therefore, this technology is essentially based on the prompt word to calculate a semantic representation of possibility, and the obtained semantic representation may not conform to the fact. In the field of auxiliary design, the maximum probability response is a good feature; but in the field of auxiliary decision-making, it is more important to generate a response that conforms to the fact.

[0007] The patent discloses a method for constructing a thinking map in an interpretable way, and a knowledge base device established based on the method. The knowledge base device can interact with the user based on natural language text, create and maintain a knowledge base organized in the form of a thinking map. The user can ask the knowledge base device questions, and the knowledge base device performs reasoning calculation with the goal of answering the questions. In the process of reasoning calculation, other task execution systems or data query systems can be operated as needed to dynamically obtain knowledge describing the results of task execution or specific data. The knowledge base device answers the user's questions in the form of natural language text.

[0008] To realize the knowledge base device, the inventor found that classical philosopher Kant divides sentences into two categories, one is synthetic sentence, and the other is analytic sentence. The meaning of the predicate of a synthetic sentence is not contained in the subject, therefore, the truth of a synthetic sentence depends on the context in which it is placed; in possible world semantics, the truth of a synthetic sentence depends on the world in which it is placed; a synthetic sentence is used to state factual knowledge; the meaning of the predicate of an analytic sentence is contained in the subject, therefore, the truth of an analytic sentence depends only on its semantic structure, and is independent of the context in which it is placed; in possible world semantics, the truth of an analytic sentence is the same in all worlds. On this basis, the inventor further found that a specific instance of an analytic sentence based on a specific context can be used to state a method of thinking. Here, the method of thinking refers to a method of reasoning from one fact to another fact. For example, "Xiaozhang's blood pressure is 150" and "Xiaozhang has high blood pressure" are synthetic sentences, which state facts; "If a person's blood pressure is a value that can be determined as high blood pressure, then he has high blood pressure" is an analytic sentence; based on the context "a blood pressure value of 150 can be determined as high blood pressure", a specific instance of the analytic sentence is "if a person's blood pressure is 150, then he has high blood pressure"; at this time, the meanings of "high blood pressure" and "150" are the meanings of them in the context synthetic sentence "a blood pressure value of 150 can be determined as high blood pressure", and the specific instance of the analytic sentence has the sufficient and necessary conditions for the truth of the analytic sentence itself and its context. The specific instance of the analytic sentence "if a person's blood pressure is 150, then he has high blood pressure" states a method of thinking, and an application of the method of thinking is to reason from the aforementioned synthetic sentence "Xiaozhang's blood pressure is 150" to the aforementioned synthetic sentence "Xiaozhang has high blood pressure". If there is a reasoning relationship between two synthetic sentences, the invention refers to the synthetic sentence as the basis of reasoning as implying the synthetic sentence as the result of reasoning; therefore, the essence of applying the method of thinking is to obtain the implication relationship between the synthetic sentences: if the synthetic sentence as the result of reasoning is not known, the reasoning result synthetic sentence is created based on the method of thinking and the basis of reasoning synthetic sentence. Conversely, if there is an implication relationship between two synthetic sentences that state facts, there must be a specific instance of an analytic sentence, and the method of thinking represented by the specific instance can obtain the implication relationship between the two synthetic sentences.

[0009] Based on the above findings, the invention designs a knowledge base device, the constituent units of which cooperatively realize the representation of synthetic sentences, i.e. the representation of factual knowledge, the representation of specific instances of analytic sentences, i.e. the representation of methods of thinking, and the representation of the way of establishing specific instances of analytic sentences based on synthetic sentences as context; these representations together constitute the representation of a complete thinking map; the knowledge base device performs target-oriented and self-driven thinking according to the thinking map. SUMMARY

[0010] The patent discloses an interactive mind map knowledge base device, which can be used to construct a fusion knowledge base based on natural language and not limited by language; or as a component module of other artificial intelligence systems, providing knowledge representation or reasoning-based question answering functions. The present application is realized by the following technical solutions:

[0011] An interactive mind map knowledge base device for storing knowledge, comprising a mind map module, which specifically includes the following types of units;

[0012] Syntax unit, simply referred to as syntax; syntax stores the common semantic structure of integrated sentences with the same semantic structure; a semantic structure represents a specific relationship between a group of unspecified concepts; a semantic structure has a group of concept placement positions; when a concept is placed in a specific position of a semantic structure, it is said that the concept appears in the position; for example, "1+2=3" is an integrated sentence conforming to the addition semantic structure, and it conforms to the common semantic structure with "4+5=9", which has three concept placement positions, the first two of which are the positions of the addends, and the last one is the position of the sum; in the above two integrated sentences, the concepts "1" and "4" appear in the first concept placement position, the concepts "2" and "4" appear in the second concept placement position, and the concepts "3" and "9" appear in the third concept placement position; the above addition semantic structure can be understood as a form of "?+?=?", which includes three concept placement positions;

[0013] Integrated unit, simply referred to as integrated; integrated stores knowledge explained by integrated sentences with authenticity; when an integrated unit stores knowledge explained by an integrated sentence, the integrated unit stores the semantic structure of the integrated sentence, the concepts appearing in the concept placement positions in the semantic structure, and the context of the integrated sentence; it is said that the integrated unit represents the integrated sentence;

[0014] Syntax equivalence unit, simply referred to as syntax equivalence; syntax equivalence stores an association between two semantic structures, the meaning of the association is that there is a one-to-one correspondence between the concept placement positions in the two semantic structures, and when the concepts appearing in the corresponding positions are the same, the two semantic structures and the concepts appearing in the concept placement positions represent the same semantics as a whole;

[0015] Substantive equivalence unit, simply referred to as substantive equivalence; the knowledge base device uses codes to represent concepts; substantive equivalence stores an association between two codes representing concepts, the meaning of the association is that the two codes represent the same concept;

[0016] Analogy isomorphism unit, simply analogy isomorphism; analogy isomorphism stores a relation between two integrated statements, the meaning of the relation is that the two integrated statements have the same context; when all integrated statements in the usage scenario of the knowledge base device have the same context, the knowledge base device chooses not to use analogy isomorphism; at this time, it is equivalent to specifying that all integrated statements represented by the integrated type in the thinking map module have the same context, and the integrated type no longer stores the context of the integrated statement.

[0017] As a further improvement, the knowledge base device is also used to store thinking methods; the thinking map module also includes the following types of optional units as needed;

[0018] Analysis unit, simply analysis; analysis alone or in cooperation with the integrated analysis type stores the specific instance description of the analysis statement with the truth; when an analysis stores a thinking method, the analysis stores the application scenario of the thinking method and the process description of the reasoning result obtained by applying the thinking method;

[0019] Integrated analysis type unit, simply integrated analysis type; integrated analysis type stores a relation between an integrated statement and an analysis statement, the meaning of the relation is that the integrated statement is the context required for the analysis statement to be transformed into a specific instance of the analysis statement;

[0020] Predicate unit, simply predicate; predicate stores the relation between an integrated statement and its subject part. When a predicate stores the relation between an integrated statement and its subject part, the predicate stores: in the integrated type representing the integrated statement, the position of the concept corresponding to the subject of the integrated statement in the semantic structure.

[0021] As a further improvement, a natural language statement can be understood as a sequence of concepts contained therein arranged in the order of their appearance, simply referred to as concept sequence; when the knowledge base device creates various types of units of the thinking map module based on the concept sequence represented by the natural language statement, various types of units of the thinking map module store the knowledge and thinking methods explained by the natural language statements corresponding to these concept sequences in the following way;

[0022] The syntax stores the common semantic structure of the integrated sentences with the same semantic structure, and is realized as follows: if there is a group of integrated sentences, the concept sequences represented by the integrated sentences are of the same length, the concepts at specific positions are different, and the concepts at other positions are the same, then the integrated sentences conform to the common semantic structure; the positions that may be different in the concept sequences are called variable positions; the variable positions are named as the "1st", "2nd",..., and "nth" variable positions according to the order of their appearance in the concept sequences; the group of variable positions are the positions for placing the concepts in the semantic structure; the core attributes stored by the syntax include: a group of variable positions; for the units of the same type and other units of the thinking map module, if there is a group of variable positions in each of the two units of the thinking map module, and the groups of variable positions correspond to each other, then the number of the two groups of variable positions is equal, the names of the variable positions are the same, and the variable positions with the same name have a corresponding relationship;

[0023] When the integrated form stores the knowledge explained by an integrated sentence, the integrated form stores the semantic structure of the concept sequence represented by the integrated sentence, the context of the integrated sentence, and the concepts appearing at the variable positions of the semantic structure; at this time, the integrated form has a group of variable positions, the group of variable positions correspond to the variable positions of the syntax storing the semantic structure, and the concepts appearing at the variable positions of the semantic structure also appear at the corresponding variable positions of the integrated form; in the units of the same type and other units of the thinking map module, the concepts appearing at the variable positions are sorted according to the names of the variable positions, and the obtained sequence is called a variable sequence; at this time, the sequence position of a concept in the variable sequence is also the name of the variable position where the concept appears, so "the variable position where a concept appears" and "the sequence position of the concept in the variable sequence" represent the same meaning; the core attributes stored by the integrated form include: a syntax, a context, a group of variable positions, and a variable sequence; the context attribute of the integrated form can be left blank when created, and then the context of the integrated form is inferred according to the identical context relationship of the integrated forms described by the analogy form; when the units of the same type and other units of the thinking map module are implemented, the variable position attribute can be inferred from the variable sequence attribute, that is, the variable position attribute includes a number equal to the length of the variable sequence, and the names of the variable positions are "1st",..., and "nth" in turn, and the variable position attribute is no longer stored separately;

[0024] The syntax equivalence stores a relationship between two semantic structures, the meaning of the relationship is that there is a one-to-one correspondence between the variable positions in the two semantic structures, when the concepts appearing in the corresponding positions are the same, the two semantic structures and the concepts appearing in the positions where the concepts are placed as a whole represent the same semantics, which is realized as follows: if there are two semantic structures, called semantic structure form one and semantic structure form two, there is a one-to-one correspondence between the variable positions; at this time, for any pair of concept sequences, if the first one conforms to the semantic structure form one, the second one conforms to the semantic structure form two, and the concepts appearing in the corresponding variable positions are the same, then they must represent the same semantics; at this time, it is said that the semantic structure form one and the semantic structure form two have equivalent meanings; it is said that the syntax form storing the semantic structure form one and the semantic structure form two is equivalent syntax form one and equivalent syntax form two respectively; at this time, the core attributes stored by the syntax equivalence include: equivalent syntax form one, equivalent syntax form two, one-to-one correspondence between the variable positions of equivalent syntax form one and equivalent syntax form two; it is said that there is a syntax equivalence relationship between equivalent syntax form one and equivalent syntax form two; the syntax equivalence relationship has transitivity, and all syntax forms connected by the syntax equivalence relationship are equivalent; for the sake of simplicity, the same or equivalent syntax forms are simply referred to as syntax equivalent forms;

[0025] The substantial equivalence stores a relationship between two codes representing concepts, the meaning of the relationship is that the two codes represent the same concept, which is realized as follows: the knowledge base device uses codes to represent concepts, if two codes represent the same concept, then the two codes are called form codes of the concept, simply form codes; if a concept corresponds to a unique code, then the code is called the substantial code of the concept, simply the substantial code; at this time, the core attributes stored by the substantial equivalence include: two form codes representing the same concept; it is said that there is a substantial equivalence relationship between the two form codes representing the concept; the substantial equivalence relationship has transitivity, and all form codes representing concepts connected by the substantial equivalence relationship are substantially equivalent;

[0026] The analogy equivalence stores a relationship between two comprehensive sentences, the meaning of the relationship is that the two comprehensive sentences have the same context, which is realized as follows: two comprehensive sentences with the same context are called equivalent comprehensive sentence one and equivalent comprehensive sentence two, and the comprehensive forms representing the two comprehensive sentences are called equivalent comprehensive form one and equivalent comprehensive form two; therefore, the core attributes stored by the analogy equivalence include: equivalent comprehensive form one and equivalent comprehensive form two; it is said that there is an analogy equivalence relationship between equivalent comprehensive form one and equivalent comprehensive form two; the analogy equivalence relationship has transitivity, and the contexts of all comprehensive forms connected by the analogy equivalence relationship are the same;

[0027] The analysis stores an application scenario of the thinking method and a process description of applying the thinking method to obtain a reasoning result, which is implemented as follows: an application scenario of a thinking method is composed of two comprehensive sentences, wherein the first comprehensive sentence contains all semantics of the second comprehensive sentence or contains part of semantics of the second comprehensive sentence; at this time, the first comprehensive sentence is said to contain or partially contain the second comprehensive sentence, and a comprehensive form representing the first comprehensive sentence is said to contain or partially contain a comprehensive form representing the second comprehensive sentence; the analysis is said to connect the two comprehensive forms; the first comprehensive form is said to be a cause comprehensive form of the analysis; the second comprehensive form is said to be a result comprehensive form of the analysis; a process of applying the thinking method to obtain a reasoning result is divided into two steps; the first step is a screening step, that is, finding knowledge meeting specific conditions from existing knowledge as a reasoning basis; the second step is a generating step, that is, generating knowledge contained or partially contained by the knowledge obtained in the screening step or the knowledge that "no existing knowledge meeting specific conditions can be found" according to the knowledge obtained in the screening step; the analysis uses two groups of variable bits and concepts thereon to respectively represent execution methods of the screening step and the generating step; at this time, the analysis contains two parts: a screening part and a generating part; the screening part contains a group of variable bits and concepts thereon, representing the execution method of the screening step; the generating part contains a group of variable bits and concepts thereon, representing the execution method of the generating step; at this time, core attributes stored by the analysis include: a cause comprehensive form, a result comprehensive form, a group of screening part variable bits, a screening part variable sequence, a group of generating part variable bits, and a generating part variable sequence;

[0028] The said mapping between the analysis and the synthesis stores a relation between a comprehensive sentence and an analysis sentence, the meaning of the said relation is that the said comprehensive sentence is the context required for the transformation of the said analysis sentence into a specific instance of the said analysis sentence, which is realized as follows: based on the concept sequence represented by an analysis sentence, the process of transforming it into the concept sequence represented by a specific instance of the said analysis sentence, that is, the process of replacing the concept at a specific position in the concept sequence represented by the said analysis sentence with another concept in its context; when the said context can be described by a comprehensive sentence, the concept in the concept sequence represented by the said comprehensive sentence is the source of the concept for the aforementioned concept replacement operation; as mentioned before, the concepts in the concept sequence represented by the specific instance sentence of the analysis sentence that are related to the application scenario of the thinking method and the process of obtaining the reasoning result are stored by the screening part variable sequence and the generating part variable sequence of the analysis sentence; therefore, the result of the aforementioned concept replacement operation is that the concept appearing at a specific analysis sentence screening part variable position and a generating part variable position will be the same as the concept appearing at a specific comprehensive sentence variable position; the mapping between the analysis and the synthesis describes this concept same relation; at this time, the core attributes stored by the mapping between the analysis and the synthesis include: an analysis sentence, a comprehensive sentence, the corresponding relation between the variable positions in the analysis sentence and the variable positions in the comprehensive sentence that have the concept same relation; at this time, it is said that the said analysis sentence and the said comprehensive sentence have the mapping relation between the analysis and the synthesis; when an analysis sentence has the mapping relation between the analysis and the synthesis with at least one comprehensive sentence, the analysis sentence can be created based on the analysis sentence rather than the concept sequence represented by the specific instance of the analysis sentence; at this time, in the screening part and the generating part of the analysis sentence: in the process of transforming the analysis sentence into its specific instance, the concept on the variable position corresponding to the position in the concept sequence of the analysis sentence that needs to be replaced should be marked as "to be determined"; the "to be determined" concept represents a concept that has not been determined yet, which is not the same as any concept, and a "to be determined" concept is also not the same as another "to be determined" concept; the meaning of the aforementioned marking is that when the analysis sentence is created based on the concept sequence represented by an analysis sentence, it must be concept replaced through the mapping relation between the analysis and the synthesis to eliminate all "to be determined" concepts, so as to obtain the analysis sentence created based on the concept sequence represented by the specific instance of the analysis sentence; on the other hand, based on the aforementioned process of transforming the analysis sentence into its specific instance, the concept sequence obtained by partially or completely replacing the position in the concept sequence of the analysis sentence that needs to be replaced with the concept to be replaced, that is, the partial or complete specific instance sentence obtained by partially or completely instantiating the analysis sentence, can also create the analysis sentence and the mapping between the analysis and the synthesis, at this time, the mapping between the analysis and the synthesis represents the way of continuing instantiation or re-instantiation of the partial or complete specific instance sentence of the analysis sentence stored by the analysis sentence;For convenience of expression, the analysis sentence is understood as a partially instantiated analysis sentence instance, which has instantiated part empty, at this time, the above several analysis formula and the creation mode of the distribution mapping formula can be uniformly expressed as: the analysis formula is based on the concept sequence represented by the specific or specific instance of the analysis sentence part to create, when the variable sequence in the filter part contains the "undetermined" concept, the "undetermined" concept needs to be eliminated through the specific mode of the distribution mapping formula representation to obtain the analysis formula corresponding to the specific instance of the analysis sentence, at this time, the analysis formula is not alone, but cooperates with the distribution mapping formula to represent the thinking method of the specific instance of the analysis sentence; when the variable sequence of the filter part of the analysis formula does not contain the "undetermined" concept, the analysis formula alone stores the thinking method of the specific instance of the analysis sentence; when the variable sequence of the filter part of the analysis formula does not contain the "undetermined" concept, the analysis formula can also cooperates with the distribution mapping formula to represent the thinking method of other specific instances of the analysis sentence;

[0029] When the predicate formula stores the association between the comprehensive sentence and its subject part, the predicate formula stores: the position of the concept corresponding to the subject of the comprehensive sentence in the semantic structure in the comprehensive formula representing the comprehensive sentence, which is implemented as: calling the comprehensive formula representing the comprehensive sentence as the framework comprehensive formula of the predicate formula; calling the variable position where the concept corresponding to the subject of the comprehensive sentence in the framework comprehensive formula as the key variable position; at this time, the core attributes stored by the predicate formula include: a framework comprehensive formula, a key variable position set;

[0030] The process of creating the thinking map module unit of the knowledge base device has the following characteristics: a thinking map module unit records its identity and characteristic attributes that distinguish it from other units of the same kind, and the characteristic attributes are referred to as attributes; the knowledge base device uses a code to represent the identity of the thinking map module unit, which is referred to as the identity code of the thinking map module unit; the identity code of a thinking map module unit is used to refer to the unit in the attributes stored in other thinking map module units; the attributes stored by the thinking map module unit have uniqueness, that is, the identity code of the thinking map module unit corresponds to the attributes one by one; similar to the code representing the concept, the identity code of the thinking map module unit is also divided into formal code and substantive code; when a thinking map module unit stores its attributes, only the substantive code representing the identity of the thinking map module unit and the substantive code representing the concept are used, and the identity code of the thinking map module unit is the substantive code; when a thinking map module unit stores its attributes, the formal code representing the identity of the thinking map module unit or the formal code representing the concept is used, and the identity code of the thinking map module unit is the formal code; a thinking map module unit with a formal code identity code corresponds to a unique thinking map module unit with a substantive code identity code; multiple thinking map module units with a formal code identity code may correspond to the same thinking map module unit with a substantive code identity code; in the above many-to-one relationship, a thinking map module unit with a formal code identity code is referred to as a formal unit of a thinking map module unit with a substantive code identity code; conversely, a thinking map module unit with a substantive code identity code is referred to as a substantive unit of each thinking map module unit with a formal code identity code; the comprehensive context attribute saves a code representing a context concept; one comprehensive corresponds to a specific context, and multiple comprehensives may correspond to the same context, so the identity code of the comprehensive, including the formal code and the substantive code, is the formal code of the context concept; therefore, the analogy isomorphic type also explains the substantive equivalence relationship between the formal codes representing the context concept; based on the substantive equivalence relationship between the concept codes and the equivalence relationship of the syntax type, first, the many-to-one relationship between the formal units and the substantive units of the thinking map module units containing only concepts and syntax types in the attributes is established, and then the attributes of the thinking map module units using the formal code are converted to be saved using only the substantive code based on the many-to-one relationship between the formal units and the substantive units that have been established, and the substantive units of all thinking map module units can be gradually established.

[0031] As further improvement, the said synthesis includes three types, i. e. norm synthesis, provable synthesis and hypothesis synthesis; the norm synthesis stores certain fact; the provable synthesis stores possibility, which represents a set of possible norm synthesis conforming to its meaning; the hypothesis synthesis stores thinking target, which instructs the knowledge base device to make inference calculation to obtain a set of possible norm synthesis conforming to its meaning; the meanings of the norm synthesis conforming to the provable synthesis or hypothesis synthesis mean that they have same context, same syntax and same or provable concepts on corresponding variable positions, and the concepts of the provable synthesis or hypothesis synthesis on the corresponding variable positions are "undetermined"; the meaning of the corresponding variable position is that because the variable positions of a syntax correspond to the variable positions of the synthesis with the said syntax as attribute, the one-to-one correspondence between the variable positions of the two equivalent syntaxes indicates the one-to-one correspondence between the variable positions of the two syntheses with the two syntaxes as attribute respectively, and the said corresponding variable position means the variable positions of the two syntheses with the said correspondence.

