Block chain-oriented credible coal mine safety knowledge graph collaborative construction method and system

Abstract

The invention discloses a block chain-oriented credible coal mine safety knowledge graph collaborative construction method and system. According to the related scheme, 1) the knowledge quality problem and the safety problem faced by multiple knowledge submitters (contributors) during construction of a knowledge graph in a coal mine safety knowledge base construction process can be solved; (2) the adopted knowledge quality evaluation scheme can estimate the credibility of knowledge by utilizing the consistency between the knowledge, so that the influence of wrong knowledge uploaded in the collaborative construction process on the graph quality is guaranteed; 3) the method can be adapted to construction of various coal mine safety field knowledge maps, does not depend on information of a specific field, is a collaborative construction framework of the knowledge maps, and can provide a low-cost collaborative construction implementation method for complex and large-scale coal mine knowledge integration and knowledge base generation; and 4) the coal mine safety knowledge can be searched based on the coal mine safety knowledge graph, and the efficiency and accuracy of searching, updating and applying the knowledge are improved.

Description

Technical Field

[0001] The present invention relates to the field of coal mine safety technology, and in particular to a blockchain-oriented trusted coal mine safety knowledge graph collaborative construction method and system. Background Art

[0002] Coal mine safety is the top priority of the coal industry because of the many safety hazards in the coal mine operating environment. In order to reduce safety hazards, it is necessary to integrate and master a wide range of coal mine safety knowledge, including work procedures, operating skills, use of safety equipment, and emergency plans. However, this knowledge is often scattered in various information sources such as texts, files, and databases, lacking a unified and structured knowledge base, which makes it difficult to find, update, and apply knowledge. To solve this problem, it is particularly necessary to build a coal mine safety knowledge graph. The knowledge graph can integrate the knowledge from various information sources and connect them in a graphical way to form a large, structured knowledge base, thereby improving the efficiency and accuracy of knowledge search, update, and application.

[0003] Therefore, constructing a coal mine safety knowledge graph is of great significance for improving coal mine safety and preventing accidents. However, since the manpower and time costs of knowledge acquisition are usually very high, it is difficult for individuals or small teams to build a knowledge graph that meets application requirements when faced with the task of building a large coal mine safety knowledge graph. Therefore, allowing multiple contributors to build a shareable knowledge graph becomes an effective way to reduce the construction and application costs of the graph. However, in the face of a coal mine safety knowledge graph construction system with multiple contributors, how to ensure the security of the knowledge graph and how to avoid the impact of erroneous knowledge and conflicting knowledge submitted by different knowledge submitters on the quality of the coal mine knowledge graph are urgent issues to be solved. Therefore, constructing a coal mine safety knowledge graph storage and construction method for multi-contributor collaborative construction is of great significance to coal mine safety information management.

[0004] In the scheme disclosed in the Chinese invention patent "Method and system for constructing knowledge graph in the field of coal mine safety production" with the authorization announcement number CN116821376B, the knowledge graph ontology structure is constructed according to the characteristics of documents in the field of coal mine safety production, where the ontology structure includes the design of the knowledge graph text ontology structure and the hierarchical design of the knowledge graph in the field of coal mine safety production. The method of automatically obtaining the entities, entity attributes and the affiliation between entities in the documents in the field of coal mine safety production is designed to improve the efficiency and accuracy of constructing the knowledge graph.

[0005] In the solution disclosed in the Chinese invention patent "A Method and System for Constructing a Knowledge Graph of Multiple Coal Mine Disasters" with the authorization announcement number CN116521944B, by constructing a domain object model, a graph ontology model, and a graph instance library, the disaster knowledge graph designs an abstract graph model as a directed graph, and stores, retrieves, and maintains the knowledge graph in the form of a graph database.

[0006] However, the above traditional solutions for constructing a knowledge graph in the coal mine safety field mainly solve problems such as how to structurally design the constructed knowledge graph and how to extract coal mine knowledge. It ignores the quality risks and safety management risks faced in the knowledge integration and knowledge collection processes in the construction of the coal mine safety knowledge graph.

[0007] In view of this, the present invention is specifically proposed. Summary of the Invention

[0008] The purpose of the present invention is to provide a collaborative construction method and system for a trusted coal mine safety knowledge graph oriented to blockchain, aiming at the collaborative construction and secure storage of the coal mine safety knowledge graph, using the coal mine safety knowledge contributed by different knowledge submitters to form a coal mine safety knowledge graph and store it in the blockchain distributed ledger.

[0009] The purpose of the present invention is achieved through the following technical solutions:

[0010] A collaborative construction method for a trusted coal mine safety knowledge graph oriented to blockchain, including:

[0011] Collect the knowledge of each fact in the coal mine safety field submitted by knowledge submitters;

[0012] Assign a unique ID to the knowledge of each fact, organize the corresponding knowledge according to the fact, and store it in the blockchain in the form of an ID; and store the knowledge corresponding to each fact submitted by each knowledge submitter in the blockchain in the form of an ID, and each piece of knowledge corresponding to each stored fact contains a corresponding quality score; wherein, for the newly submitted knowledge of each fact, perform a validity check to determine whether the corresponding fact is valid; for each piece of knowledge of each valid fact, use the credibility of the corresponding knowledge submitter and the consistency with other knowledge under the same valid fact to perform a quality evaluation to obtain a quality score; the credibility of the knowledge submitter is calculated using the quality scores of the knowledge corresponding to each fact submitted by it.

[0013] For each valid fact, select the triple with the highest quality score, and comprehensively select the triples with the highest quality scores of all valid facts to construct a coal mine safety knowledge graph.

[0014] A collaborative construction system for a trusted coal mine safety knowledge graph oriented to blockchain, used to implement the aforementioned method, the system includes:

[0015] A knowledge collection unit for collecting knowledge of various facts in the field of coal mine safety submitted by knowledge submitters;

[0016] A knowledge storage and quality management unit for assigning a unique ID to the knowledge of each fact, organizing the corresponding knowledge according to the facts, and storing it in the blockchain in the form of an ID; and storing the knowledge corresponding to each fact submitted by each knowledge submitter in the blockchain in the form of an ID, and each piece of knowledge corresponding to the stored fact contains a corresponding quality score; wherein, for the newly submitted knowledge of each fact, an effectiveness check is performed to determine whether the corresponding fact is valid; for each piece of knowledge of each valid fact, the credibility of the corresponding knowledge submitter and the consistency with other knowledge under the same valid fact are used for quality evaluation to obtain a quality score; the credibility of the knowledge submitter is calculated using the quality scores of the knowledge corresponding to the facts submitted by it;

[0017] A coal mine safety knowledge graph construction unit for selecting the triple with the highest quality score for each valid fact and constructing a coal mine safety knowledge graph by integrating the triples with the highest quality scores of all valid facts.

[0018] A processing device includes: one or more processors; a memory for storing one or more programs;

[0019] Wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the foregoing method.

[0020] A readable storage medium stores a computer program, and when the computer program is executed by a processor, the foregoing method is implemented.

[0021] As can be seen from the technical solution provided by the present invention above, 1) it can solve the knowledge quality problems and security problems faced by multiple knowledge submitters (contributors) in the process of constructing a coal mine safety knowledge graph; 2) the adopted knowledge quality evaluation scheme can utilize the consistency between knowledge to estimate the credibility of knowledge and ensure the impact of uploaded incorrect knowledge on the graph quality during the collaborative construction process; 3) it can be adapted to the construction of various coal mine safety field knowledge graphs, does not depend on information in a specific field, is a collaborative construction framework for knowledge graphs, and can provide a low-cost collaborative construction implementation method for complex and large-scale coal mine knowledge integration and knowledge base generation; 4) based on the coal mine safety knowledge graph, coal mine safety knowledge can be searched, and the efficiency and accuracy of knowledge search, update, and application can be improved. Description of the Drawings

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0023] Figure 1 It is a flowchart of a method for collaborative construction of a trusted coal mine safety knowledge graph for the blockchain provided by the embodiments of the present invention;

[0024] Figure 2 It is a schematic diagram of a method for collaborative construction of a trusted coal mine safety knowledge graph for the blockchain provided by the embodiments of the present invention;

[0025] Figure 3 It is a schematic diagram of a system for collaborative construction of a trusted coal mine safety knowledge graph for the blockchain provided by the embodiments of the present invention;

[0026] Figure 4 It is a schematic diagram of a processing device provided by the embodiments of the present invention. Detailed implementation manners

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0028] First, the following explanations will be made for the terms that may be used in this article:

[0029] The term "and / or" means that either one or both of the two can be realized. For example, X and / or Y means that it includes both the case of "X" or "Y" and the three cases of "X and Y".