[0032] As a further improvement, the analysis formulae include two types: positive analysis formulae and negative analysis formulae; the basis for one application of the thinking method stored in the positive analysis formulae is a set of normative hypothesis formulae; the basis for one application of the thinking method stored in the negative analysis formulae is a negative hypothesis formula, which refers to a hypothesis formula that does not conform to the meaning of any known normative hypothesis formula; the screening part variable positions of the analysis formulae correspond to the variable positions of the cause hypothesis formulae connected thereto, and the generating part variable positions correspond to the variable positions of the effect hypothesis formulae connected thereto; the inference result of the analysis formulae has a set of variable positions corresponding to the variable positions of the generating part of the analysis formulae; the generating part of the analysis formulae includes a set of variable positions and concepts thereon, which are used to represent the execution method of the generating step, and are implemented in that the concept on a generating part variable position represents the inference method of obtaining the concept on the corresponding variable position in the inference result; the inference method includes: known, i.e. fixedly obtaining a specific concept; reference, i.e. identical to the concept on a specific variable position in the normative hypothesis formula or hypothesis formula on which the analysis formula is applied; set, i.e. clustering the set of normative hypothesis formulae on which the analysis formula is applied according to the concept on a specific variable position, in each category, synthesizing the concepts on the variable positions not according to the specific variable position into a set, and confirming the existence of a concept for reference to the set; decomposition, i.e. confirming the existence of a concept for representing the normative hypothesis formula itself finally obtained by inference; pending, i.e. fixedly obtaining the concept of “pending”, which represents that the application of the thinking method stored in the analysis formula cannot obtain the concept on the variable position; at the same time, the analysis formula indicates the association between the specific generating part variable position and the screening part variable position, which is referred to as the source relationship of the generating part variable position; if the generating part is a concept representing the “set”, “decomposition” or “reference” method on the i-th variable position, it is associated with the j-th variable position of the screening part, and the source relationship indicates that the basis for obtaining the concept on the i-th variable position in the inference result is: the concept on the variable position corresponding to the j-th variable position of the screening part in the hypothesis formula on which the thinking method is applied this time; for simplicity of representation, the concept representing the “known” inference method is directly used in the generating part to represent the “known” inference method, i.e. fixedly obtaining the concept, and the concept of “pending” is used to represent the “pending” inference method; the variable positions of the generating part representing the “known”, “reflection” and “pending” inference methods have no source relationship.

[0033] The predicate is also used to connect the knowledge base device connected task execution system or data query system, to achieve: when a hypothesis synthesis formula in the form of predicate is created in the reasoning process, the knowledge base device operates the task execution system or the data query system, executes the task and confirms that the standard synthesis formula storing the task execution result exists in the mind map module, or queries the data and confirms that the standard synthesis formula storing the query result exists in the mind map module; the hypothesis synthesis formula conforms to the form of the predicate, which means that the following conditions are met simultaneously: 1) the context of the hypothesis synthesis formula is the same as that of the frame synthesis formula of the predicate, and the syntax is equivalent; 2) the concept on the key variable position in the frame synthesis formula is the same as that on the corresponding variable position in the hypothesis synthesis formula, or the concept in the frame synthesis formula is "to be determined"; 3) the concept on the variable position corresponding to the key variable position of the frame synthesis formula in the hypothesis synthesis formula is not "to be determined"; at this time, the concept on the variable position corresponding to the key variable position of the frame synthesis formula in the hypothesis synthesis formula constitutes the parameter of task execution or the condition of data query; the predicate can also be used to implement logical consistency checking, which requires that all reasoning results to be synthesized must explicitly specify the subject: when multiple analysis formulas are connected to the same result synthesis formula, if the concepts representing the reasoning methods on the mutually corresponding variable positions in the generation part of these analysis formulas do not contain "to be determined", these variable positions should be able to cover the subject of the synthesis sentence represented by the result synthesis formula; the significance of this logical consistency is that if multiple analysis formulas all obtain the partial implication relationship of the result synthesis formula, their partial conclusions can be connected to form a complete conclusion under the condition of consistent subject; a synthesis sentence can have multiple definitions of subject, for example, if a synthesis sentence can be rewritten in passive voice, the object of this synthesis sentence can also be defined as the subject; the above logical consistency checking can be implemented as follows: if multiple analysis formulas are connected to the same result synthesis formula, and the result synthesis formula is also the frame synthesis formula of at least one predicate, then there is at least one predicate with the result synthesis formula as the frame synthesis formula, which defines a group of variable positions in the result synthesis formula as key variable positions, at this time, the concept on the variable position corresponding to the key variable position of the result synthesis formula in the generation part variable position of the analysis formula connected to the result synthesis formula should not represent the "to be determined" reasoning method.

[0034] In the synthesis mapping relationship, if the synthesis formula is a provable synthesis formula, there is a synthesis mapping relationship between the analysis formula and each standard synthesis formula represented by the provable synthesis formula.

[0035] As a further improvement, the knowledge base device can also be used for reasoning calculation; the knowledge base device further comprises a reasoning calculation module; the reasoning calculation module performs reasoning calculation based on the units of the mind map module, and creates or deletes the units of the mind map module to record the process or result of the reasoning calculation according to the following method;

[0036] The thinking map module also includes a unit used in the reasoning process, called the reasoning result unit, abbreviated as reasoning result. The second step of the application analysis storage thinking method to obtain the reasoning result is the generation step, that is, according to the knowledge obtained in the screening step or the knowledge that "no existing knowledge can be found according to the specific conditions", generate the knowledge they contain or partially contain, which is implemented as: the reasoning calculation module confirms that the reasoning result in the thinking map module already exists to store the generated knowledge, otherwise create a reasoning result to store the generated knowledge; for simplicity of expression, "confirm that it exists" means "confirm that it already exists, otherwise create it"; the reasoning result includes a set of variable bits and a variable sequence composed of concepts on the variable bit in order of variable bit name; the reasoning result obtained based on an analysis formula has a set of variable bits corresponding to the fruit comprehensive formula connected with the analysis formula; the reasoning calculation module performs a comprehensive process on the reasoning result to obtain a standard comprehensive formula; the syntax of the standard comprehensive formula obtained by the comprehensive process is the same as that of the fruit comprehensive formula connected with the analysis formula, and the variable bits correspond to each other. The variable sequence is obtained by the following method: divide the reasoning results connected with the same fruit comprehensive formula into a group; in each group of reasoning results, find all possible combinations so that in each combination, all reasoning results have only one non-determined concept on the corresponding variable bit, and these non-determined concepts are sorted by variable bit name to obtain a variable sequence, which is a variable sequence of a standard comprehensive formula obtained by the comprehensive process; at this time, the core attributes stored by the reasoning result include: a grouping mark, a set of variable bits, and a variable sequence; the grouping mark of the reasoning result obtained from an analysis formula is the fruit comprehensive formula connected with the analysis formula.

[0037] The inference calculation module executes the process of inference calculation, which is composed of transaction units; the execution of the transaction units is triggered by the thinking map module units marked with states; the states include newly created, newly deleted, and to be integrated; the transaction units are divided into generating transaction units and deleting transaction units; the generating transaction units are triggered and executed by the thinking map module units marked with the newly created or to be integrated states; when a specific step of the generating transaction units needs to delete a thinking map module unit, it does not directly delete it, but only marks it with the newly deleted state; in all transaction units, all operation steps of "the inference calculation module finds… thinking map module units" ignore the thinking map module units marked with the newly deleted state; when a specific step of a transaction unit confirms the existence of a thinking map module unit marked with the newly deleted state, the newly deleted state is cancelled; when a specific step of a transaction unit creates a new thinking map module unit, it is marked with the newly created state; the deleting transaction units are triggered and executed by the thinking map module units marked with the newly deleted state, and the units are actually deleted in the transaction units; the thinking map module units marked with the states lose the states immediately after triggering the execution of the corresponding transaction units; the inference calculation module records the process in which the thinking map module units that do not exist in the knowledge base device at the beginning of inference are confirmed to exist by the generating transaction units; the execution of the generating transaction units is triggered by a thinking map module unit marked with a state, which is called a triggering unit, and multiple thinking map module units may be confirmed to exist, each of which is called a confirmed unit; the inference calculation module records the process in which a generating transaction unit confirms the existence of a group of thinking map module units as a group of proof relations; the proof relations are a type of thinking map module units that need to be used by the inference calculation module when performing inference calculation, and the core attributes stored therein include a triggering unit and a confirmed unit; a proof relation is supported by its triggering unit and owned by its confirmed unit; for all thinking map module units that do not exist at the beginning of inference, if they lose all proof relations owned by them after the completion of a transaction unit, they are marked with the newly deleted state; when the inference calculation module actually deletes a thinking map module unit, it also deletes all proof relations supported by the unit; the proof relations are not deleted or marked with states because they do not own proof relations, and are immediately deleted when needed;

[0038] When a hypothesis synthesis is marked as newly created, a transaction unit is triggered to be generated: let the hypothesis synthesis be H; first, all the proof relations supported by H are deleted; then, the inference computing module finds all the analysis that can confirm the existence of inference results conforming to the meaning of H; for each of them, let it be Ah, the inference computing module generates a filter condition Fa according to the inference method represented by the inference method part of Ah, so that the canonical synthesis filtered by Fa can possibly obtain inference results conforming to the meaning of H after applying the thinking method of Ah; if Fa conforms to the meaning of the filter part of Ah, it is confirmed that there exists a hypothesis synthesis Ha, which has the same context and syntax as the cause synthesis of Ah, and its variable sequence represents Fa, and a proof relation (H, Ha) is created, and it is confirmed that there exists an association between Ha and Ah, which is called a loop association; the loop association is a kind of thinking map module unit that needs to be used by the inference computing module when performing inference calculation, and the core attributes stored therein include a hypothesis synthesis and an analysis; let the loop association Sah describe the loop association between Ha and Ah, and create a proof relation (H, Sah); then, the inference computing module finds all the part synthesis mapping of Ah, and for each of them, let it be Ma, let Ma describe the part synthesis mapping relationship between the analysis Ah and the synthesis Ta, and if the meaning of Ta is not contradictory to Fa, it is confirmed that there exists a hypothesis synthesis Ht, which has the same context and syntax as Ta, and its variable sequence represents the meaning of Ta limited by Fa, and a proof relation (H, Ht) is created, and it is confirmed that there exists an association between Ma, Ht and Ah, which is called a transformation association; the transformation association is also a kind of thinking map module unit that needs to be used by the inference computing module when performing inference calculation, and the core attributes stored therein include a part synthesis mapping, a hypothesis synthesis and an analysis; let the transformation association Cmha describe the transformation association between Ma, Ht and Ah, and create a proof relation (H, Cmha); then, the inference computing module finds all the predicate that conforms to the meaning of H, and for each of them, let it be Ph, if Ph is associated with the task execution system or the data query system connected by the inference computing module, the inference computing module executes the task or queries the data, and it is confirmed that there exists a group of canonical synthesis that stores the task results or the required data, and for each of them, let it be Nh, a proof relation (H, Nh) is created; a kind of judgment method for the analysis that can confirm the existence of inference results conforming to the meaning of the hypothesis synthesis is that the syntax of the fruit synthesis connected by the analysis is equivalent to the syntax of the hypothesis synthesis, and the context is the same, and the variable sequence of the generation part of the analysis satisfies: 1) the concept of the hypothesis synthesis is "undetermined", or 2) the concept of the hypothesis synthesis is not "undetermined" and the concept of the generation part represents a "known" inference method, or 3) the concept of the hypothesis synthesis is not "undetermined" and is the same as the concept of the generation part;The screening condition according to the analysis formula makes the screened standard comprehensive formula possibly apply the analysis formula storage thinking method to obtain the reasoning result conforming to the hypothesis comprehensive formula meaning. One implementation is that the screening condition is recorded as a variable sequence, and the variable positions thereof correspond to the variable positions of the analysis formula screening part. At this time, since the variable positions of the analysis formula generation part are associated with the variable positions of the screening part in the source relationship, the variable positions of the analysis formula generation part are also simultaneously associated with the variable positions of the screening condition. The concept on each variable position of the screening condition is determined by the following method: the variable positions of the hypothesis comprehensive formula correspond to the variable positions of the generation part; for each pair of corresponding variable positions, if the concept of the hypothesis comprehensive formula is not "undetermined" and the concept of the generation part represents the "reference" reasoning method, the concept on the variable position of the screening condition associated with the variable position of the generation part is equal to the concept of the hypothesis comprehensive formula. The concept on the variable position of the screening condition which cannot be determined according to the method is equal to the non-"undetermined" concept on the corresponding variable position of the screening part. If the concept on the corresponding variable position of the screening part is "undetermined", the concept thereon is "arbitrary". One judgment method of the screening condition conforming to the meaning of the analysis formula screening part is that the concepts on the corresponding variable positions of the screening condition variable sequence and the analysis formula screening part variable sequence are the same or the concept of the screening part is "arbitrary". One implementation of the variable sequence of the hypothesis comprehensive formula representing the screening condition is that the variable sequence of the hypothesis comprehensive formula is obtained by replacing "arbitrary" in the screening condition variable sequence with "undetermined". One judgment method of the meaning of the comprehensive formula not contradicting the screening condition is that the score comprehensive relationship contains the concept same relationship between the i th variable position of the analysis formula screening part and the j th variable position of the comprehensive formula. For each group of i and j, the following conditions are met: 1) the concept on the j th variable position of the comprehensive formula is not "undetermined", and the concept on the i th variable position of the screening condition is the same as or "arbitrary", or 2) the concept on the j th variable position of the comprehensive formula is "undetermined". The variable sequence of the hypothesis comprehensive formula represents the meaning of the comprehensive formula having the score mapping relationship with the analysis formula obtained by limiting the screening condition. One implementation is that the variable sequence of the hypothesis comprehensive formula is obtained by the following method: 1) for each group of the foregoing i and j, if the concept on the j th variable position of the comprehensive formula is "undetermined" and the concept on the i th variable position of the screening condition is "arbitrary", the concept on the j th variable position of the hypothesis comprehensive formula is "undetermined", 2) at the same time, for each group of the foregoing i and j, if the concept on the j th variable position of the mapping comprehensive formula is "undetermined" and the concept on the i th variable position of the screening condition is not "arbitrary", the concept on the j th variable position of the hypothesis comprehensive formula is the same as the concept on the i th variable position of the screening condition, 3) the concept on the variable position of the hypothesis comprehensive formula which cannot be determined according to the foregoing method is equal to the concept on the corresponding variable position of the mapping comprehensive formula.

[0039] When an analysis formula is marked with a newly created or newly deleted state, trigger the execution of the corresponding transaction unit, and in the generation or deletion transaction unit, perform the following operations: record the analysis formula as A; the reasoning calculation module finds the hypothesis synthesis formula that may confirm the existence of the reasoning result conforming to the meaning of the analysis formula, and marks the newly created state for the hypothesis synthesis formula;

[0040] When a loop evidence association is marked with a newly created state, trigger the execution of the generation transaction unit (as shown in Figure 1 , Figure 1 The following describes the operation of the 22nd step of the specific implementation case): record the loop evidence association as S, S describes the loop evidence association between the hypothesis synthesis formula H and the analysis formula A; first, delete all proof relations supported by S; then, the reasoning calculation module finds the normative synthesis formula conforming to the meaning of H, and records its set as E; if A is a positive analysis formula, take E as the basis of the thinking method stored by applying A; if E is an empty set and A is a negative analysis formula, take H as the basis of the thinking method stored by applying A; the reasoning calculation module confirms the existence of each reasoning result based on the above basis and applies the thinking method stored by A; for each reasoning result, record it as Rs, and create the proof relation (S, Rs);

[0041] When a transformation association is marked with a newly created state, trigger the execution of the generation transaction unit (as shown in Figure 2 , Figure 2 The following describes the operation of the 27th step of the specific implementation case): record the transformation association as C, C describes the transformation association between the hypothesis synthesis formula H, the analysis formula A, and the mapping formula M; first, delete all proof relations supported by C; then, find the normative synthesis formula conforming to the meaning of H, and for each of them, record it as Nt; the reasoning calculation module confirms the existence of an analysis formula, whose cause synthesis formula and effect synthesis formula are the same as A, and its screening part and generation part variable sequence is obtained by replacing the concept in Nt in the concept relationship described by M according to the screening part and generation part variable sequence of A; record the analysis formula as At, and create the proof relation (C, At);

[0042] When a normative synthesis formula is marked with a newly created or newly deleted state, trigger the execution of the corresponding transaction unit, and in the generation or deletion transaction unit, perform the following operations: record the normative synthesis formula as N; the reasoning calculation module finds the hypothesis synthesis formula conforming to the meaning of N, and for each of them, record it as Hn; the reasoning calculation module finds the loop evidence association in which Hn participates, and marks the newly created state for it; the reasoning calculation module finds the transformation association in which Hn participates, and marks the newly created state for it;

[0043] When a reasoning result is marked as newly created or newly deleted, a corresponding transaction unit is triggered to be executed, and the following operations are performed in the generation or deletion transaction unit: let R be the reasoning result, and let G be the group mark of R; mark G as to-be-combined state;

[0044] When a combination is marked as to-be-combined state, a generation transaction unit is triggered to be executed (as shown in Figure 3 , and the following operations are performed: Figure 3 The following describes the operation of the 24th step of the specific implementation case): let N be the combination, and first, delete the proof relation supported by N; then, the reasoning calculation module finds all reasoning results marked as N, and performs the foregoing combination process on the reasoning results; the reasoning calculation module confirms the existence of the canonical combination obtained by the combination process, and for each of the canonical combinations, let Nc be the canonical combination, and create a proof relation (N, Nc).

[0045] As a further improvement, the knowledge base device creates various types of mind map module units based on sequences composed of symbols representing specific concepts, and stores the knowledge and thinking methods represented by the sequences in the mind map module units; the knowledge base device assigns different codes to different sequences, which are referred to as sequence codes; assigns a code to each position in each sequence, which is referred to as a position code; uses the sequence code as the identity form code of the mind map module unit, and uses the position code as the form code of the concept; and the knowledge base device records the symbol sequence on which the mind map module unit is created when the mind map module unit is created.

[0046] As a further improvement, the knowledge base device creates various types of mind map module units based on natural language texts entered by a user in the form of symbol sequences, and responds to a thinking target specified by the user in the form of a natural language text in the form of a symbol sequence.

[0047] The knowledge base device records the symbol sequence on which the thinking map module units are created while creating the thinking map module units, and is implemented as follows: the knowledge base device further comprises an auxiliary module for storing natural language texts in the form of symbol sequences input by a user; a text block in the natural language texts represents a specific concept; a natural language sentence is a sequence composed of symbols, and the symbol sequence corresponds to a concept sequence arranged in the symbol order of the concepts represented by each symbol in the symbol sequence; the natural language texts are organized in articles, and an article comprises an original text symbol sequence, referred to as an original sequence, and a set of explanation symbol sequences, referred to as explanation sequences; the original sequence refers to a text used in natural language communication; the explanation sequences refer to text representations of semantic structures, knowledge and thinking methods that need to be known in advance for understanding the meaning expressed by the text; the explanation sequences comprise syntax sequences, synthesis sequences and analysis sequences, which represent semantic structures, knowledge and thinking methods stored in the syntax type, the synthesis type and the analysis type of the thinking map module respectively; meanwhile, specific subsequences of the original sequence also constitute sentences for explaining semantic structures, knowledge and thinking methods, and the contents explained by the subsequences of the original sequence are also stored in the syntax type, the synthesis type and the analysis type; as described above, the knowledge base device assigns a sequence code to each sequence and a position code to each position of each sequence; the knowledge base device creates the thinking map module units based on the codes; meanwhile, the knowledge base device uses a type of symbol units in the auxiliary module to store text block symbols, uses a type of position units to store the association between a position code and a symbol appearing at the position represented by the position code, and uses a type of article units to record the arrangement of position codes corresponding to the original sequence and the explanation sequences of an article; meanwhile, the knowledge base device supplements the record of the arrangement of position codes corresponding to the symbol sequence on which the units are created in the thinking map module units.

[0048] The knowledge base device responds to a thinking target provided by a user with natural language texts in the form of symbol sequences, and is implemented as follows: when reporting a synthesis type, the knowledge base device refers to a symbol sequence, referred to as a reference sequence of the synthesis type; a reference sequence refers to any symbol sequence that can establish a one-to-one correspondence with the variable positions of the synthesis type, and such symbol sequences include syntax sequences or original sequence subsequences corresponding to all equivalent syntax types of the syntax type of the synthesis type; the knowledge base device keeps the symbols in the reference sequence that do not correspond to the variable positions of the synthesis type unchanged, and replaces the symbols corresponding to the variable positions of the synthesis type with any symbol that can represent the concept in the variable position.

[0049] As a further improvement, the knowledge base device uses the following types of codes; using a type of mind map module unit, a form code is used to represent a set of such units in a mind map module, the attributes of this set of such units use form codes, so the identity code of this set of such units is also a form code; using a type of mind map module unit, a substance code is used to represent a set of such units in a mind map module, the attributes of this set of such units only use substance codes, so the identity code of this set of such units is also a substance code; the knowledge base device uses the following types of codes and corresponding types of units;

[0050] locf: global location code / concept form code; if the symbol at a specific position in a symbol sequence that can be accurately understood refers to a concept, it must refer to a unique concept; symbols at different sequence positions may refer to the same concept; therefore, the global location code is also the form code of the concept;

[0051] stxf: code of a syntactic sequence or a sub-sequence of a text sequence that describes semantic structure / syntactic form code; a syntactic sequence or a sub-sequence of a text sequence that describes semantic structure corresponds to a unique syntactic form; different syntactic sequences or sub-sequences of text sequences that describe semantic structure may correspond to the same syntactic form; therefore, the code of a syntactic sequence or a sub-sequence of a text sequence that describes semantic structure is also the form code of the syntactic form;

[0052] synf: code of a synthetic sequence or a sub-sequence of a text sequence that describes knowledge / synthetic form code / grouping marker substance code; a synthetic sequence or a sub-sequence of a text sequence that describes knowledge corresponds to a unique synthetic form; different synthetic sequences or sub-sequences of text sequences that describe knowledge may correspond to the same synthetic form; therefore, the code of a synthetic sequence or a sub-sequence of a text sequence that describes knowledge is also the form code of the synthetic form; on the other hand, each synthetic sentence derived by reasoning in the original text must have its complete reasoning path, if the synthetic sentence is derived by synthesizing the partial reasoning results of multiple reasoning paths, the synthetic sentence uniquely indicates the grouping marker when the partial reasoning results synthesized by this set of reasoning paths, therefore the code of the synthetic sequence of the knowledge derived by reasoning, the code of the sub-sequence of the text sequence, the code of the synthetic form are also the substance code of the grouping marker;

[0053] anaf: code of an analysis sequence or a sub-sequence of a text sequence that describes thinking methods / analysis form code; an analysis sequence or a sub-sequence of a text sequence that describes thinking methods corresponds to a unique analysis form; different analysis sequences or sub-sequences of text sequences that describe thinking methods may correspond to the same analysis form; therefore, the code of an analysis sequence or a sub-sequence of a text sequence that describes thinking methods is also the form code of the analysis form;

[0054] prdf: form code of a predicate form;

[0055] stxef: form code of syntactic equivalence form;

[0056] synef: form code of analog equivalence form;

[0057] cptef: form code of substantial equivalence form;

[0058] asef: form code of analytic mapping form;

[0059] stxs: substantial code of syntactic form;

[0060] ctxs: substantial code of context;

[0061] cpts: substantial code of concept;

[0062] syns: substantial code of synthetic form;

[0063] anas: substantial code of analytic form;

[0064] prds: substantial code of predicate form;

[0065] ases: substantial code of analytic mapping form;

[0066] seks: substantial code of evidence association;

[0067] tfms: substantial code of transformation association;

[0068] ress: substantial code of reasoning result.

[0069] As a further improvement, the workflow of the knowledge base device is as follows:

[0070] First, acquire and store knowledge, thinking methods, and thinking goals; the thinking map module of the knowledge base device receives the natural language text in the form of symbol sequence input by the user, assigns a sequence code and a position code to each symbol sequence and each position in it, takes the sequence code as the form code of the thinking map module unit, and takes the position code as the form code of the concept; create a thinking map module unit based on these form codes, and the thinking map module unit is a form unit, and the identity code of the thinking map module unit is the form code;

[0071] Second, start reasoning calculation; the reasoning calculation module of the knowledge base device starts reasoning calculation according to the user's indication, and first completes the preparation work of reasoning calculation, that is, obtains the substantial unit of all thinking map module units, and the method is as follows;

[0072] First, based on the substantial equivalence form, create the substantial code of the concept represented by the form code of the concept represented by all form codes having a substantial equivalence relationship, and record the form code-substantial code correspondence relationship of the form code of the concept represented by the code; the form code-substantial code correspondence relationship of the form code of the concept represented by the code is called the concept correspondence relationship;

[0073] Then, based on the analogy equivalence, a substantial code of the common context of all the synthetic form codes with analogy equivalence is created, and the correspondence between the synthetic form code and the context substantial code is recorded; the correspondence between the synthetic form code and the context substantial code is called the synthetic context correspondence;

[0074] Then, based on the syntactic equivalence, the order of the variable bits of the syntactic form with syntactic equivalence is unified, and the correspondence between the unified variable bit order and the original variable bit order is recorded; a syntactic substantial code is created for all the syntactic form codes with syntactic equivalence, and the unified variable bit order is represented by the substantial code; a syntactic substantial unit is created; the correspondence between the syntactic form code and the substantial code, and the correspondence between the variable bits of the syntactic form unit and the syntactic substantial unit are recorded; the correspondence between the syntactic form code and the substantial code, and the correspondence between the variable bits of the syntactic form unit and the syntactic substantial unit are called the syntactic correspondence;

[0075] Then, based on the aforementioned concept correspondence, syntactic correspondence, and synthetic context correspondence, the attributes of the synthetic form represented by the form code are converted into the attributes represented by the substantial code; a synthetic substantial code is assigned to each different attribute, and a synthetic substantial unit is created; the correspondence between the synthetic form code and the substantial code is recorded; the correspondence between the synthetic form code and the substantial code is called the synthetic correspondence;

[0076] Then, based on the aforementioned concept correspondence, synthetic correspondence, and syntactic correspondence, the attributes of the analytic form represented by the form code are converted into the attributes represented by the substantial code; an analytic substantial code is assigned to each different attribute, and an analytic substantial unit is created; the correspondence between the analytic form code and the substantial code is recorded; the correspondence between the analytic form code and the substantial code is called the analytic correspondence;

[0077] Then, based on the aforementioned synthetic correspondence, syntactic correspondence, and analytic correspondence, the attributes of the mapping form represented by the form code are converted into the attributes represented by the substantial code; a mapping form substantial code is assigned to each different attribute, and a mapping form substantial unit is created; the correspondence between the mapping form code and the substantial code is recorded;

[0078] Then, based on the aforementioned synthetic correspondence and syntactic correspondence, the attributes of the predicate form represented by the form code are converted into the attributes represented by the substantial code; a predicate form substantial code is assigned to each different attribute, and a predicate form substantial unit is created; the correspondence between the predicate form code and the substantial code is recorded;

[0079] Third step, reasoning calculation module according to the essence unit in the thinking map module, carries out reasoning calculation according to the preceding steps;At the beginning of reasoning, the essence unit corresponding to the hypothesis comprehensive form unit entered by the user is marked with a newly created state;During the reasoning process, when the normative comprehensive form corresponding to the hypothesis comprehensive meaning entered by the user is created, found or deleted, the reasoning calculation module immediately reports to the user: the reasoning calculation module finds the available reference sequence and the symbol that can represent the specific concept according to the form unit and the essence unit of the thinking map module unit according to the preceding steps, and forms a natural language text report to the user in the form of symbol sequence.

[0080] The beneficial effects of the present application are:

[0081] 1. Allow to express knowledge and thinking method based on natural language text, express ability is strong, easy to understand and use;Support the gradual fusion of knowledge expressed in different languages;Support the establishment of large knowledge base in a gradual fusion manner;

[0082] 2. Knowledge and thinking method are expressed in a mutually related manner, so that the results of reasoning calculation can further drive new reasoning calculation process, and realize self-driven thinking;New information can be actively obtained during the reasoning calculation process;

[0083] 3. In the case that the thinking map content in the knowledge base conforms to the fact, the reasoning calculation process is based on the fact, and the reasoning calculation result also reflects the fact;

[0084] 4. Known knowledge, thinking method, new knowledge generated by reasoning can be reported in the form of natural language text, the reasoning process is clear and can be explained, which is convenient for manual checking and correcting the knowledge base. BRIEF DESCRIPTION OF DRAWINGS

[0085] Figure 1 It is an example diagram of a newly created transaction unit triggered by a evidence correlation;

[0086] Figure 2 It is an example diagram of a newly created transaction unit triggered by a transformation correlation;

[0087] Figure 3 It is an example diagram of a transaction unit triggered by a to-be-combined grouping marker comprehensive form;

[0088] Figure 4 It is an example diagram of a user-entered thinking map module unit;

[0089] Figures 5-7 It is a data record diagram formed by the reasoning process of the patent embodiment in the database. DETAILED DESCRIPTION

[0090] The following merely illustrates the embodiments of the present application, and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0091] The following describes an embodiment of a knowledge base apparatus in software form, which uses data tables of a relational database to implement the aforementioned thinking map module units and auxiliary module units; the primary key of the data table is the identity code of the unit, and other fields of the data table are the attributes of the unit; and the following descriptions and conventions are made:

[0092] 1) Capital letter marks represent table names or field names; the tail letter C represents a field, and the field with an asterisk represents the primary key; the tail letter F represents a data table, and the table only uses form codes or other codes or information to represent thinking map module units, which is referred to as a form table; the tail letter S represents a data table, and the table only uses substantive codes or other codes or information to represent thinking map module units, which is referred to as a substantive table; the tail letter M represents a data table, and the table simultaneously uses form codes or substantive codes or other codes or information to represent thinking map module units, which is referred to as a comparison table; the tail letter X represents a data table, and the table represents auxiliary module units; the form code field in the comparison table is the primary key of the table;

[0093] 2) Small letter marks represent codes; the tail letter f represents a code as a form code of the identity of a thinking map module unit; the tail letter s represents a code as a substantive code of the identity of a thinking map module unit; the tail letter x represents a code as a code of the identity of an auxiliary module unit;

[0094] 3) An underscore starting small letter mark represents other information that is not a code;

[0095] 4) The "XXXF(AAAC*:aaaf, BBBSC: <bbbf>The structure of a data table is described in the form of "XXXF(AAAC, BBBSC, ZZZC, _zzz,...) " where XXXF is the table name; AAAC, BBBSC, ZZZC are field names; the AAAC field with an asterisk indicates the primary key; _zzz indicates a specific type of non-code information; "<>" indicates a sequence, <bbbf>represents a sequence of codes of type bbbf; the content of the field is specified in the form "field: code or sequence or other information";

[0096] 5) a mixed case letter mark represents a record in a data table;

[0097] 6) the "[] " operator is used to represent a search, and "XXXF[AAAC=aaaf] " represents the records obtained by searching the data table with a specific condition, which can be one or more records;

[0098] 7) the ". " operator is used to represent the field in a data record, and when the operator is in front of a record, the value of a field is taken out, and when the operator is in front of a group of data records, the set of field values in all records is taken out;

[0099] 8) the " () " operator is used to represent the sequence elements in order;

[0100] 9) Len( ) is used to represent the length of the sequence;

[0101] 10) " " is used to represent the value of a data table record, and the values of the fields in the " " are arranged in the order of the fields in the data table definition;

[0102] 11) " converted to substantive code " is used to represent the following operation: for a specific formal code, find the record in which the formal code is located in the comparison table whose primary key is the formal code of this type, and extract the substantive code field of the record;

[0103] 12) if the arrangement of a sequence of concept codes follows a specific form of a syntax, that is, follows the order of the " i-th variable position " used in the field description of the data record in the following syntax form table, it is said that the sequence of concept codes is arranged in form order; if the arrangement follows the essence of a syntax, that is, follows the order of the " i-th variable position " used in the field description of the data record in the following syntax essence table, it is said that the sequence of concept codes is arranged in essence order; " arranged in essence order " is used to represent the following operation: for a specific sequence of concept codes, record its syntax form as stxf; in the following STXM table, find the record whose primary key value is stxf, and its variable position name comparison field gives the information that the i-th variable position of the syntax form corresponds to the j-th variable position of the syntax essence, and the arrangement order of the sequence of concept codes is adjusted according to this information;

[0104] 13) " substantivize " is used to represent the following operation: if the object of substantivization is a formal code, it is converted into a substantive code; if the object of substantivization is a sequence of formal codes, each element is first converted into a substantive code, and then the sequence is arranged in essence order;

[0105] 14) Since the aforementioned variable positions are the 1st…nth, there is no need to record the variable positions themselves in the data table, only the corresponding relationship between the variable positions; this design is applied to the design of all the following data tables, and will not be repeated;

[0106] The knowledge base device uses the aforementioned thought map module codes locf, stxf, synf, anaf, prdf, stxef, synef, cptef, asef, stxs, ctxs, cpts, syns, anas, prds, ases, seks, tfms, ress, and the following types of auxiliary module codes:

[0107] 1) symx: symbol / text block code;

[0108] 2) artx: article code;

[0109] 3) setx: code of a set composed of a group of concepts, the set itself is not a concept, a set concept refers to a set;

[0110] The knowledge base device uses the following types of other information that are not codes:

[0111] 1) _txt: the content of the symbol / text block, represented in the database as a string or a serialized form of a general text block such as an image;

[0112] 2) _ind: position in the sequence, represented in the database as an integer, with the first sequence position being 1 and invalid sequence positions being 0, using positive and negative integers to represent the sequence positions in two different sequences in the same sequence sequence;

[0113] 3) _typ: comprehensive type, with three possible values, namely normative comprehensive, provable comprehensive, and hypothesis comprehensive, represented as N, P, and H respectively;

[0114] 4) _mod: analysis type, with two possible values, positive analysis and negative analysis, represented as P and N respectively;

[0115] 5) _sta: state of the thought map module unit, with three possible values, representing new creation, new deletion, and pending synthesis;

[0116] The knowledge base device uses the following data tables:

[0117] 1) SYMX (IDC*:symx, TXTC:_txt): SYMX, text block content table; IDC, symbol code field; TXTC, text content field; a record indicates that a text block of specific content corresponds to a specific symbol code; this table is modified when the knowledge base device receives a symbol sequence: if a text block of a specific content is not in the table, a record of this table is created;

[0118] 2) LOCF (IDC*:locf, SYMC:symx): LOCF, global location symbol table; IDC, global location code field; SYMC, symbol field; a record indicates that a specific symbol appears at a specific global location; this table is modified when the knowledge base device receives a symbol sequence: the knowledge base device creates a global location code distinguishable from all locations in all sequences for each location in each received symbol sequence, finds the symbol code of the text content appearing at the location in the SYMX table, creates one if none exists, and records in this table: the symbol appears at the global location;

[0119] 3) ARTX (ARTC:artx, INDC:_ind, LOCC*:locf): ARTX, article sequence table; ARTC, article field; INDC, sequence position field; LOCC, global location field; a record indicates that global location locf is related to article artx; if _ind is not 0, locf is the location in the original article sequence, and _ind is the sequence of the global location in the original article sequence; if _ind is 0, locf is the location in the article explanation sequence; the aforementioned knowledge base device creates a comprehensive formula describing the task execution result or data query result according to the predicate formula associated task execution system or data query system, and creates these comprehensive formulas also according to the symbol sequence returned by the task execution system or data query system, which is also considered as the explanation sequence of the article to which the predicate formula belongs; this table is modified when the knowledge base device receives a symbol sequence, and the text block at global location locf is SYMX [IDC=LOCF [IDC=locf].SYMC].TXTC;

[0120] 4) STXF (IDC*:stxf, ARTC:artx, SYMSC: <locf>, VARSC: <locf>STXF, syntax form table; IDC, code field; ARTC, article field; SYMSC, text block global position sequence field; VARSC, variable bit global position sequence field; one record indicates that in the article artx, a syntax sequence or a text sequence sub-sequence indicates a syntax, which shares the syntax form code stxf; the sequence is composed of a set of global positions, which are stored in the SYMSC field; the global positions corresponding to the variable bits in the sequence are stored in the VARS field <locf>In this case, the i-th variable bit in the syntax representation corresponds to the global position <locf>(i); the table is modified as the knowledge base device receives the syntactic sequence or a subsequence of the original text sequence representing the syntactic form;

[0121] 5) SYNF (IDC*:synf, ARTC:artx, STXC:stxf, VALSC: <locf>TYPC: typ, SYMSC: symc <locf>SYNF, integrated form code; IDC, code field; ARTC, article field; STXC, syntax code field; VALSC, variable sequence field; TYPC, type field; SYMSC, text block global position sequence field; a record represents that in the article artx, an integrated sequence or a text sequence sub-sequence represents an integrated form, which shares the integrated form code synf; the sequence is the source sentence of the integrated form, which consists of a set of global positions and is saved in the SYMSC field; the meaning of the source sentence can be expressed by the syntax form stxf, and when the form is used, the global positions of the text block of the source sentence corresponding to the variable bit are saved in the VALSC field in the order of the variable bit name; the type of the integrated form is _typ; the table is modified when the integrated sequence or the text sequence sub-sequence representing the integrated form is received by the knowledge base device;

[0122] 6) ANAF(IDC*:anaf, ARTC:artx, CSYNC:synf, RSYNC:synf, FVALSC: <locf>GVALSC: <locf>,SRCC: <locf>TYPC: _mod, SYMSC: <locf>ANAF, analysis form field; IDC, code field; ARTC, article field; CSYNC, cause synthesis field; RSYNC, result synthesis field; FVALSC, filter variable sequence field; GVALSC, generator variable sequence field; SRCC, source relation field; TYPC, type field; SYMSC, text block global position sequence field; one record represents, in an article artx, an analysis sequence or a source sequence subsequence of a text sequence, which share an analysis form code anaf; the source sentence of the analysis is composed of a set of global positions, which are stored in the SYMSC field; one form of the cause synthesis of the analysis is synf in the CSYNC field; one form of the result synthesis of the analysis is synf in the RSYNC field; the global positions in the source sentence corresponding to the filter variable positions of the analysis are arranged in the order of the variable positions of the syntax form of this form of the cause synthesis, and the resulting sequence is stored in the FVALSC field; the global positions in the source sentence corresponding to the generator variable positions of the analysis are arranged in the order of the variable positions of the syntax form of this form of the result synthesis, and the resulting sequence is stored in the GVALSC field; the SRCC field describes the source relation, which <locf>with the GVALSC field <locf>is of equal length and its constituent elements are the FVALSC fields <locf>compositional elements or a code representing an invalid global position, the SRCC field is recorded <locf>For <slocf>, the GVALSC field is recorded <locf>For <glocf>, the FVALSC field is recorded <locf>For <flocf>; source relationship <slocf>description, <glocf>correlation of the middle ith position <flocf>occurrence <slocf>the position of (i) if <slocf>(i) is a null global location, then the associated location; the type of this analysis is _mod; this table is modified when the knowledge base facility receives an analysis sequence or a text sequence representing an analysis expression subsequence;

[0123] 7) PRDF(IDC*:prdf, FSYNC:synf, KEYSC:<_ind>) : PRDF, predicate form table; IDC, code field; FSYNC, frame synthesis form field; KEYSC, key variable position field; a record represents that in the form synf of the frame synthesis form, its variable sequence is represented as the concept form code sequence SYNF [IDC=synf].VALSC, in which the sequence positions of the key variable positions are arranged in size to obtain <_ind>; this table is modified when the knowledge base facility receives the recorded synthesis sentence structure knowledge;

[0124] 8) STXEF(IDC*:sxef, XSTXC:stxf, YSTXC:stxf, MAPC:<_ind>) : STXEF, syntax equivalent form table; IDC, code field; XSTXC, equivalent syntax form one field; YSTXC, equivalent syntax form two field; MAPC, variable position name correspondence field; a record represents that the two stxf of the XSTXC field and the YSTXC field have an equivalent relationship, and the values of the two fields are recorded as stxxf and stxyf respectively; the i-th variable position in the syntax form stxyf corresponds to the <_ind>(i)-th variable position in the syntax form stxxf; the syntax forms with equivalent relationship are said to have the same syntax form essence; this table is modified when the knowledge base facility receives the recorded syntax equivalent relationship;

[0125] 9) SYNEF(IDC*:synef, XSYNC:synf, YSYNC:synf) : SYNEF, analog equivalent form table; IDC, code field; XSYNC, equivalent synthesis form one field; YSYNC, equivalent synthesis form two field; a record represents that the two synf of the XSYNC and YSYNC fields represent the synthesis forms with the same source sentence context; this table is modified when the knowledge base facility receives the recorded analog equivalent relationship;

[0126] 10) CPTEF(IDC*:cptef, XCPTC:locf, YCPTC:locf) : CPTEF, essence equivalent form table; IDC, code field; XCPTC, equivalent concept one field; YCPTC, equivalent concept two field; a record represents that the two locf of the XCPTC and YCPTC fields represent the same concept on the global location; this table is modified when the knowledge base facility receives the recorded essence equivalent relationship;

[0127] 11) ASEF (IDC* : asef, ANAC : anaf, SYNC : synf, RPLC: <locf>ASEF, Subtotal Mapping Form Table; IDC, Code Field; ANAC, Analytic Expression Field; SYNC, Synthetic Expression Field; RPLC, Substitution Relationship Field; A record indicates that there is a subtotal mapping relationship between an analytic expression with form anAF and a synthetic expression with form SYNC. The RPLC field describes the substitution relationship; the RPLC field of this ASEF table record... <locf>For <rlocf> ； <rlocf>is equal to the global location code sequence of the SYNF [IDC=synf].VALSC field, and is recorded as <slocf> ； <rlocf>With <slocf>is the same length and whose constituent elements are the constituent elements of the sequence of the FVALSC field and the GVALSC field in the ANAF [IDC=anaf] record or a code representing an invalid global position, denoted as <aflocf> 、 <aglocf>; substitution relationship <rlocf>instructions, <slocf>Mid i-th variable bit substitution <aflocf> 、 <aglocf>occurrence <rlocf>the position of (i) if <rlocf>(i) if the global position is invalid, then no replacement position; this table is modified when the knowledge base device receives an entered mapping of the subsumption pair;

[0128] 12) STXM (SSTXC:stxs, FSTXC*:stxf, MAPC:<_ind>): STXM, syntax form correspondence table; SSTXC, substantive code field; FSTXC, formal code field; MAPC: variable bit name correspondence field; one record represents that the syntax form stxf represents has the substantive code stxs, and the syntax forms with equivalent relation have the same substantive code; <_ind> represents that the variable bit name "the i-th" used in the data record field definition of the syntax form corresponds to the variable bit name "the <_ind>(i)-th" used in the attribute (i.e. the syntax substantive) description of the syntax form; in the preparation stage of the inference computation, the knowledge base device generates the table content: 1) using the set Eq to record the records of the STXEF table that have been processed; using the set Tdo to record the syntax form codes to be processed; 2) if Tdo is empty, taking one record in the STXEF table that is not in the set Eq, recording its XSTXC field as stxif, creating a substantive code for the formal code stxif, recording as stxis, creating an integer sequence from 1 to Len(STXF[IDC=stxif].VARSC), recording as <_ordi>, indicating that the corresponding way of the variable bit names of the syntax substantive and the variable bit names of the syntax form is the same name correspondence; creating a table record {stxis,stxif,<_ordi>}; recording stxif in the set Tdo; 3) if Tdo is not empty, removing one code in it, recording as stxaf; then, finding the syntax equivalence relation composed of stxaf, i.e. the union set of the record set STXEF[XSTXC=stxaf] and the record set STXEF[YSTXC=stxaf]; after removing the records in Eq in the set, for each record in the set, recording as Rec, performing the operation: 3.1) if Rec.YSTXC is equal to stxaf, recording Rec.XSTXC as stxbf, otherwise recording Rec.YSTXC as stxbf; because of the syntax equivalence, stxbf represents the same substantive code as the substantive code of the syntax form represented by stxaf, recording as stxas, which can be obtained through STXM[FSTXC=stxaf].SSTXC; at this time, the data record STXM[FSTXC=stxaf] describes the name correspondence between the variable bits of the syntax form stxaf and the variable bits of the syntax substantive stxas in the MAPC field, the data record Rec describes the correspondence between the variable bits of the syntax form stxaf and the variable bits of the syntax form stxbf in the MAPC field, through the two correspondence relations, the correspondence between the variable bits of the syntax form stxbf and the variable bits of the syntax substantive stxas can be obtained, recording as <_ordb>; 3.2) If STXM[FSTXC=stxbf] does not exist, create a record {stxas, stxbf, <_ordb>} in this table, and put stxbf into the set Tdo; 3.3) put Rec into the set Eq; 4) repeat the above steps 2)-3) until the operation cannot continue; 5) in the syntax form code appearing in the IDC field of the STXF table, remove the syntax form code appearing in the FSTXC field of this table, and for each of the remaining syntax form codes, record it as stxjf; create a substantive code for the form code stxjf, record it as stxjs; create a sequence of integers from 1 to Len(STXF[IDC=stxjf].VARSC), record it as <_ordj>; create a record {stxjs, stxjf, <_ordj>} in this table.

[0129] 13) CTXM(SCTXC:ctxs, FSYNC*:synf): CTXM, context contrast table; SCTXC, context field; FSYNC, sync form field; a record indicates that the context of the sync form synf can be represented by the substantive code ctxs; in the preparation stage of inference computation, the knowledge base device generates the content of this table: 1) construct a graph data structure, whose nodes are all the codes appearing in the IDC field of the SYNF table, and whose edges correspond to the records of the SYNEF table, connecting the code in the XSYNC field and the code in the YSYNC field; 2) calculate all the maximum connected subgraphs in the graph, and the nodes without edges are regarded as being in the single-node subgraph consisting of themselves; 3) for each maximum connected subgraph, record it as Gr, create a context substantive code, record it as ctxrs, and for each node in Gr, record its code as synrf, create a record {ctxrs, synrf} in this table;

[0130] 14) CPTM(CPTC:cpts, LOCC*:locf): CPTM, concept contrast table; CPTC, concept field; LOCC, global location field; a record indicates that the concept substantive code cpts and the global location / concept form code locf represent the same concept; in the preparation stage of inference computation, the knowledge base device generates the content of this table: 1) construct a graph data structure, whose nodes are all the codes appearing in the IDC field of the LOCF table, and whose edges correspond to the records of the CPTEF table, connecting the code in the XCPTC field and the code in the YCPTC field; 2) calculate all the maximum connected subgraphs in the graph, and the nodes without edges are regarded as being in the single-node subgraph consisting of themselves; 3) for each maximum connected subgraph, record it as Gr, create a concept substantive code, record it as cptrs, and for each node in Gr, record its code as locrf, create a record {cptrs, locrf} in this table;

[0131] 15) SYNS (IDC*:syns, STXC:stxs, CTXC:ctxs, VALSC: <cpts>TYPC:_typ,STAC:_sta):SYNS, synthetic essence table; IDC, identity code field; STXC, syntax field; CTXC, context field; VALSC, concept sequence field; TYPC, type field; STAC, state mark field; a record is a synthetic form represented by an essence code, including its identity code, attribute, state; in the preparation stage of inference calculation, the knowledge base device generates the initial content of the table; at the beginning of inference calculation, the knowledge base device changes the STAC field of the record of the table with the essence code of the hypothesis synthetic form entered by the user as the primary key to newly created; in the inference process, the knowledge base device uses the table as the record table of synthetic forms; the initial content of the table is obtained by essence operation on the records of the SYNF table; a record in the SYNF table is denoted as Rec, and the synthetic form code in the IDC field of the record is synrf; the essence operation on the record is as follows: the syntax form code in Rec.STXC is essenced to obtain a syntax essence code, denoted as stxrs; the context essence code of the synthetic form is looked up in CTXM[FSYNC=Rec.IDC].SCTXC, denoted as ctxrs; the concept form code sequence of Rec.VALSC is arranged in accordance with the syntax form Rec.STXC, denoted as <locrf>The concept essence code sequence of the essence of the concept is denoted as <cptrs>and records the correspondence between the positions before the rearrangement and the positions after the rearrangement in the materialization process, denoted as <_indr>, which represents, <locrf>The i-th position becomes after substantialization <cptrs>The i-th position in the middle of the inductor; Rec. TYPC is unchanged in the substantial operation, denoted as _typr; the STAC field of the initial record in the SYNS table is null, indicating no marked state; the knowledge base device confirms that there is a record in the table, and the stxrs, ctxrs, <cptrs>, record of _typr, null, record its IDC field as synrs; add record {synrs, synrf, <_indr>} in the following SYNM;

[0132] 16) SYNM (SSYNC: syns, FSYNC: synf, MAPC: <_ind>) : SYNM, Syn Form Control Table; SSYNC, substantive code field; FSYNC, form code field; MAPC, position control field; a record represents that the substantive code syns of the syn form and the form code synf represent the same syn form, the position correspondence relationship between the concept sequence represented by the form code of the syn form form synf recorded in the SYNF table and the concept sequence represented by the substantive code of the syn form syns recorded in the SYNS table is <_ind>; in the preparation stage of reasoning calculation, the knowledge base device generates the content of this table when generating the initial content of the SYNS table;

[0133] 17) ANAS (IDC*: anas, CSYNC: syns, RSYNC: synf, FVALSC: <cpts>GVALSC: <cpts>, MAPC: _ind, TYPC: _mod, STAC: _sta): ANAS, analysis formula essence table; IDC, identity code field; CSYNC, as a result of synthesis field; RSYNC, as a result of synthesis field; FVALSC, filter part concept sequence field; GVALSC, generator part concept sequence field; MAPC, source relation field; TYPC, type field; STAC, state mark field; a record is an analysis formula represented by an essence code, including its identity code, attribute, state; in the preparation stage of inference calculation, the knowledge base device generates the initial content of the table; in the inference calculation process, the knowledge base device uses the table as the record table of analysis formula; the initial content of the table is obtained by essence operation on the record of the ANAF table, and a record in the ANAF table is Rec, and the analysis formula form code of the IDC field of Rec is anarf; the essence operation on the record is to record the synthesis formula form codes of Rec.CSYNC and Rec.RSYNC as csynrf and rsynrf, and essence them to obtain the synthesis formula essence code, recorded as csynrs and rsynrs; in the STXC field of the record with csynrs and rsynrs as the primary key in the SYNF table, the syntax formula form codes of the two synthesis formula forms are found, recorded as cstxrf and rstxrf, at this time, the filter part concept sequence and the generator part concept sequence represented by the concept form code sequence in Rec.FVALSC and Rec.GVALSC are recorded as <flocrf> 、 <glocrf>, the code arrangement order of which respectively follows the syntaxes cstxrf and rstxrf, and the concept sequences of the screening unit and the generating unit are represented by the concept substantive code sequences, denoted as <fcptrs> 、 <gcptrs>and record the position correspondence before and after rearrangement in the materialization process of the two concept sequences, record them as <_findr>, <_gindr>; <_findr> represents, <flocrf>the i-th position in the vector x becomes <fcptrs>the <_findr>(i)th position in the <_gindr> denotes, <glocrf>the i-th position in the vector x becomes <gcptrs>the <gindr>(i)th position in the Rec.SRCC field; and the correspondence between the two positions and the Rec.SRCC field record <flocrf> 、 <glocrf>The positional correspondence between them is represented as follows. <fcptrs> 、 <gcptrs>The sequence of the position correspondence relationship between the two positions is denoted as <_sind>; <_sind> and <gcptrs>isometric, denoted <gcptrs>the i-th position in the <fcptrs>the <_sind>(i)-th position in the _sind array; Rec.TYPC is unchanged in the materialization operation, denoted as _modr; the STAC field of the initial record in the ANAS table is empty, indicating no annotation state; the knowledge base device confirms that there is a record in the table whose fields except IDC are respectively csynrs, rsynrf, <fcptrs> 、 <gcptrs>, record its IDC field as anars; add a record {anars, anarf, <_findr>, <_gindr>} in the following analysis formula correspondence table;

[0134] 18) ANAM (SANAC: anas, FANAC*: anaf, FMAPC: <_ind>, GMAPC: <_ind>) : ANAM, analysis formula correspondence table; SANAC, substantive code field; FANAC, formal code field; FMAPC, filter part position correspondence field; GMAPC, generator part position correspondence field; a record represents that the substantive code anas and the formal code anaf of an analysis formula represent the same analysis formula; the position correspondence relationship between the filter concept sequence and the generator concept sequence represented by the formal code of the analysis formula form anaf in the ANAF table and the filter concept sequence and the generator concept sequence represented by the substantive code of the analysis formula anas in the ANAS table is <_ind> of the FMAPC field and <_ind> of the GMAPC field, respectively; in the preparation stage of reasoning calculation, the knowledge base device generates the content of this table when generating the initial content of the ANAS table;

[0135] 19) PRDS (IDC*: prds, FSYNC: syns, KEYSC: <_ind>) : PRDS, predicate formula substantive table; IDC, code field; FSYNC, frame synthesis formula field; KEYSC, key variable position field; a record is a predicate formula represented by a substantive code, including its identity code, attribute, and state; in the preparation stage of reasoning calculation, the knowledge base device generates the initial content of this table; the initial content of this table is obtained by performing a substantiation operation on a record Rec in the PRDF table, whose analysis formula formal code in the IDC field is prdrf; perform a substantiation operation on it, record the frame synthesis formula formal code of Rec.FSYNC as fsynrf, and the frame synthesis formula substantive code obtained by substantiating it is recorded as fsynrs; record the concept sequence represented by the substantive code of fsynrs in the record of the SYNS table with fsynrs as the primary key as <fcptrs>; the concept sequence represented by the form code in the record of the SYNF table whose primary key is the form synrf in the form composite form fsynrf is <flocrf>, fsynrf in the SYNM table record whose primary key is fsynrf, the position of the matching field is <_mindr>, at this time, the sequence of the record Rec.KEYSC field is <_findr>, the sequence in <_findr> indicates the <flocrf>The position of the middle key variable bit, <_mindr>, indicates <flocrf>With <fcptrs>The position correspondence between the two images is obtained according to the <_findr> and the <_mindr> <fcptrs>The sequence of the key variable bit is arranged in size to obtain a sequence, denoted as <_sindr>; the knowledge base device confirms that there is a record in the table, in which the fields other than the IDC field are respectively fsynrs and <_sindr>; the IDC field is denoted as prdrs; a record {prdrs, prdrf} is added in the following predicate formula comparison table;

[0136] 20) PRDM (SPRDC: prds, FPRDC*: prdf): PRDM, predicate formula comparison table; SPRDC, substantive code field; FPRD, formal code field; a record indicates that the substantive code prds and the formal code prdf of the predicate formula represent the same analysis formula; in the preparation stage of the reasoning calculation, the knowledge base device generates the content of the table when generating the initial content of the PRDS table;

[0137] 21) ASES (IDC*: ases, ANAC: anas, SYNC: syns, MAPC: <_ind>): ASES, analysis-synthesis formula substantive table; IDC, code field; ANAC, analysis formula field; SYNC, synthesis formula field; MAPC, replacement relationship field; a record is a split-synthesis formula represented by a substantive code, including its identity code, attribute, and state; in the preparation stage of the reasoning calculation, the knowledge base device generates the initial content of the table; the initial content of the table is obtained by performing a substantiation operation on the record of the ASEF table, denoted as Rec, in which the split-synthesis formula formal code of the IDC field is aserf; the analysis formula formal code and the synthesis formula formal code in anas and syns are denoted as anarf and synrf, which are substantiated to obtain the analysis formula substantive code and the synthesis formula substantive code, denoted as anars and synrs; anars is arranged in the screening part concept sequence and the generating part concept sequence represented by the substantive code in the record of the ANAS table with anars as the primary key <afcptrs> 、 <agcptrs>, the concept sequence expressed in the substantive code in the record of the SYNS table with the synrs as the primary key <scptrs>; record analysis form anarf the filter concept sequence and the generator concept sequence indicated by the form code in the ANAF table record whose primary key is the form anarf <aflocrf> 、 <aglocrf>, the concept sequence represented by the form code in the location correspondence field record of the SYNF table whose form synrf is the primary key is <slocrf>; the screening part position reference field and the generating part position reference field in the ANAM table record whose main key is in the form of anarf are <_afmindr>, <_agmindr>, and the position reference field <_smindr> in the SYNM table record whose main key is in the form of synrf; at this time, the RPLC field of Rec describes <slocrf>With <aflocrf> 、 <aglocrf>The position correspondence between the two is indicated by <smindr> which indicates <slocrf>With <scptrs>The position correspondence between the two is indicated by the <afmindr> tag <aflocrf>With <afcptrs>The position correspondence between the two is indicated by <agmindr> <aglocrf>With <agcptrs>corresponding relationship between the positions, and according to the position corresponding relationship, the position of the object in the image is obtained <scptrs>With <afcptrs> 、 <agcptrs>The position correspondence between the two is recorded as <_indr, which represents: <scptrs>the i-th position of the <indr>(i) when <indr>(i) is a positive integer, <afcptrs>the <_indr>(i)th position; 2) when <_indr>(i) is a negative integer, let its absolute value be |<_indr>(i)|, <agcptrs>the <_indr)(i)th position in <_indr); 3) when <indr>(i) corresponds to 0 <aglocrf> 、 <agcptrs>IDC: idc, ANAC: anac, STAC: sta) : RESM, relation evidence set map; IDC, code field; ANAC, analysis field; STAC, state mark field; a record represents a relation evidence set, including its analysis and state; in the preparation stage of inference calculation, the knowledge base device generates the content of this table when generating the initial content of the RESM table;

[0138] 22) ASEM (SASEC: ases, FASEC: asef) : ASEM, abstract mapping form correspondence table; SASEC, substantive code field; FASEC, form code field; a record represents that the substantive code ases and the form code asef of the abstract mapping form represent the same abstract mapping form; in the preparation stage of inference calculation, the knowledge base device generates the content of this table when generating the initial content of the ASES table;

[0139] 23) SEKS (IDC: seks, HYPC: syns, ANAC: anas, STAC: _sta) : SEKS, relation evidence set substantive table; IDC, code field; HYPC, hypothesis synthesis field; ANAC, analysis field; STAC, state mark field; a record is a relation evidence set represented by a substantive code, including its associated hypothesis synthesis and analysis, and the state of the relation evidence set; in the inference process, the knowledge base device uses this table as the record table of the relation evidence set;

[0140] 24) TFMS (IDC: tfms, ASEC: ases, HYPC: syns, ANAC: anas, STAC: _sta) : TFMS, transformation relation substantive table; IDC, code field; ASEC, abstract mapping form field; HYPC, hypothesis synthesis field; ANAC, analysis field; STAC, state mark field; a record is a transformation relation represented by a substantive code, including its associated abstract mapping form, hypothesis synthesis, analysis, and the state of the transformation relation; in the inference process, the knowledge base device uses this table as the record table of the transformation relation;

[0141] 25) RESS (IDC: ress, GRPC: syns, VALSC: <cpts>STAC:STAC, status tag; one record for one inference result represented by the essence code, including its group tag synthesis, concept sequence, and the status of the inference result; during the inference process, the knowledge base device uses this table as the record table of the inference result;

[0142] 26) SETX(IDC*:setx, ITMSC: <cpts>SETX, set composition table; IDC, code field; ITMSC, constituent concept field; a record represents that the set setx defined by its elements includes a group of concepts whose substantial codes are arranged in size order <cpts>; In the reasoning process, the knowledge base device uses this table to record the concept sets generated by the "set" reasoning method;

[0143] 27) ENSS (CPTC: cpts, SETC: setx): ENSS set concept essence table; CPTC, concept field; SETC, set field; a record indicates that the meaning of concept essence code cpts refers to set setx; in the reasoning process, the knowledge base device uses this table to record the reference relationship between the set concept generated by the "set" reasoning method and the specific concept set;

[0144] 28) EVDS (SUPC: anys, HASC: anys): EVDS, evidence essence table; SUPC, supported field; HASC: owned field; a record indicates that there is a piece of evidence, which is supported by the thought map module unit anys in the SUPC field and owned by the thought map module unit anys in the HASC field; anys refers to a non-specific category of thought map module unit; in the reasoning process, the knowledge base device uses this table to record the proof relationship; since the proof relationship is not the attribute of any other thought map module unit, there is no need to record its identity code;

[0145] After starting reasoning, the knowledge base device uses the aforementioned thought map module unit essence table to perform reasoning calculation; the reasoning calculation module operates the data records in the essence table according to the aforementioned method; and reports the results required by the user;

[0146] When the knowledge base device reports the comprehensive type represented by the essence code, it is recorded as syns; the syntax essence code is recorded as stxs, and the concept sequence is recorded as <cpts>; remember <cpts>The i-th element of the set is cptis; the set of formal codes that represent the same syntax as the syntax code stxs can be obtained by STXM[SSTXC=stxs].FSTXC, which is called the equivalent syntax form set, and a code in the set is denoted as stxjf; a reference sequence that can be used to represent the synthesis syns is STXF[IDC=stxjf].SYMSC, which is based on the global position code representation, and can be converted into a text block sequence by means of the LOCF table and the SYMX table; the variable bit of the synthesis syns, i.e. <cpts>The correspondence between the sequence positions in the sequence and the sequence positions in the text block sequence can be obtained based on STXM[FSTXC=stxjf].MAPC and STXF[IDC=stxjf].VARSC; the set of form codes of the integrated form entered by the user using the syntax form stxjf can be obtained by SYNF[STXC=stxjf].IDC, which is called the equivalent integrated form set, and a code in the set is denoted as synkf; a reference sequence that can be used to represent the integrated form syns is SYNF[IDC=synkf].SYMSC, which is based on the global position code representation and can be converted into a text block sequence by the LOCF table and the SYMX table; the correspondence between the variable positions of the integrated form syns and the sequence positions of the text block sequence can be obtained based on STXM[FSTXC=stxjf].MAPC, STXF[IDC=stxjf].VARSC, and SYNF[IDC=synkf].VALSC; on the other hand, the text block for representing the concept cptis can be obtained by SYMX[IDC=LOCF[IDC=CPTM[CPTC=cptis].LOCC].SYMC].TXTC; based on the above reference sequence and the concept text block option, the user is reported the synthesized text block sequence.

[0147] Specific knowledge base device use cases

[0148] In the following case, the knowledge base device uses the following 6 special locf codes to represent 6 special concepts: i) 1: pending; ii) 2: arbitrary; iii) 3: reference; iv) 4: set; v) 5: decomposition; vi) 6: reflection.

[0149] The following describes a use case of the knowledge base device, in which the user reads an article, understands it, and instructs the knowledge base device to create a database table structure.

[0150] The article read by the user is as follows:

[0151] On the first day of school, the staff need to be tested for some indicators. The test results are as follows: Zhang San's fasting blood glucose is 7.8 mmol / L, Li Si's fasting blood glucose is 8.2 mmol / L, and Wang Wu's fasting blood glucose is 6.9 mmol / L. The diagnostic cutoff value for fasting blood glucose is 7.0 mmol / L. If the staff's fasting blood glucose value is greater than the diagnostic cutoff value, the staff is diagnosed as having diabetes based on fasting blood glucose. If the staff does not have a postprandial 2h blood glucose test result, he has not had a postprandial 2h blood glucose test, which may lead to misdiagnosis. In this case, it is recommended that the staff supplement the postprandial 2h blood glucose test. Which staff need to submit a new test application to the school medical center?

[0152] The following comprehensive sentences exist in this article, and the user enters the original text sequence subsequences based on these comprehensive sentences:

[0153] 1) Zhang San's fasting blood glucose is 7.8 mmol / L

[0154] 2) Li Si's fasting blood glucose is 8.2 mmol / L

[0155] 3) Wang Wu's fasting blood glucose is 6.9 mmol / L

[0156] 4) The diagnostic cutoff value for fasting blood glucose is 7.0 mmol / L

[0157] 5) The value of the staff's fasting blood glucose ("if greater than the diagnostic cutoff value") The faculty member's value for fasting blood glucose 6) The diagnostic cutoff value for fasting blood glucose ("if the staff's value is greater than

[0158] ") The faculty member's value for fasting blood glucose The diagnostic cutoff value 7) The value of the staff's fasting blood glucose diagnostic cutoff value ("if greater than ")

[0159] The faculty member's value for fasting blood glucose The diagnostic cutoff value

[0160] 8) The value greater than the diagnostic cutoff value ("if the staff's fasting blood glucose The value is greater than the diagnostic cutoff value ")

[0161] 9) If the staff's fasting blood glucose value is greater than the diagnostic cutoff value

[0162] 10) The staff is diagnosed as diabetic according to fasting blood glucose

[0163] 11) The staff does not have a postprandial 2h blood glucose test result

[0164] 12) He did not do a postprandial 2h blood glucose test

[0165] 13) The staff is diagnosed as diabetic according to fasting blood glucose, and he did not do a postprandial 2h blood glucose test ("then The faculty member is diagnosed with diabetes based on fasting blood glucose . If the staff does not have a postprandial 2h blood glucose test result He does not have a 2-hour postprandial glucose test then ， ") The faculty member is diagnosed with diabetes based on fasting blood glucose He does not have a 2-hour postprandial glucose test, which can lead to misdiagnosis )

[0166] 14) The staff is diagnosed as diabetic according to fasting blood glucose, and he did not do a postprandial 2h blood glucose test, which may lead to misdiagnosis ("then Figure 4 . If the staff does not have a postprandial 2h blood glucose test result ， then Figure 5 ")

[0167] ​​15) In this case, the staff member is recommended to have a 2h postprandial glucose test.

[0168] 16) These people need to submit a new test request to the hospital

[0169] 17) Which staff members need to submit a new test request to the hospital?

[0170] The syntactic sequences of the syntactic patterns created by the user based on their understanding of the article include:

[0171] 1) The test result is the staff member's index

[0172] 2) The staff member's index is the test result

[0173] 3) The threshold of the index is the diagnostic cutoff

[0174] 4) The staff member's index is the test result, and its threshold is the diagnostic cutoff

[0175] 5) There is a relationship between the test result and the diagnostic cutoff

[0176] 6) The staff member's index is the test result, and its threshold is the diagnostic cutoff, and there is a relationship between them

[0177] 7) The staff member is diagnosed based on the index

[0178] 8) The staff member has not been tested

[0179] 9) The staff member is diagnosed based on the index and has not been tested

[0180] 10) This situation refers to the staff member being diagnosed based on the index and not having been tested

[0181] 11) In this case, the staff member is encouraged to have the recommended test

[0182] 12) The population needs to submit a new test request to the hospital

[0183] The analysis sequences of the analysis patterns created by the user based on their understanding of the article include:

[0184] 1) If a staff member's index is a test result, then the staff member's index is the test result and the index has a diagnostic cutoff

[0185] 2) If the threshold of the index is the diagnostic cutoff, then the staff member's index is the test result and the index has the diagnostic cutoff

[0186] 3) If a staff member's index is a test result and the index has a diagnostic cutoff, then the staff member's index is the test result and the index has the diagnostic cutoff, and there is a relationship between the index and the test result

[0187] 4) If the result of the detection and the diagnostic limit value exist in the relationship, the index of the staff is the result of the detection and the index exists in the diagnostic limit value, and there is a relationship between the result of the detection and the diagnostic limit value

[0188] 5) If the staff's fasting blood glucose has a detection result and is greater than the diagnostic limit value, the staff is diagnosed as diabetic according to the index

[0189] 6) If the index of the staff has no detection result, the staff has not done the check

[0190] 7) If the staff is diagnosed according to the index, the staff is diagnosed according to the index and has not done the check

[0191] 8) If the staff has not done the check, the staff is diagnosed according to the index and has not done the check

[0192] 9) If the staff is diagnosed according to the index and has not done the check, this case means that the staff is diagnosed according to the index and has not done the check

[0193] 10) If the case means that the staff is diagnosed according to the index and has not done the check, the staff is encouraged to do the recommended check in this case 11) If the staff is encouraged to do the recommended check in this case, these people need to submit a new check application to the school hospital Figure 6 As shown in the following steps:

[0194] 1) The user divides the natural language text block, enters the above-mentioned original text sequence, original text sequence sub-sequence and explanation sequence; the knowledge base device creates 74 records in SYMX based on the above-mentioned sequence entered by the user (see the first 74 records of the SYMX table for details Figure 5 ), creates 365 records in LOCF (see the first 365 records of the LOCF table for details Figure 6 ), and creates 365 records in the ARTX table (see the first 365 records of the ARTX table for details Figure 7 ); the knowledge base device creates 12 records in STXF according to the syntax sequence marked by the user (see the STXF table for details Figure 6 ); the knowledge base device creates 11 records in ANAF according to the analysis sequence marked by the user (see the ANAF table for details Figure 6 ); and the knowledge base device creates 17 records in SYNF according to the comprehensive sequence marked by the user (see the SYNF table for details Figure 6 );

[0195] 2) The user understands the knowledge of the original text and marks the predicate form on the original text sequence, and the knowledge base device creates 1 record in PRDF according to the user's marking (see the PEDF table for details Figure 7 );

[0196] 3) User based on the original text knowledge understanding, in the original text sequence and interpretation sequence marked the same concept, the knowledge base device according to the user marked interpretation sequence, in the ASEF created 2 records (see details Figure 6 ASEF table);

[0197] 4) User based on the original text knowledge understanding, in the original text sequence and interpretation sequence marked the same concept, the knowledge base device according to the user marked interpretation sequence, in the CPTEF created 95 records (see details Figure 6 CPTEF table of the first 95 records);User based on the original text knowledge understanding, in the original text sequence and interpretation sequence marked the same concept, the knowledge base device according to the user marked interpretation sequence, in the STXEF created 1 records (see details Figure 7 STXEF table);

[0198] 5) User based on the original text knowledge understanding, in the original text sequence marked the comprehensive sequence with context, the knowledge base device according to the user marked

[0199] interpretation sequence, in the SYNEF created 16 records (see details Figure 5 SYNEF table);

[0200] In the preparation stage of reasoning calculation, the knowledge base device generates the following data table content:

[0201] 6) The knowledge base device establishes the association between the concept form code and the essence code, and creates 106 records in the CPTM (see details Figure 5 CPTM table of the first 106 records);

[0202] 7) The knowledge base device establishes the association between the syntax form code and the essence code, and creates 12 records in the STXM (see details Figure 7 STXM table);

[0203] 8) The knowledge base device establishes a graph data structure, and generates the context code corresponding to the comprehensive form according to the SYNEF, and creates

[0204] 17 records in the CTXM (see details Figure 5 CTXM table);

[0205] 9) The knowledge base device converts the comprehensive form code into the essence code, and creates 17 records in the SYNS (see details Figure 7 SYNS table)。

[0206] The knowledge base device establishes the association between the comprehensive form code and the essence code based on the newly generated essence code, and creates 17 records in the SYNM (see details Figure 7 SYNM table)。

[0207] 10) The KB device converts the analytical formal code into the substantial code, creating 11 records in ANAS (see the ANAS table below). Figure 7

[0208] The KB device establishes the association between the analytical formal code and the substantial code based on the newly generated substantial code, creating 11 records in ANAM (see the ANAM table below). Figure 5

[0209] 11) The KB device converts the predicate formal code into the substantial code, creating 1 record in PRDS (see the PRDS table below). The KB device establishes the association between the predicate formal code and the substantial code based on the newly generated substantial code, creating 1 record in PRDM (see the PRDM table below). Figure 7 Figure 7

[0210] 12) The KB device converts the aspect mapping formal code into the substantial code, creating 2 records in ASES (see the ASES table below). The KB device establishes the association between the aspect mapping formal code and the substantial code based on the newly generated substantial code, creating 2 records in ASEM (see the ASEM table below). ​ ​

[0211] The KB device performs reasoning according to the reasoning target (hypothesis synthesis) proposed by the user:

[0212] 13) When the hypothesis synthesis 660 is marked as newly created, the execution transaction unit is triggered and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 660 are deleted, and the analytical code 671 that can confirm the existence of a reasoning result that meets the meaning of 660 is found by the reasoning calculation module. Since the screening condition meets the meaning of 671 screening unit, it is confirmed that there is a hypothesis synthesis:

[0213] {675, 641, 643, <1, 1, 1>, H, newly created}, and a proof relation (660, 675) is created, and a looped association {676, 675, 671, newly created} is confirmed to exist, and a proof relation (660, 676) is created.

[0214] ​​​​​​14) Since the hypothesis composition 675, which was confirmed to exist in the above step 13), is marked with a newly created state, a transaction unit is triggered to be executed and immediately loses the state after the transaction unit is executed: all proof relations supported by 675 are deleted, and the inference calculation module finds the analysis 670 that can confirm to exist a reasoning result that conforms to the meaning of 675. Since the screening condition conforms to the meaning of the screening unit 670, a hypothesis composition is confirmed to exist: {677, 640, 643, <1, 1, 1, 1, 1>, H, newly created}, and a proof relation (675, 677) is created, and a looped evidence association {678, 677, 670, newly created} is confirmed to exist, and a proof relation (675, 678) is created.

[0215] 15) Since the hypothesis composition 677, which was confirmed to exist in the above step 14), is marked with a newly created state, a transaction unit is triggered to be executed and immediately loses the state after the transaction unit is executed: all proof relations supported by 677 are deleted, and the inference calculation module finds the analysis 669 that can confirm to exist a reasoning result that conforms to the meaning of 677. Since the screening condition conforms to the meaning of the screening unit 669, a hypothesis composition is confirmed to exist: {679, 639, 643, <1, 1, 1, 1>, H, newly created}, and a proof relation (677, 679) is created, and a looped evidence association {680, 679, 669, newly created} is confirmed to exist, and a proof relation (677, 680) is created.

[0216] 16) Since the hypothesis composition 679, which was confirmed to exist in the above step 15), is marked with a newly created state, a transaction unit is triggered to be executed and immediately loses the state after the transaction unit is executed: all proof relations supported by 679 are deleted, and the inference calculation module finds the analyses 667 and 668 that can confirm to exist a reasoning result that conforms to the meaning of 679. For the analysis 667, since the screening condition conforms to the meaning of the screening unit, a hypothesis composition is confirmed to exist: {681, 637, 643, <1, 1, 1>, H, newly created}, and a proof relation (679, 681) is created, and a looped evidence association {682, 681, 667, newly created} is confirmed to exist, and a proof relation (679, 682) is created. For the analysis 668, since the screening condition conforms to the meaning of the screening unit, a hypothesis composition is confirmed to exist:

[0217] {683, 638, 643, <1, 1>, H, newly created}, and a proof relation (679, 683) is created, and a looped evidence association

[0218] {684, 683, 668, newly created} is created, and a proof relation (679, 684) is created.

[0219] 17) Since the hypothesis composition 681 confirmed to exist in the above step 16) is marked with a newly created state, a transaction unit is triggered to be executed and lost immediately after the execution of the transaction unit: all the proof relations supported by 681 are deleted, and the inference computing module finds the analysis 665 that can confirm to exist the inference result that conforms to the meaning of 681. Since the screening condition conforms to the meaning of the screening part, a hypothesis composition is confirmed to exist: {685, 636, 643, <1, 609, 1, 611>, H, newly created}, and a proof relation (681, 685) is created, and a looped evidence association {686, 685, 665, newly created} is confirmed to exist, and a proof relation (681, 686) is created.

[0220] 18) Since the hypothesis composition 683 confirmed to exist in the above step 16) is marked with a newly created state, a transaction unit is triggered to be executed and lost immediately after the execution of the transaction unit: all the proof relations supported by 683 are deleted, and the inference computing module finds the analysis 666 that can confirm to exist the inference result that conforms to the meaning of 683. Since the screening condition does not conform to the meaning of the screening part of 666, the knowledge base device finds the sub-composition mapping 674 participated by 666, and confirms to exist a hypothesis composition:

[0221] {687, 637, 643, <1, 612, 610>, H, newly created}, and a proof relation (683, 687) is created. At the same time, a transformation association {688, 674, 687, 666, newly created} is confirmed to exist, and a proof relation (683, 688) is created.

[0222] 19) Since the hypothesis composition 685 confirmed to exist in the above step 17) is marked with a newly created state, a transaction unit is triggered to be executed and lost immediately after the execution of the transaction unit: all the proof relations supported by 685 are deleted, and the inference computing module finds the analysis 663 and 664 that can confirm to exist the inference result that conforms to the meaning of 685. For the analysis 663, since the screening condition conforms to the meaning of the screening part, a hypothesis composition is confirmed to exist: {689, 634, 643, <1, 609, 1, 1>, H, newly created}, and a proof relation (685, 689) is created, and a looped evidence association {690, 689, 663, newly created} is confirmed to exist, and a proof relation (685, 690) is created. For the analysis 664, since the screening condition does not conform to the meaning of the screening part of 664, the knowledge base device finds the sub-composition mapping 673 participated by 664, confirms to exist a recorded hypothesis composition 689, and confirms to exist a transformation association {692, 673, 689, 664, newly created}, and a proof relation (685, 692) is created.

[0223] 20) Since the hypothesis synthesis 689 marked as newly created in the above step 19) is confirmed to exist, a transaction unit is triggered to be executed and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 689 are deleted, the inference computing module finds the analysis formulas 661 and 662 that can confirm the existence of the inference result that conforms to the meaning of 689. For the analysis formula 661, since the screening condition conforms to the meaning of the screening unit, a hypothesis synthesis is confirmed to exist: {693, 632, 643, <1, 609, 1>, H, newly created}, and a proof relation (689, 693) is created, and a looped evidence association {694, 693, 661, newly created} is confirmed to exist, and a proof relation (689, 694) is created. For the analysis formula 662, since the screening condition conforms to the meaning of the screening unit, a hypothesis synthesis is confirmed to exist:

[0224] {695, 633, 643, <609, 1>, H, newly created}, and a proof relation (689, 695) is created, and a looped evidence association {696, 695, 662, newly created} is confirmed to exist, and a proof relation (689, 696) is created.

[0225] 21) Since the looped evidence association 694 marked as newly created in the above step 20) is triggered to be executed and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 694 are deleted; then, the inference computing module finds the canonical synthesis formulas 644, 645 and 646 that conform to the meaning of the hypothesis synthesis 693, since 661 is a positive analysis formula, 644, 645 and 646 are used as the basis of the thinking method stored by applying 661, and the thinking method stored by applying 661 is used to confirm the existence of 3 inference result records {697, 450, <621, 609, 622, 1>, newly created}, {698, 450, <623, 609, 624, 1>, newly created},

[0226] {699, 450, <625, 609, 626, 1>, newly created} and the corresponding 3 proof relations (694, 697), (694, 698),

[0227] (694, 699).

[0228] 22) Since the looped evidence association 696 marked as newly created in the above step 20) is triggered to be executed and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 696 are deleted; then, the inference computing module finds the canonical synthesis formula 648 that conforms to the meaning of the hypothesis synthesis 695, since 662 is a positive analysis formula, 648 is used as the basis of the thinking method stored by applying 662, and the thinking method stored by applying 662 is used to confirm the existence of 1 inference result record

[0229] {700, 450, <1, 609, 1, 627>, new creation} and one corresponding proof relation (696, 700).

[0230] 23) Since the reasoning result 700 in the above step 22) is marked as new creation, a transaction unit is triggered to be executed and the state is lost immediately after the execution of the transaction unit: mark 450 as pending synthesis.

[0231] 24) Since the synthesis 450 in the above step 23) is marked as pending synthesis, a transaction unit is triggered to be executed and the state is lost immediately after the execution of the transaction unit: delete all proof relations supported by 450, the reasoning calculation module finds all the reasoning results 697, 698, 699 and 700 marked with the synthesis 450 in the group, performs the synthesis process on them, and confirms that there are three canonical syntheses

[0232] {701, 634, 643, <621, 609, 622, 627>, N, new creation}, {702, 634, 643, <623, 609, 624, 627>, N, new creation}, {703, 634, 643, <625, 609, 626, 627>, N, new creation} and three corresponding proof relations (450, 701),

[0233] (450, 702), (450, 703).

[0234] 25) Since the canonical synthesis 701 in the above step 24) is marked as new creation, a transaction unit is triggered to be executed and the state is lost immediately after the execution of the transaction unit: the reasoning calculation module finds the hypothesis synthesis 689 that meets the meaning of 701, since 689 participates in the evidence association 690 and the transformation association 692, therefore mark 690 and 692 as new creation.

[0235] 26) Since the evidence association 690 in the above step 25) is marked as new creation, a transaction unit is triggered to be executed and the state is lost immediately after the execution of the transaction unit: delete all proof relations supported by 690; then, the reasoning calculation module finds the canonical syntheses 701, 702 and 703 that meet the meaning of the hypothesis synthesis 689, since 663 is a positive analysis, therefore 701, 702 and 703 are used as the basis of the thinking method stored by 663, and the thinking method stored by 663 is applied to confirm that there are three reasoning result records {704, 452, <621, 609, 622, 627, 1>, new creation}, {705, 452, <623, 609, 624, 627, 1>, new creation},

[0236] {706, 452, <625, 609, 626, 627, 1>, newly created} and the corresponding 3 proof relations (690, 704), (690, 705), (690, 706).

[0237] 27) Since the above step 25) the transformation association 692 is labeled with the newly created state, triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: 692 describes the transformation association of the abstracted mapping 673, the hypothetical generalized 689 and the analyzed 664; first, delete all the proof relations supported by 692; then, the reasoning calculation module finds the canonical generalizations 701, 702 and 703 that comply with the meaning of 689, and confirms that there are 3 analyzed 664 that comply with the meaning of 692.

[0238] {707, 651, 652, <622, 627, 2>, <1, 1, 3, 3, 3>, <0, 0, 1, 2, 3>, P},

[0239] {708, 651, 652, <624, 627, 2>, <1, 1, 3, 3, 3>, <0, 0, 1, 2, 3>, P}, {709, 651, 652, <626, 627, 2>, <1, 1,

[0240] 3, 3, 3>, <0, 0, 1, 2, 3>, P} and create 3 corresponding proof relations (692, 707), (692, 708), (692, 709);

[0241] 28) Since the above step 27) the analyzed 707 is labeled with the newly created state, triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: find all the reasoning results that 707 can confirm the existence of the hypothetical generalization 685 that complies with its meaning, and label the newly created state.

[0242] 29) Since the hypothesis synthesis 685 identified in step 28) above is confirmed to exist, it triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 685 are deleted, the analysis 664, 707, 708 and 709 that can confirm the existence of a reasoning result that is consistent with the meaning of 685 are found by the reasoning computation module. For the analysis 707, since the filter condition is consistent with the meaning of the filter part, a hypothesis synthesis is confirmed to exist: {710, 635, 643, <622, 627, 611>, H, newly created}, and a proof relation (685, 710) is created, and a looped evidence association {711, 710, 707, newly created} is confirmed to exist, and a proof relation (685, 711) is created. For the analysis 708, since the filter condition is consistent with the meaning of the filter part, a hypothesis synthesis is confirmed to exist: {712, 635, 643, <624, 627, 611>, H, newly created}, and a proof relation (685, 712) is created, and a looped evidence association {713, 712, 708, newly created} is confirmed to exist, and a proof relation (685, 713) is created. For the analysis 709,

[0243] Since the filter condition is consistent with the meaning of the filter part, a hypothesis synthesis is confirmed to exist:

[0244] {714, 635, 643, <626, 627, 611>, H, newly created}, and a proof relation (685, 714) is created, and a looped evidence association {715, 714, 709, newly created} is confirmed to exist, and a proof relation (685, 715) is created.

[0245] 30) Since the hypothesis synthesis 710 identified in step 29) above is confirmed to exist, it triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 710 are deleted, the predicate 672 that is consistent with the meaning of 710 is found, since 672 is associated with the "size comparison" task execution system that can be operated by the knowledge base, the computation engine returns the symbol sequence <"7.8", "7.0", "greater than">, and new records are created in LOCF, CPTEF, ARTX and CPTM: the primary keys 785-787 of LOCF, the primary keys 798, 795, 793 of CPTEF, the primary keys 785-787 of ARTX, and the primary keys 785-787 of CPTM. And a canonical synthesis {716, 635, 643, <622, 627, 611>, N, newly created} that describes the task result is confirmed to exist, and a proof relation (710, 716) is created.

[0246] 31) Since the hypothesis synthesis 712 confirmed to exist in the above step 29) is marked with the new creation state, a transaction unit is triggered to execute and loses the state immediately after the transaction unit is executed: all the proof relations supported by 712 are deleted, the predicate expression 672 that meets the meaning of 712 is found, and since 672 is associated with the "size comparison" task execution system that the knowledge base can operate, the computing engine returns the symbol sequence <"7.2", "7.0", "greater than">, and new records are created in LOCF, CPTEF, ARTX, and CPTM: the primary keys 788-790 of LOCF, the primary keys 794, 796, 799 of CPTEF, the primary keys 788-790 of ARTX, and the primary keys 788-790 of CPTM. And the normative synthesis {717, 635, 643, <624, 627, 611>, N, new creation} describing the task result is confirmed to exist, and the proof relation (712, 717) is created.

[0247] 32) Since the hypothesis synthesis 714 confirmed to exist in the above step 29) is marked with the new creation state, a transaction unit is triggered to execute and loses the state immediately after the transaction unit is executed: all the proof relations supported by 714 are deleted, the predicate expression 672 that meets the meaning of 714 is found, and since 672 is associated with the "size comparison" task execution system that the knowledge base can operate, the computing engine returns the symbol sequence <"6.9", "7.0", "less than">, and the following new records are created in SYMX, LOCF, CPTEF, ARTX, and CPTM: the primary key 718 of SYMX, the primary keys 791, 792, 719 of LOCF, the primary keys 797, 800 of CPTEF, the primary keys 791, 792, 719 of ARTX, and the primary keys 791, 792, 719 of CPTM. And the normative synthesis

[0248] {721, 635, 643, <626, 627, 720>, N, new creation} describing the task result is confirmed to exist, and the proof relation (714, 721) is created.

[0249] 33) Since the normative synthesis 716 in the above step 30) is marked with the new creation state, a transaction unit is triggered to execute and loses the state immediately after the transaction unit is executed: the hypothesis synthesis 710 that meets the meaning of 716 is found, and since 710 participates in the evidence association 711, 711 is marked with the new creation state.

[0250] 34) Since the normative synthesis 717 in the above step 30) is marked with the new creation state, a transaction unit is triggered to execute and loses the state immediately after the transaction unit is executed: the hypothesis synthesis 712 that meets the meaning of 717 is found, and since 712 participates in the evidence association 713, 713 is marked with the new creation state.

[0251] 35) Since the above-mentioned step 33) proof association 711 is marked with a newly created state, trigger the execution of a transaction unit and immediately lose the state after the execution of the transaction unit: delete all proof relationships supported by 711; Then, the reasoning calculation module finds the normative synthesis formula 716 that meets the meaning of the hypothesis synthesis formula 710, since 707 is a positive analysis formula, take 716 as the basis for applying the thinking method stored in 707, and confirm that there is 1 reasoning result record

[0252] {722, 452, <1, 1, 622, 627, 611>, newly created} and 1 corresponding proof relationship (711, 722).

[0253] 36) Since the above-mentioned step 34) proof association 713 is marked with a newly created state, trigger the execution of a transaction unit and immediately lose the state after the execution of the transaction unit: delete all proof relationships supported by 713; Then, the reasoning calculation module finds the normative synthesis formula 717 that meets the meaning of the hypothesis synthesis formula 712, since 708 is a positive analysis formula, take 717 as the basis for applying the thinking method stored in 708, and confirm that there is 1 reasoning result record

[0254] {723, 452, <1, 1, 624, 627, 611>, newly created} and 1 corresponding proof relationship (713, 723).

[0255] 37) Since the above-mentioned step 36) reasoning result 723 is marked with a newly created state, trigger the execution of a transaction unit and immediately lose the state after the execution of the transaction unit: mark 452 with a to-be-synthesized state.

[0256] 38) Since the above-mentioned step 37) synthesis formula 452 is marked with a to-be-synthesized state, trigger the execution of a transaction unit and immediately lose the state after the execution of the transaction unit: delete all proof relationships supported by 452, and find all reasoning results 704, 705, 706, 722 and 723 marked with synthesis formula 452, perform the synthesis process on them, and confirm that there are 2 normative synthesis formulas

[0257] {724, 636, 643, <621, 609, 622, 627, 611>, N, newly created}, {725, 636, 643, <623, 609, 624, 627, 611>,

[0258] N, newly created} and 2 corresponding proof relationships (452, 724), (452, 725).

[0259] 39) Since the above-mentioned step 38) normative synthesis 725 is marked with a newly created state, a transaction unit is triggered to be executed and immediately loses the state after the execution of the transaction unit: find 725 the hypothesis synthesis 685 that meets the meaning, since 685 participates in the evidence association 686, therefore mark 686 with a newly created state.

[0260] 40) Since the above-mentioned step 39) evidence association 686 is marked with a newly created state, a transaction unit is triggered to be executed and immediately loses the state after the execution of the transaction unit: delete all proof relations supported by 686; Then, the reasoning calculation module finds the normative synthesis 724 and 725 that meets the meaning of the hypothesis synthesis 685, since 665 is a positive analysis, then 724 and 725 are used as the basis of the thinking method stored by 665, and the thinking method stored by 665 is applied to confirm that there are 2 reasoning result records

[0261] {726,453,<621,609,610>,newly created}, {727,453,<623,609,610>,newly created} and the corresponding 2 proof relations (686,726), (686,727).

[0262] 41) Since the above-mentioned step 40) reasoning result 727 is marked with a newly created state, a transaction unit is triggered to be executed and immediately loses the state after the execution of the transaction unit: mark 453 with a to-be-synthesized state.

[0263] 42) Since the above-mentioned step 41) synthesis 453 is marked with a to-be-synthesized state, a transaction unit is triggered to be executed and immediately loses the state after the execution of the transaction unit: delete all proof relations supported by 453, and find all grouped reasoning results 726 and 727 marked with synthesis 453, perform the synthesis process on them, and confirm that there are 2 normative syntheses

[0264] {728,637,643,<621,609,610>,N,newly created}, {729,637,643,<623,609,610>,N,newly created} and 2 corresponding proof relations (453,728), (453,729).

[0265] 43) Since the above-mentioned step 42) normative synthesis 729 is marked with a newly created state, a transaction unit is triggered to be executed and immediately loses the state after the execution of the transaction unit: find 729 the hypothesis synthesis 681 that meets the meaning, since 681 participates in the evidence association 682, therefore mark 682 with a newly created state.

[0266] 44) Since the newly created state is marked by the above-mentioned step 43), the transaction unit is triggered to execute and immediately lose the state after the execution of the transaction unit: delete all the proof relations supported by 682; then, the inference calculation module finds the canonical generalizations 728 and 729 that conform to the meaning of the hypothesis generalization 681, and since 667 is a positive analysis formula, 728 and 729 are taken as the basis for applying the thinking method stored in 667, and the existence of 2 inference result records is confirmed by applying the thinking method stored in 667

[0267] {730, 456, <621, 609, 610, 1>, newly created}, {731, 456, <623, 609, 610, 1>, newly created} and the corresponding 2 proof relations (682, 730), (682, 731).

[0268] 45) Since the newly created state is marked by the above-mentioned step 18), the transformation association 688 is triggered to execute the transaction unit and immediately lose the state after the execution of the transaction unit: 688 describes the transformation association of the division generalization 674, the hypothesis generalization 687 and the analysis formula 666; first, delete all the proof relations supported by 688; then, the inference calculation module finds the canonical generalizations 728 and 729 that conform to the meaning of 687, and confirms the existence of 2 analysis formulas {732, 654, 655, <621, 612, 2>, <3, 3>, <1, 2>, N},

[0269] {733, 654, 655, <623, 612, 2>, <3, 3>, <1, 2>, N} and create 2 corresponding proof relations (688, 732), (688, 733); 46) Since the newly created state is marked by the above-mentioned step 45), the transaction unit is triggered to execute and immediately lose the state after the execution of the transaction unit:

[0270] find all the hypothesis generalizations 683 that conform to the meaning of the inference result that 733 may confirm the existence of, and mark the newly created state.

[0271] 47) Since the hypothesis synthesis 683 marked as newly created in the above step 46) is confirmed to exist, a transaction unit is triggered to be executed and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 683 are deleted, the analysis formulas 666, 732 and 733 are found by the inference calculation module to possibly confirm the existence of a reasoning result that conforms to the meaning of 683. For the analysis formula 732, since the screening condition conforms to the meaning of the screening part, a hypothesis synthesis is confirmed to exist: {734, 632, 643, <621, 612, 1>, H, newly created}, and a proof relation (683, 734) is created, and a looped evidence association {735, 734, 732, newly created} is confirmed to exist, and a proof relation (683, 735) is created. For the analysis formula 733, since the screening condition conforms to the meaning of the screening part, a hypothesis synthesis is confirmed to exist: {736, 632, 643, <623, 612, 1>, H, newly created}, and a proof relation (683, 736) is created, and a looped evidence association {737, 736, 733, newly created} is confirmed to exist, and a proof relation (683, 737) is created.

[0272] 48) Since the looped evidence association 735 marked as newly created in the above step 47) is triggered to be executed and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 735 are deleted; since there is no canonical synthesis that conforms to the meaning of the hypothesis synthesis 734 and 732 is a negative analysis formula, 734 is taken as the basis of the thinking method stored by applying 732, and the thinking method stored by applying 732 is applied, and it is confirmed that there is 1 reasoning result record {738, 455, <621, 612>, newly created} and the corresponding 1 proof relation (735, 738).

[0273] 49) Since the looped evidence association 737 marked as newly created in the above step 47) is triggered to be executed and loses the state immediately after the execution of the transaction unit: all the proof relations supported by 737 are deleted; since there is no canonical synthesis that conforms to the meaning of the hypothesis synthesis 736 and 733 is a negative analysis formula, 736 is taken as the basis of the thinking method stored by applying 733, and the thinking method stored by applying 733 is applied, and it is confirmed that there is 1 reasoning result record {739, 455, <623, 612>, newly created} and the corresponding 1 proof relation (737, 739).

[0274] 50) Since the reasoning result 739 marked as newly created in the above step 49) is triggered to be executed and loses the state immediately after the execution of the transaction unit: mark 438 as a state to be synthesized.

[0275] 51) Since the above step 50) the synthesis formula 455 is marked as to-be-synthesized state, trigger the execution of transaction unit and immediately lose the state after the execution of the transaction unit: delete all the proof relations supported by 455, and find all the reasoning results 738 and 739 marked synthesis formula as 455 of the group, execute the synthesis process on them, and confirm that there are 2 canonical synthesis formulas {740, 638, 643, <621, 612>, N,

[0276] {741, 638, 643, <623, 612>, N, newly created} and 2 corresponding proof relations (455, 740), (455, 741).52) Since the above step 51) the synthesis formula 741 is marked as newly created state, trigger the execution of transaction unit and immediately lose the state after the execution of the transaction unit: find the hypothesis synthesis formula 683 that meets the meaning of 741, and since 683 participates in the evidence association 684, mark 684 as newly created state.

[0277]

[0278] 53) The above step 52) the evidence association 684 is marked as newly created state, trigger the execution of transaction unit and immediately lose the state after the execution of the transaction unit: delete all the proof relations supported by 684; then, the reasoning calculation module finds the canonical synthesis formulas 740 and 741 that meet the meaning of the hypothesis synthesis formula 683, and since 668 is a positive analysis formula, takes 740 and 741 as the basis of the thinking method stored by 668, and applies the thinking method stored by 668, confirms that there are 2 reasoning result records

[0279] {742, 456, <621, 1, 1, 612>, newly created}, {743, 456, <623, 1, 1, 612>, newly created} and 2 corresponding proof relations (684, 742), (684, 743).

[0280] 54) Since the above step 53) the reasoning result 743 is marked as newly created state, trigger the execution of transaction unit and immediately lose the state after the execution of the transaction unit: mark 456 as to-be-synthesized state.

[0281] 55) Since the above step 54) the synthesis formula 456 is marked as to-be-synthesized state, trigger the execution of transaction unit and immediately lose the state after the execution of the transaction unit: delete all the proof relations supported by 456, and find all the reasoning results 730, 731, 742 and 743 marked synthesis formula as 456 of the group, execute the synthesis process on them, and confirm that there are 2 canonical synthesis formulas

[0282] {744, 639, 643, <621, 609, 610, 612>, N, new created}, {745, 639, 643, <623, 609, 610, 612>, N, new created}

[0283] and 2 corresponding proof relations (456, 744), (456, 745).

[0284] 56) Since the above step 55) norm synthesis 745 is marked as new created, trigger to execute a transaction unit and lose the state immediately after the transaction unit is executed: find a norm synthesis 679 that is consistent with the meaning of 745, since 679 participates in the looped association 680, mark 680 as new created.

[0285] 57) The above step 56) looped association 680 is marked as new created, trigger to execute a transaction unit and lose the state immediately after the transaction unit is executed: delete all proof relations supported by 680; then, the reasoning calculation module finds norm syntheses 744 and 745 that are consistent with the meaning of the hypothesis synthesis 679, since 669 is a positive analysis, take 744 and 745 as the basis of the thinking method stored by 669, and apply the thinking method stored by 669, confirm that there are 2 reasoning result records

[0286] {749, 457, <748, 621, 609, 610, 612>, new created}, {753, 457, <752, 623, 609, 610, 612>, new created} and 2 corresponding proof relations (680, 749), (680, 753).

[0287] 58) Since the above step 57) reasoning result 753 is marked as new created, trigger to execute a transaction unit and lose the state immediately after the transaction unit is executed: mark 457 as to be synthesized.

[0288] 59) Since the above step 58) synthesis 457 is marked as to be synthesized, trigger to execute a transaction unit and lose the state immediately after the transaction unit is executed: delete all proof relations supported by 457, and find all reasoning results 749 and 753 that group mark synthesis for 457, execute the synthesis process for them, and confirm that there are 2 norm syntheses

[0289] {754, 640, 643, <748, 621, 609, 610, 612>, N, new created}, {755, 640, 643, <752, 623, 609, 610,

[0290] 612>, N, new created} and 2 corresponding proof relations (457, 754), (457, 755).

[0291] 60) Since the above step 59) normatively annotated the newly created state to the specification synthesis 755, triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: finds the normative synthesis 677 that is consistent with the meaning of the hypothesis synthesis 755, since 677 participates in the evidential association 678, therefore annotates the newly created state to 678.

[0292] 61) The above step 60) evidential association 678 annotated the newly created state, triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: deletes all the proof relations supported by 678; Then, the reasoning computation module finds the normative syntheses 754 and 755 that are consistent with the meaning of the hypothesis synthesis 677, since 670 is a positive analysis, therefore takes 754 and 755 as the basis of the thinking method stored by 670, and applies the thinking method stored by 670, confirms that there are 2 reasoning result records

[0293] {758, 458, <748, 621, 612>, newly created}, {759, 458, <752, 623, 612>, newly created} and the corresponding 2 proof relations (678, 758), (678, 759).

[0294] 62) Since the above step 61) reasoning result 759 annotated the newly created state, triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: annotates the to-be-synthesized state to 458.

[0295] 63) Since the above step 62) synthesis 458 annotated the to-be-synthesized state, triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: deletes all the proof relations supported by 458, and finds all the reasoning results 758 and 759 that group mark the synthesis for 458, performs the synthesis process on them, and confirms that there are 2 normative syntheses

[0296] {760, 641, 643, <748, 621, 612>, N, newly created}, {761, 641, 643, <752, 623, 612>, N, newly created} and 2 corresponding proof relations (458, 760), (458, 761).

[0297] 64) Since the above step 63) normative synthesis 761 annotated the newly created state, triggers the execution of a transaction unit and loses the state immediately after the execution of the transaction unit: finds the hypothesis synthesis 675 that is consistent with the meaning of 761, since 675 participates in the evidential association 676, therefore annotates the newly created state to 676.

[0298] 65) Step 64) above, evidence-based association 676, marks the newly created state, triggers the execution of a transaction unit, and immediately loses its state after the execution of this transaction unit: all proof relations supported by 676 are deleted; then, the inference calculation module finds canonical synthesis formulas 760 and 761 that conform to the meaning of hypothesis synthesis formula 675. Since 671 is an affirmative analytic formula, 760 and 761 are used as the basis for applying the thinking method stored in 671, and the thinking method stored in 671 is applied to confirm the existence of one inference result record.

[0299] {764,459, <763> , Newly created} and the corresponding 1 proof relation (676,764).

[0300] 66) Since the reasoning result 764 in step 65) above marks the newly created state, the execution of the transaction unit is triggered and the state is lost immediately after the execution of the transaction unit: mark 459 as pending synthesis state.

[0301] 67) Since step 66) above marks the state to be synthesized as synthesis formula 459, an execution transaction unit is triggered, and the state is immediately lost after the execution of this transaction unit: all proof relations supported by 459 are deleted, and all inference results 764 grouped and marked as synthesis formula 459 are found. The synthesis process is performed on them, and it is confirmed that there exists one canonical synthesis formula {765, 642, 643,} <763> ,N,newly created} and 1 corresponding proof relation (459,765).

[0302] 68) Since step 67) above specifies the newly created state for synthesis formula 765, an execution transaction unit is triggered, and the state is immediately lost after the execution of this transaction unit: a hypothesis synthesis formula 660 that matches the meaning of 765 is found, i.e., the reasoning goal proposed by the user. Therefore, the knowledge base device reports the synthesis formula with substance code 765 as the reasoning result. Since the syntactic substance code of 765 is 642, the concept sequence is... <763> The equivalent syntactic form of 642 is 443, with a reference sequence of <233,234>. It can be converted into a text block sequence <"crowd", "needs to submit a new examination request to the school hospital"> using the LOCF and SYMX tables, and a concept sequence. <763>

[0303] The order of the symbols corresponds to the position of the text block "crowd". On the other hand, the sequence of symbols used to represent concept 763 is...

[0304] <"Zhang San", "Li Si">. Based on the above reference sequence and symbol sequence, the synthesized symbol sequence "Zhang San and Li Si need to submit a new examination application to the school hospital" reports to users.

[0305] The above description is not a limitation of the present application, and it should be pointed out that for those skilled in the art, without departing from the essential scope of the present application, a number of changes, modifications, additions or substitutions can be made, and these improvements and refinements should be considered as the protection scope of the present application.< / cpts> < / cpts> < / cpts> < / cpts> < / cpts> < / cpts> < / agcptrs> < / aglocrf> < / indr> < / agcptrs> < / afcptrs> < / scptrs> < / agcptrs> < / afcptrs> < / scptrs> < / agcptrs> < / aglocrf> < / afcptrs> < / aflocrf> < / scptrs> < / slocrf> < / aglocrf> < / aflocrf> < / slocrf> < / slocrf> < / aglocrf> < / aflocrf> < / scptrs> < / agcptrs> < / afcptrs> < / fcptrs> < / fcptrs> < / flocrf> < / flocrf> < / flocrf> < / fcptrs> < / gcptrs> < / fcptrs> < / fcptrs> < / gcptrs> < / gcptrs> < / gcptrs> < / fcptrs> < / glocrf> < / flocrf> < / gcptrs> < / glocrf> < / fcptrs> < / flocrf> < / gcptrs> < / fcptrs> < / glocrf> < / flocrf> < / cpts> < / cpts> < / cptrs> < / cptrs> < / locrf> < / cptrs> < / locrf> < / cpts> < / rlocf> < / rlocf> < / aglocf> < / aflocf> < / slocf> < / rlocf> < / aglocf> < / aflocf> < / slocf> < / rlocf> < / slocf> < / rlocf> < / rlocf> < / locf> < / locf> < / slocf> < / slocf> < / flocf> < / glocf> < / slocf> < / flocf> < / locf> < / glocf> < / locf> < / slocf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / locf> < / bbbf> < / bbbf>

Claims

1. An interactive mind mapping knowledge base device, characterized in that the knowledge base device is used to store knowledge; it includes a mind mapping module, which specifically includes the following types of units; 1.

1. Syntactic unit, or syntactic unit for short; a syntactic unit stores the common semantic structure of synthetic statements with the same semantic structure; a semantic structure represents a specific relationship between a set of unspecified concepts; a semantic structure has a set of positions for placing concepts; when a concept is placed in a specific position in a semantic structure, the concept is said to appear in that position; 1.

2. A comprehensive unit, or simply a comprehensive unit; a comprehensive unit stores knowledge described by a comprehensive statement with real validity; when a comprehensive unit stores knowledge described by a comprehensive statement, the comprehensive unit stores the semantic structure of the comprehensive statement, the concept at the position of the concept appearing in the semantic structure, and the context of the comprehensive statement; the comprehensive unit is said to represent the comprehensive statement. 1.

3. Syntactic equivalence unit, or syntactic equivalence for short; Syntactic equivalence stores a kind of association between two semantic structures. The meaning of the association is that there is a one-to-one correspondence between the positions where concepts are placed in the two semantic structures. When the concepts appearing in the corresponding positions are the same, the two semantic structures and the concepts appearing in the positions where the concepts are placed represent the same semantics as a whole. 1.

4. Substantial equivalence unit, or substantive equivalence for short; the knowledge base device uses codes to represent concepts; a substantial equivalence stores an association between two codes representing concepts, the meaning of which is that the two codes represent the same concept; 1.

5. Analogical equivalence unit, or analogical equivalence for short; analogical equivalence stores an association between two synthesized statements, the meaning of which is that the two synthesized statements have the same context; when all synthesized statements in the usage scenario of the knowledge base device have the same context, the knowledge base device chooses not to use analogical equivalence; at this time, it is equivalent to specifying that all synthesized statements represented by synthesized expressions in the mind map module have the same context, and the synthesized expression no longer stores the context of the synthesized statement; The knowledge base device creates various mind map module units based on sequences of symbols representing specific concepts to store the knowledge and thinking methods represented by these symbol sequences. The knowledge base device assigns different codes to different sequences, called sequence codes; it assigns a globally distinguishable code to each position in each sequence, called a position code; it uses sequence codes as identity form codes for mind map module units and position codes as concept form codes; the knowledge base device records the symbol sequences on which these units are based while creating them. The knowledge base device creates various mind map module units based on natural language text in the form of symbol sequences entered by the user, and responds to the thinking goals specified by the user in the form of natural language text in the form of symbol sequences.

2. The interactive mind mapping knowledge base device according to claim 1, characterized in that, The knowledge base device is also used to store thinking methods; The mind mapping module may also include the following types of optional units as needed; 2.

1. Analytical unit, or simply analytical unit; an analytical unit, alone or in conjunction with a sub-synthesis, stores specific instances of analytical statements that illustrate a thinking method; when an analytical unit stores a thinking method, the analytical unit stores the application scenario of the thinking method and the process description of obtaining the reasoning result by applying the thinking method. 2.

2. Sub-synthesis mapping unit, abbreviated as sub-synthesis mapping; Sub-synthesis mapping stores a relationship between a synthesis statement and an analysis statement. The meaning of the relationship is that the synthesis statement is the context required for the analysis statement to be transformed into a specific instance of the analysis statement. 2.

3. Predicate-based unit, or simply predicate-based; A predicate stores the association between a synthetic statement and its subject. When a predicate stores the association between a synthetic statement and its subject, the predicate stores the position of the concept corresponding to the subject of the synthetic statement in the semantic structure within the synthetic expression representing the synthetic statement.

3. The interactive mind mapping knowledge base device according to claim 1 or 2, characterized in that, When the knowledge base device creates various units of the mind map module based on concept sequences represented by natural language statements, the various units of the mind map module store the knowledge and thinking methods described by the natural language statements corresponding to these concept sequences in the following manner; 3.

1. The common semantic structure of synthesized statements with the same semantic structure stored in the syntactic form is implemented as follows: If there exists a set of synthesized statements that represent concept sequences of the same length, differing only in concepts at specific positions while having the same concepts at other positions, then they conform to a common semantic structure; the positions where these concept sequences differ are called variable positions; these variable positions are named "the 1st", "the 2nd", ... "the nth" variable positions according to their order of appearance in the concept sequence; this set of variable positions represents the placement of concepts in the semantic structure; at this point, the core attributes of the syntactic form storage include: A set of variable positions; for this type of unit and other mind map module units, if two mind map module units are said to have a set of variable positions and they correspond to each other, it means that the two sets of variable positions have the same number of positions, the same name, and a corresponding relationship between variable positions with the same name. 3.

2. When a synthetic expression stores the knowledge described by a synthetic statement, the synthetic expression stores the semantic structure of the synthetic statement, the concepts appearing at the positions of the concepts in the semantic structure, and the context of the synthetic statement, implemented as follows: A synthesis stores the semantic structure of the concept sequence represented by a synthesis statement, the context of the synthesis statement, and the concepts appearing in the variable positions of the semantic structure. In this case, the synthesis is said to have a set of variable positions, which correspond to the variable positions storing the syntactic form of the semantic structure. The concepts appearing in the variable positions of the semantic structure also appear in the corresponding variable positions of the synthesis. In this type of unit and other mind map module units, the concepts appearing in a set of variable positions are sorted according to their variable position names, and the resulting sequence is called a variable sequence. In this variable sequence, the position of a concept is also the name of the variable position in which the concept appears. Therefore, "the variable position in which the concept appears" and "the position of the concept in the variable sequence" have the same meaning. At this point, the core attributes of integrated storage include: A syntax, a context, a set of variable positions, and a variable sequence; the context attribute of the comprehensive expression is empty at creation time, and then the context of the comprehensive expression is inferred based on the same relationship of the comprehensive context described by the analogy equivalence; when specifically implementing this type of unit and other mind map module units, the dependent variable position attribute is inferred from the variable sequence attribute, that is, it contains a number of variable positions equal to the length of the variable sequence, and the names are "the 1st"..."the nth" variable positions in sequence, and the variable position attribute is no longer stored separately; 3.

3. The syntactic equivalence stored in the above statement represents a relationship between two semantic structures. This relationship signifies a one-to-one correspondence between the variable positions in these two semantic structures. When the concepts appearing at corresponding positions are identical, the two semantic structures and the concepts appearing at their respective positions represent the same semantics. This is achieved as follows: If there are two semantic structures, referred to as semantic structure form one and semantic structure form two, and their variable positions have a one-to-one correspondence; then, for any pair of concept sequences, if the first one conforms to semantic structure form one and the second one conforms to semantic structure form two, and the concepts appearing at corresponding variable positions are identical, then they necessarily represent the same semantics; in this case, semantic structure form one and semantic structure form two are said to be equivalent; the syntactic expressions storing semantic structure form one and semantic structure form two are respectively called equivalent syntactic expression one and equivalent syntactic expression two. At this point, the core properties of syntactic equivalence storage include: a one-to-one correspondence between equivalent syntax one, equivalent syntax two, and the variable positions between equivalent syntax one and equivalent syntax two; equivalent syntax one and equivalent syntax two are said to have a syntactic equivalence relation; syntactic equivalence relations are transitive, and all syntaxes connected by syntactic equivalence relations are said to be equivalent; for the sake of simplicity, the same or equivalent syntaxes are simply referred to as syntactic equivalence. 3.

4. The aforementioned substantial equivalence storage establishes an association between two codes representing concepts. This association signifies that the two codes represent the same concept. It is implemented as follows: the knowledge base device uses codes to represent concepts. If two codes represent the same concept, these two codes are called the formal codes of that concept, or simply formal codes. If a concept corresponds to a unique code, then the code is called the substantial code of the concept, or simply substantial code. In this case, the core attributes of substantial equivalence storage include: two formal codes representing the same concept; a substantial equivalence relationship between these two formal codes representing the concept; and transitivity of the substantial equivalence relationship, meaning that all formal codes representing concepts connected by the substantial equivalence relationship are substantially equivalent. 3.

5. The analog equivalence storage describes an association between two synthesized statements. This association signifies that the two synthesized statements share the same context. Specifically, two synthesized statements with the same context are called equivalence synthesized statement one and equivalence synthesized statement two; the syntheses representing these two synthesized statements are called equivalence synthese one and equivalence synthese two. Therefore, the core attributes of analog equivalence storage include: equivalence synthese one and equivalence synthese two; an analog equivalence relationship exists between equivalence synthese one and equivalence synthese two; the analog equivalence relationship is transitive, meaning all syntheses connected by the analog equivalence relationship share the same context. 3.

6. When an analytical expression stores a thinking method, the analytical expression stores the application scenario of the thinking method and the process description of obtaining the reasoning result by applying the thinking method, which is implemented as follows: An application scenario of the thinking method consists of two synthetic statements, wherein the first synthetic statement contains all the semantics of the second synthetic statement or contains part of the semantics of the second synthetic statement; in this case, the first synthetic statement is said to contain or partially contain the second synthetic statement, and the synthetic expression representing the first synthetic statement is said to contain or partially contain the synthetic expression representing the second synthetic statement; the analytical expression is said to connect the two synthetic expressions; the first synthetic expression is called the causal synthetic expression of the analytical expression; the second synthetic expression is called the effect synthetic expression of the analytical expression; the process of obtaining the reasoning result by applying the thinking method is divided into two steps; the first step is the screening step. The first step is to find knowledge that meets specific conditions from existing knowledge as the basis for reasoning; the second step is the generation step, which generates knowledge implied or partially implied by the knowledge obtained from the screening step or the knowledge that "existing knowledge cannot be found based on specific conditions". The analytic expression uses two sets of variable positions and the concepts above them to represent the execution methods of the screening and generation steps respectively. At this time, the analytic expression contains two parts: the screening part and the generation part. The screening part contains a set of variable positions and the concepts above them, representing the execution method of the screening step; the generation part contains a set of variable positions and the concepts above them, representing the execution method of the generation step. At this time, the core attributes stored in the analytic expression include: a causal synthesis, an effect synthesis, a set of screening part variable positions, a screening part variable sequence, a set of generation part variable positions, and a generation part variable sequence. 3.

7. The aforementioned subsynthesis-mapped storage provides an association between a synthesis statement and an analysis statement. The significance of this association is that the synthesis statement is the context required for the analysis statement to be transformed into a specific instance of the analysis statement. This is implemented as follows: based on a concept sequence represented by an analysis statement, it is transformed into a concept sequence represented by a specific instance of the analysis statement; that is, replacing a concept at a specific position in the concept sequence represented by the analysis statement with another concept in its context. When the context can be described by a synthesis statement, the concepts in the concept sequence represented by the synthesis statement are the source of the aforementioned concept replacement operation. As mentioned above, subsynthesis... In the concept sequence represented by the specific instance of the analytical statement, the concepts related to the application scenario of the thinking method and the process of obtaining the reasoning result are stored by the variable sequence of the filtering part and the variable sequence of the generating part of the analytical expression; therefore, the result of the aforementioned concept substitution operation is that the concepts appearing at the variable positions of the specific filtering part and generating part of the analytical expression will be the same as the concepts appearing at the variable positions of the specific synthetic expression; the subsynthetic mapping describes this concept identity relationship; at this time, the core attributes stored in the subsynthetic mapping include: the correspondence between the variable positions in an analytical expression, a synthetic expression, and the variable positions in the synthetic expression that have a concept identity relationship; at this time, it is said that there is a subsynthetic mapping relationship between the analytical expression and the synthetic expression. 3.

8. When a predicate stores the association between a synthetic statement and its subject, the predicate storage is implemented as follows: the position of the concept corresponding to the subject of the synthetic statement in the semantic structure of the synthetic expression representing the synthetic statement is implemented as follows: the synthetic expression representing the synthetic statement is called the frame synthetic expression of the predicate; the variable position where the concept corresponding to the subject of the synthetic statement in the frame synthetic expression is located is called the key variable position; at this time, the core attributes of the predicate storage include: a frame synthetic expression and a set of key variable positions; 3.

9. The process of creating mind graph module units by the knowledge base device has the following characteristics: the knowledge base device uses codes to represent the identity of mind graph module units, called the identity code of the mind graph module unit; the identity code of a mind graph module unit is used to refer to this unit in the attributes stored by other mind graph module units; the attributes stored by a mind graph module unit are unique, that is, the identity code of the mind graph module unit corresponds one-to-one with the attribute; similar to the code representing a concept, the identity code of a mind graph module unit is also divided into a formal code and a substantive code; when a mind graph module unit uses only the substantive code representing the identity of the mind graph module unit and the substantive code representing a concept when storing its attributes, the identity code of this mind graph module unit is a substantive code; when a mind graph module unit uses either the formal code representing the identity of the mind graph module unit or the formal code representing a concept when storing its attributes, the identity code of this mind graph module unit is a formal code; a mind graph module unit with a formal identity code corresponds to a unique mind graph module unit with a substantive identity code; multiple mind graph module units with formal identity codes correspond to the same identity. The mind map module unit with the code as the substance code; in the above many-to-one relationship, a mind map module unit with an identity code as the formal code is called a formal unit of a mind map module unit with an identity code as the substance code; conversely, a mind map module unit with an identity code as the substance code is called a substance unit of every mind map module unit with an identity code as the formal code; the context attribute of the synthesis stores a code representing a context concept; one synthesis corresponds to a specific context, and multiple synthesiss correspond to the same context. Therefore, the identity code of the synthesis, including the formal code and the substance code, are all formal codes of context concepts; therefore, the analogy equivalence also illustrates the substantial equivalence relationship between formal codes representing context concepts; based on the substantial equivalence relationship between concept codes and the equivalence relationship of syntax, firstly, a many-to-one relationship between formal units and substance units of mind map module units whose attributes only contain concepts and syntax is established. Then, based on the established many-to-one relationship between formal units and substance units, the attributes of mind map module units that use formal codes are converted to be stored using only substance codes, and the substance units of all mind map module units can be gradually established.

4. The interactive mind mapping knowledge base device according to claim 3, characterized in that the synthesis includes three types: normative synthesis, verifiable synthesis, and hypothetical synthesis; the normative synthesis stores necessary facts; the verifiable synthesis stores possibilities, representing a set of possible normative synthesiss that conform to its meaning; the hypothetical synthesis stores the thinking goal, instructing the knowledge base device to perform reasoning calculations to obtain a set of possible normative synthesiss that conform to its meaning; the normative synthesis conforms to the meaning of verifiable synthesis or hypothetical synthesis in that they have the same context, are syntactically equivalent, and have the same concept at the corresponding variable position, or the concept of verifiable synthesis or hypothetical synthesis is "to be determined"; the meaning of the corresponding variable position is: since the variable position of a syntax corresponds to the variable position of a synthesis with the syntax as an attribute, the one-to-one correspondence between the variable positions of two equivalent syntaxes simultaneously indicates the one-to-one correspondence between the variable positions of two synthesiss with the two syntaxes as attributes respectively, and the corresponding variable position refers to the variable position in the two synthesiss that has this correspondence.

5. The interactive mind mapping knowledge base device according to claim 4, characterized in that, 5.

1. The analytic formula includes two types: affirmative analytic formula and negative analytic formula; the basis for the application of the thinking method stored in the affirmative analytic formula is a set of normative synthetic formulas; the basis for the application of the thinking method stored in the negative analytic formula is a negative hypothesis synthetic formula, which refers to a hypothesis synthetic formula for which no known normative synthetic formula conforms to its meaning; the variable positions of the screening part of the analytic formula correspond to the variable positions of the causal synthetic formulas it connects to, and the variable positions of the generating part correspond to the variable positions of the effect synthetic formulas it connects to; the reasoning result of the analytic formula has a set of variable positions that correspond to the variable positions of the generating part of the analytic formula; The generation part of the analytic expression contains a set of variable positions and the concepts thereon, used to represent the execution method of the generation step. It is implemented such that a concept at a variable position in the generation part represents the reasoning method for obtaining the corresponding concept at the variable position in the reasoning result. The reasoning method includes: known, i.e., a specific concept is fixed; referencing, i.e., the concept is the same as the concept at the specific variable position in the canonical synthesis or hypothetical synthesis upon which the analytic expression is based; and set, i.e., clustering the set of canonical synthesis upon which the analytic expression is based according to the concepts at the specific variable positions, and in each category, synthesizing the concepts at the specific non-clustered variable positions into a set, while simultaneously determining... The existence of a concept to refer to the set is acknowledged; decomposition, i.e., knowing that the concept at a specific variable position in the canonical synthesis or hypothetical synthesis upon which the analytic expression is based refers to a set, decomposes the set into its element concepts; reflection, i.e., confirming the existence of a concept to represent the canonical synthesis itself ultimately derived by reasoning; undetermined, i.e., fixing the "undetermined" concept, indicating that the thought method stored in this analytic expression cannot obtain the concept at the variable position; at the same time, the analytic expression indicates the relationship between a specific generator variable position and a filter variable position, called the source relationship of the generator variable position; if the generator is a representation at the i-th variable position... The concept of "set", "decomposition" or "substitution" method is associated with the j-th variable position in the filtering part. The source relationship indicates that the basis for obtaining the concept at the i-th variable position in the reasoning result is: the concept at the variable position corresponding to the j-th variable position in the filtering part in the synthetic formula on which the thinking method is applied this time; for simplification, the concept that does not represent the reasoning method is directly used in the generation part to represent the "known" reasoning method, that is, the concept is fixed, and the concept of "to be determined" is used to represent the "to be determined" reasoning method; the variable positions where the concepts representing the "known", "reflection" and "to be determined" reasoning methods are located in the generation part have no source relationship; 5.

2. The predicate form is also used to associate the task execution system or data query system connected to the knowledge base device, implemented as follows: when a hypothesis synthesis conforming to the predicate form is created during the reasoning process, the knowledge base device operates the task execution system or the data query system to execute the task and confirm that the canonical synthesis of the stored task execution result exists in the mind graph module, or to query the data and confirm that the canonical synthesis of the stored query result exists in the mind graph module; the hypothesis synthesis conforming to the predicate form means that the following conditions are met simultaneously: 1) the context of the hypothesis synthesis and the frame synthesis of the predicate form are the same and the syntax is equivalent; 2) the concept at the key variable position in the frame synthesis is the same as the concept at the corresponding variable position in the hypothesis synthesis or the concept in the frame synthesis is "pending"; 3) the concept at the variable position corresponding to the key variable position in the frame synthesis is not "pending"; at this time, the concept at the variable position corresponding to the key variable position in the frame synthesis constitutes the parameters for task execution or the conditions for data query; 5.

3. In the above-mentioned subsynthesis correspondence relationship, if the synthesis formula is a provable synthesis formula, it means that there is a subsynthesis correspondence relationship between the analytical formula and each normative synthesis formula represented by the provable synthesis formula.

6. The interactive mind graph knowledge base device according to claim 5, characterized in that the knowledge base device can also be used for reasoning calculation; the knowledge base device further includes a reasoning calculation module; the reasoning calculation module performs reasoning calculation based on the units of the mind graph module, and creates or deletes mind graph module units in the following manner to record the process or result of reasoning calculation; 6.

1. The mind mapping module also includes a type of unit used in the reasoning process, called the reasoning result unit, or simply the reasoning result; the second step of obtaining the reasoning result using the analytical storage method is the generation step, which generates the knowledge implied or partially implied by the knowledge obtained from the screening step or the knowledge that "cannot find existing knowledge based on specific conditions". This is implemented as follows: the reasoning calculation module confirms that the reasoning result storing the generated knowledge already exists in the mind mapping module; otherwise, it creates the reasoning result storing the generated knowledge. For simplicity, "confirms existence..." is used to represent "confirms existence... otherwise, creates..."; the reasoning result includes a set of variable positions and a variable sequence composed of the concepts on these variable positions ordered by variable position names; the reasoning result obtained based on an analytical expression has a set of variable positions that correspond to the result synthesis expression connected to the analytical expression. The reasoning calculation module performs a synthesis process on the reasoning results to obtain a canonical synthesis. The syntax and context of the canonical synthesis obtained by the synthesis process are the same as the result synthesis connected by the analytic expression, and its variable positions correspond to the variable positions of the result synthesis. Its variable sequence is obtained as follows: the reasoning results of the analytic expressions connected with the same result synthesis are grouped together; in each group of reasoning results, all combinations are found such that in each combination, all reasoning results have one and only one non-"pending" concept in the corresponding variable position. These non-"pending" concepts are sorted by variable position name to obtain a variable sequence, which is the variable sequence of a canonical synthesis obtained by the synthesis process. At this time, the core attributes stored in the reasoning results include: a grouping label, a group of variable positions, and a variable sequence; the grouping label of the reasoning results obtained from an analytic expression is the result synthesis connected by the analytic expression. 6.

2. The reasoning calculation module's reasoning calculation process consists of transaction units; the execution of a transaction unit is triggered by a mind graph module unit marked with a state; the states include newly created, newly deleted, and pending synthesis; the transaction units are divided into generating transaction units and deleting transaction units; generating transaction units are triggered by mind graph module units marked with a newly created or pending synthesis state; when a specific step of generating a transaction unit requires deleting a mind graph module unit, it is not directly deleted, but only marked with a newly deleted state; in all transaction units, all "reasoning calculation module finds...mind graph module unit" operation steps ignore mind graph module units marked with a newly deleted state; when a specific step of a transaction unit confirms the existence of a mind graph module unit marked with a newly deleted state, its newly deleted state is canceled; when a specific step of a transaction unit creates a new mind graph module unit, it is marked with a newly created state; deleting transaction units are triggered by mind graph module units marked with a newly deleted state, and the unit is actually deleted in the transaction unit; the mind graph module units marked with a state... A block unit loses its state immediately after triggering the execution of the corresponding transaction unit. For mind graph module units that do not exist in the knowledge base device at the start of reasoning, the reasoning computation module records the process by which they are confirmed to exist by the generating transaction unit. The execution of the generating transaction unit is triggered by a mind graph module unit that has been marked with a state, called the triggering unit, and confirms the existence of multiple mind graph module units, each of which is called a confirmation unit. The reasoning computation module records the process by which a generating transaction unit confirms the existence of a set of mind graph module units as a set of proof relations. Proof relations are a type of mind graph module unit that the reasoning computation module needs to use when performing reasoning computation. Its core attributes include a triggering unit and a confirmation unit. A proof relation is said to be supported by its triggering unit and owned by its confirmation unit. For all mind graph module units that do not exist at the start of reasoning, if they lose all the proof relations they own after a transaction unit is completed, they are marked as newly deleted. When the reasoning computation module actually deletes a mind graph module unit, it also deletes all the proof relations supported by the unit. The proof relation itself is not affected by the transaction unit. It is deleted because it does not have a proof relationship, nor is it marked as a state; it is deleted immediately when deletion is required. 6.

3. When a hypothesis synthesis is marked as newly created, the execution of the generation transaction unit is triggered: Let the hypothesis synthesis be H; First, delete all proof relations supported by H; Then, the reasoning calculation module finds all analytical expressions that confirm the existence of reasoning results that conform to the meaning of H; For each of them, denoted as Ah, the reasoning calculation module, based on the reasoning method represented by the generation part of Ah, reverse-engineers a screening condition, denoted as Fa, so that the canonical synthesis obtained by screening by Fa, after applying the thinking method of Ah, obtains a reasoning result that conforms to the meaning of H; If Fa conforms to the meaning of the screening part of Ah, then it is confirmed that there is a hypothesis synthesis, denoted as Ha, whose context and syntax are the same as the factor synthesis of Ah, and its variable sequence list Let Fa be an example, and simultaneously create a proof relation (H, Ha), confirming the existence of an association between Ha and Ah, called an evidence-based association. An evidence-based association is a type of mind map module unit used by the reasoning module during reasoning computation. Its core attributes include: a hypothesis synthesis and an analytic expression. Let Sah describe the evidence-based association between Ha and Ah, and create a proof relation (H, Sah). Then, the reasoning module finds all the syntaxy correspondences involving Ah, and for each of them, let Ma be an example. Let Ma describe the syntaxy correspondence between the analytic expression Ah and the synthesis expression Ta. If the meaning of Ta does not contradict Fa, then a hypothesis synthesis, denoted as Ht, is confirmed, whose context and syntax are consistent with Fa. Ta is the same, and its variable sequence represents the meaning of Ta after being restricted by Fa. At the same time, a proof relation (H, Ht) is created, and it is confirmed that there is an association between Ma, Ht and Ah. This association is called a deformed association. The deformed association is also a type of mind map module unit used by the reasoning and calculation module when performing reasoning and calculation. Its core attributes include: a subsynthetic enumeration, a hypothesis synthesis, and an analytical expression. Let the deformed association Cmha describe the deformed association between Ma, Ht and Ah, and create a proof relation (H, Cmha). Then, the reasoning and calculation module finds all predicates of H that conform to its meaning. For each of them, it is denoted as Ph. If Ph is associated with the task execution system or data query system connected to the reasoning and calculation module, then the reasoning and calculation module will determine the relationship between Ma, Ht and Ah. If the system is in place, the reasoning and calculation module executes the task or queries the data, and confirms the existence of a set of normalized synthesis formulas for storing the task results or required data. For each of these formulas, denoted as Nh, a proof relation (H, Nh) is created. The analytical formula confirms the existence of a reasoning result that conforms to the meaning of the hypothetical synthesis formula. One method of judgment is that the result synthesis formula connected by the analytical formula is syntactically equivalent to the hypothetical synthesis formula and has the same context. Furthermore, the variable sequence of the generator part of the analytical formula and the variable sequence of the hypothetical synthesis formula satisfy the following conditions at the corresponding variable positions: 1) the concept of the hypothetical synthesis formula is "to be determined", or 2) the concept of the hypothetical synthesis formula is not "to be determined" and the concept of the generator part represents a non-"known" reasoning method, or 3) the concept of the hypothetical synthesis formula is not "to be determined" and is the same as the concept of the generator part.Based on the analytical formula's backward deduction of screening conditions, the filtered canonical synthesis, after applying the analytical formula's stored thinking methods, yields a reasoning result that conforms to the meaning of the hypothetical synthesis. One implementation is to denote the screening conditions as a sequence of variables, with each variable corresponding to a variable in the analytical formula's screening part. In this case, since the variable in the analytical formula's generating part is associated with the variable in the screening part in the source relation, the variable in the analytical formula's generating part is also simultaneously associated with the variable in the screening conditions. The concept at each variable in the screening conditions is determined as follows: the variable in the hypothetical synthesis corresponds to the variable in the generating part; for each pair of corresponding variable positions, if the concept of the hypothetical synthesis is not "to be determined" and the concept of the generating part... The "referential" reasoning method states that the concept at the screening condition variable position associated with the generator variable position is equal to the concept of the hypothesis synthesis. For screening condition variable positions whose concepts cannot be determined according to the method, the concept at that position is equal to the non-"to be determined" concept at the corresponding screening variable position. If the concept at the corresponding screening variable position is "to be determined," then the concept at that position is "arbitrary." One method for determining whether the screening conditions conform to the meaning of the analytic screening part is that the concepts at corresponding variable positions of the screening condition variable sequence and the analytic screening part variable sequence are the same or the concept of the screening part is "arbitrary." One implementation of the hypothesis synthesis variable sequence representing the screening conditions is that the hypothesis synthesis variable... The quantity sequence is obtained by replacing "any" with "to be determined" in the screening condition variable sequence; one method to determine whether the meaning of the comprehensive formula contradicts the screening conditions is that the scoring comprehensive correspondence contains the concept sameness relationship between the i-th variable position of the analytical formula screening part and the j-th variable position of the comprehensive formula. For each pair i and j, the following conditions are met: 1) the concept at the j-th variable position of the comprehensive formula is not "to be determined", and the concept at the i-th variable position of the screening condition is the same as it or is "any", or 2) the concept at the j-th variable position of the comprehensive formula is "to be determined"; the variable sequence of the hypothetical comprehensive formula represents the meaning obtained by restricting the comprehensive formula with the analytical formula having a comprehensive correspondence relationship with the analytical formula through screening conditions. One implementation is... The variable sequence of the hypothesis synthesis is obtained as follows: 1) For each group of i and j mentioned above, if the concept in the j-th variable position of the synthesis is "to be determined" and the concept in the i-th variable position of the screening condition is "any", then the concept in the j-th variable position of the hypothesis synthesis is "to be determined". 2) At the same time, for each group of i and j mentioned above, if the concept in the j-th variable position of the corresponding synthesis is "to be determined" and the concept in the i-th variable position of the screening condition is not "any", then the concept in the j-th variable position of the hypothesis synthesis is the same as the concept in the i-th variable position of the screening condition. 3) The concept in the variable position of the hypothesis synthesis that cannot be determined according to the above method is equal to the concept in the corresponding variable position of the corresponding synthesis. 6.

4. When an analytic expression is marked as newly created or newly deleted, the corresponding transaction unit is triggered. In both the creation and deletion transaction units, the following operations are performed: Let the analytic expression be A; the reasoning and calculation module finds the hypothesis synthesis expressions that confirm the existence of reasoning results that conform to the meaning of analytic expression A, and marks these hypothesis synthesis expressions as newly created. 6.

5. When an evidence-based association marks a newly created state, the execution of a transaction generation unit is triggered: Let the evidence-based association be S, which describes the evidence-based association between the hypothesis synthesis H and the analytic A; First, delete all proof relations supported by S; Then, the reasoning computation module finds the canonical synthesis that conforms to the meaning of H, and denotes its set as E; If A is an affirmative analytic, then E is used as the basis for applying the thought methods stored in A; If E is an empty set and A is a negative analytic, then H is used as the basis for applying the thought methods stored in A; Based on the above basis, the reasoning computation module applies the thought methods stored in A to confirm the existence of each reasoning result; For each reasoning result, denoted as Rs, a proof relation (S, Rs) is created; 6.

6. When a variant association marks a newly created state, the execution of the generation transaction unit is triggered: Let the variant association be C, where C describes the variant association of the syntactic mapping M, the hypothetical synthesis H, and the analytical expression A; First, delete all proof relations supported by C; Then, find the canonical synthesis that conforms to the meaning of H, and for each of them, let Nt. The reasoning and calculation module confirms the existence of an analytical expression whose causal synthesis and effect synthesis are the same as A, and whose filtering part and generating part variable sequence are the variable sequence obtained after the filtering part and generating part variable sequence of A are partially replaced by the concept in Nt according to the same relation described by M. Let the analytical expression be At, and create the proof relation (C, At); 6.

7. When a canonical synthesis is marked as newly created or newly deleted, the corresponding transaction unit is triggered. In both the creation and deletion transaction units, the following operations are performed: Let the canonical synthesis be N; the inference calculation module finds the hypothetical synthesis that N is consistent with its meaning, and for each of them, it is denoted as Hn; the inference calculation module finds the evidence-based associations in which Hn participates and marks them as newly created; the inference calculation module finds the variant associations in which Hn participates and marks them as newly created. 6.

8. When a reasoning result is marked as newly created or newly deleted, the corresponding transaction unit is triggered. In both the creation and deletion transaction units, the following operations are performed: Let the reasoning result be R, and its grouping label be the synthesis formula G; mark G as pending synthesis. 6.

9. When a synthesis formula is marked as a state to be synthesized, the execution of the transaction generation unit is triggered: Let the synthesis formula be N; First, delete the proof relations supported by N; Then, the inference calculation module finds all inference results grouped and marked as N, and performs the aforementioned synthesis process on them; The inference calculation module confirms the existence of a canonical synthesis formula obtained from the synthesis process, and for each of them, it is denoted as Nc, and a proof relation (N, Nc) is created.

7. The interactive mind mapping knowledge base device according to claim 6, characterized in that, 7.

1. The knowledge base device, while creating mind map module units, records the symbol sequences upon which these units are based, and implements this as follows: The knowledge base device also includes an auxiliary module to store natural language text in the form of symbol sequences entered by the user; a text block representing a specific concept in the natural language text is a symbol; a natural language sentence is a sequence of symbols, and the symbol sequence corresponds to a concept sequence obtained by arranging the concepts represented by each symbol in the order of their symbols; the natural language text is organized in articles, and an article includes a sequence of original symbols, referred to as the original sequence, and a set of explanatory symbol sequences, referred to as the explanatory sequence; the original sequence refers to the text used in natural language communication; the explanatory sequence refers to the text representation of semantic structures, knowledge, and thinking methods that need to be known in advance to understand the meaning expressed by the text; the explanatory sequence includes syntactic sequences, synthetic sequences, and analytical sequences, which explain... The semantic structure, knowledge, and thinking methods are stored in the syntactic, synthetic, and analytic formats of the mind graph module, respectively. Simultaneously, specific subsequences of the original text sequence also constitute statements explaining the semantic structure, knowledge, and thinking methods; the content explained by these original text sequence subsequences is also stored in the syntactic, synthetic, and analytic formats. As mentioned earlier, the knowledge base device assigns a sequence code to each sequence and a position code to each position in each sequence. The knowledge base device creates mind graph module units based on these codes. Furthermore, in its auxiliary modules, the knowledge base device uses a type of symbol unit to store text block symbols, a type of position unit to store the association between a position code and the symbol appearing at its represented position, and a type of article unit to record the arrangement of position codes corresponding to the original text sequence of the article. Additionally, the knowledge base device supplements its mind graph module units with the arrangement of position codes corresponding to the symbol sequences on which the aforementioned units are based. 7.

2. The knowledge base device responds to the user's thinking objectives with natural language text in the form of symbol sequences, implemented as follows: When reporting a synthesis, the knowledge base device refers to a symbol sequence, called the reference sequence of the synthesis; a reference sequence refers to a symbol sequence in which any variable position of the synthesis can establish a one-to-one correspondence with its variable position, such symbol sequences include syntactic sequences or original text sequence subsequences corresponding to all equivalent syntactic forms of the synthesis; the knowledge base device keeps the symbols in the reference sequence that do not correspond to the variable positions of the synthesis unchanged, and replaces the symbols that correspond to the variable positions of the synthesis with any symbol that can represent the concept at the variable position.

8. The interactive mind mapping knowledge base device according to claim 6 or 7, characterized in that, The knowledge base device uses the following types of codes: A formal code for a type of mind graph module unit indicates that a mind graph module contains a group of such units, the attributes of which use formal codes, and therefore the identity code of this group of units is also a formal code; A substantive code for a type of mind graph module unit indicates that a mind graph module contains a group of such units, the attributes of which only use substantive codes, and therefore the identity code of this group of units is also a substantive code; The knowledge base device uses the following types of codes and corresponding types of units. locf: Global positional code / conceptual formal code; a symbol at a specific position in a sequence of symbols that can be accurately understood must refer to a unique concept if it refers to a concept; symbols at different positions in the sequence refer to the same concept; therefore, the global positional code is also the formal code of the concept. stxf: The code / syntactic form code of a syntactic sequence or a subsequence of a plaintext sequence describing a semantic structure; a syntactic sequence or a subsequence of a plaintext sequence describing a semantic structure corresponds to a unique syntactic form; different syntactic sequences or subsequences of plaintext sequences describing semantic structures correspond to the same syntactic form; Therefore, the code of a syntactic sequence or a subsequence of a plaintext sequence that describes a semantic structure is also the formal code of the syntactic form; synf: The code / synthetic formal code / grouping mark essence code of the synthesized sequence or the original sequence subsequence describing knowledge; A synthetic sequence or a subsequence of a textual sequence explaining knowledge corresponds to a unique synthetic formula; different synthetic sequences or subsequences of a textual sequence explaining knowledge correspond to the same synthetic formula. Therefore, the code of a synthetic sequence or a subsequence of a textual sequence that describes knowledge is also a synthetic formal code; anaf: The code / analytic formula for an analytic sequence or a subsequence of a primitive sequence illustrating a method of thinking; each analytic sequence or subsequence of a primitive sequence illustrating a method of thinking corresponds to a unique analytic formula; different analytic sequences or subsequences of primitive sequences illustrating a method of thinking correspond to the same analytic formula; Therefore, the code of the original sequence or subsequence of the analytical sequence or the explanation of the thinking method is also the formal code of the analytical form; prdf: the formal code of the predicate; stxef: The formal code for syntactic equivalence; synef: A formal code analogous to equivalence; cptef: Formal code of substantial equivalence; asef: A format code for a composite mapping; stxs: The essential code of the syntax; ctxs: The underlying code of the context; cpts: The essential code of the concept; syns: Synthetic subcode; anas: analytic essence code; prds: the essential code of the predicate; ases: Subsidiary mapping of the core code; seks: the essential code for evidence-based association; tfms: The essential code for transformation association; ress: The essential code of the reasoning result.

9. The interactive mind mapping knowledge base device according to claim 8, characterized in that, The workflow of the knowledge base device is as follows: 9.

1. Acquire and store knowledge, thinking methods, and thinking goals; the mind mapping module of the knowledge base device receives natural language text in the form of symbol sequences entered by the user, assigns a sequence code and a position code to each symbol sequence and each position therein, uses the sequence code as the formal code of the mind mapping module unit, and uses the position code as the formal code of the concept; creates mind mapping module units based on these formal codes, the mind mapping module units being formal units, and their identity codes being formal codes; 9.

2. Start reasoning calculation; The reasoning calculation module of the knowledge base device starts reasoning calculation according to the user's instructions. First, it completes the preparatory work for reasoning calculation, that is, it obtains the substantive units of all mind map module units, as follows; First, based on the substantial equivalence formula, create the substantial codes of the concepts represented by the formal codes of all concepts that have substantial equivalence relations, and record the formal code-substantial code correspondence of the codes representing concepts; the formal code-substantial code correspondence of the codes representing concepts is called the concept correspondence relationship; Then, based on analogical equivalence, a common context substance code is created for the formal codes of all synthetic codes with analogical equivalence, recording the correspondence between synthetic formal codes and context substance codes; the correspondence between synthetic formal codes and context substance codes is called synthetic context correspondence. Then, based on syntactic equivalence, the order of variable bits in syntactic expressions with syntactic equivalence relations is unified, and the comparison relationship between the variable bit order before and after unification is recorded. A syntactic substance code is created for the formal code of all syntactic expressions with syntactic equivalence relations. The substance code is used to represent the unified variable bit order, and syntactic substance units are created. The correspondence between the formal code and substance code of the syntactic expression, and the correspondence between the variable bits of the syntactic formal units and substance units are recorded. The correspondence between the formal code and substance code of the syntactic expression, and the correspondence between the variable bits of the syntactic formal units and substance units are called syntactic comparison relations. Then, based on the aforementioned concept comparison relationship, syntactic comparison relationship, and comprehensive context comparison relationship, the attributes represented by the comprehensive formal code are converted into attributes represented by the substance code. A comprehensive substance code is assigned to each different attribute, and a comprehensive substance unit is created. The comprehensive formal code-substance code comparison relationship is recorded. The comprehensive formal code-substance code comparison relationship is simply referred to as the comprehensive comparison relationship. Then, based on the aforementioned conceptual correspondence, synthetic correspondence, and syntactic correspondence, the attributes represented by the analytic expression in formal codes are converted into attributes represented by substance codes. An analytic substance code is assigned to each different attribute, and an analytic substance unit is created. The analytic formal code-substance code correspondence is recorded. The analytic formal code-substance code correspondence is simply referred to as the analytic correspondence. Then, based on the aforementioned comprehensive, syntactic, and analytical correspondences, the attributes represented by formal codes in the syntactic correspondences are converted into attributes represented by substantive codes. A subtactic correspondence substantive code is assigned to each different attribute, and a subtactic correspondence substantive unit is created. The formal code-substantive code correspondence of the syntactic correspondences is recorded. Then, based on the aforementioned comprehensive and syntactic correspondences, the attributes represented by the predicate in formal codes are converted into attributes represented by substance codes. A predicate substance code is assigned to each different attribute, and a predicate substance unit is created. The formal code-substance code correspondence of the predicate is recorded. 9.

3. The reasoning and calculation module performs reasoning and calculation according to the substantive units in the mind map module, following the steps described above. At the start of reasoning, the substantive units corresponding to the user-input hypothesis synthesis form units are marked as newly created. During the reasoning process, when a normative synthesis that conforms to the meaning of the user-input hypothesis synthesis is created, found, or deleted, the reasoning and calculation module immediately reports to the user. Following the steps described above, the reasoning and calculation module finds available reference sequences and symbols representing specific concepts based on the correspondence between the form units and substantive units in the mind map module, and constructs a natural language text report in the form of a symbol sequence to the user.

Images