[0030] The description of terms such as "include", "comprise", "contain", "have" or other similar semantics should be interpreted as non-exclusive inclusion. For example: including a certain technical feature element (such as raw material, component, ingredient, carrier, dosage form, material, size, part, component, mechanism, device, step, process, method, reaction condition, processing condition, parameter, algorithm, signal, data, product or article, etc.) should be interpreted as not only including the clearly listed certain technical feature element, but also including other well-known technical feature elements in the art that are not clearly listed.

[0031] The term "consisting of" means excluding any technical feature elements not expressly listed. If this term is used in a claim, the claim will be closed, excluding technical feature elements other than those expressly listed, except for conventional impurities associated therewith. If this term only appears in a sub-clause of a claim, it only limits the elements expressly listed in that sub-clause, and the elements recited in other sub-clauses are not excluded from the overall claim.

[0032] The following provides a detailed description of the method and system for collaborative construction of a trusted coal mine safety knowledge graph for blockchain. The content not described in detail in the embodiments of the present invention belongs to the prior art well-known to those skilled in the art. For the conditions not specified in the embodiments of the present invention, they are carried out according to the conventional conditions in the art or the conditions recommended by the manufacturer. The reagents or instruments used in the embodiments of the present invention without indicating the manufacturer are all conventional products that can be obtained through commercial purchase.

[0033] Embodiment 1

[0034] The embodiment of the present invention provides a method for collaborative construction of a trusted coal mine safety knowledge graph for blockchain, as Figure 1 shown, which mainly includes the following steps:

[0035] Step 1: Collect the knowledge of each fact in the coal mine safety field submitted by the knowledge submitter.

[0036] In the embodiment of the present invention, for the coal mine safety field, collect the knowledge of each fact submitted by the knowledge submitter in the form of triples.

[0037] Step 2: Knowledge storage and quality management.

[0038] In the embodiment of the present invention, assign a unique ID to the knowledge of each fact, organize the corresponding knowledge according to the fact, and store it in the blockchain in the form of the ID; and store the knowledge corresponding to each fact submitted by each knowledge submitter in the blockchain in the form of the ID, and the knowledge corresponding to each stored fact includes the corresponding quality score; wherein, for the newly submitted knowledge of each fact, perform a validity check to determine whether the corresponding fact is valid; for each knowledge of each valid fact, use the credibility of the corresponding knowledge submitter and the consistency with other knowledge under the same valid fact to perform a quality evaluation to obtain a quality score; the credibility of the knowledge submitter is calculated using the quality scores of the knowledge corresponding to each fact submitted by it.

[0039] Step 3: Construct a coal mine safety knowledge graph.

[0040] In the embodiments of the present invention, for each valid fact, the triple with the highest quality score is selected, and the triples with the highest quality scores of all valid facts are integrated to construct a coal mine safety knowledge graph.

[0041] In order to more clearly demonstrate the technical solutions provided by the present invention and the resulting technical effects, the following uses specific embodiments to describe in detail what is provided by the embodiments of the present invention.

[0042] I. Overall introduction of the solution.

[0043] The embodiments of the present invention provide a method for collaborative construction of a trusted coal mine safety knowledge graph for blockchain, which is oriented to the collaborative construction and secure storage of a coal mine safety knowledge graph. It utilizes the coal mine safety knowledge contributed by different knowledge submitters to form a coal mine safety knowledge graph and stores it in the blockchain distributed ledger. A trusted knowledge quality management strategy for collaborative construction of a knowledge graph is also proposed in this method to manage the possible incorrect knowledge in the knowledge submission links of multiple contributors, thereby avoiding the negative impact of low-trust knowledge submitted on the quality of the knowledge graph. It provides guarantees for constructing a trusted coal mine knowledge graph and a secure knowledge storage system. As Figure 2 shown, it shows the overall process of the method, mainly including the following four parts.

[0044] 1. Knowledge collection.

[0045] In the embodiments of the present invention, in the field of coal mine safety, knowledge submitters are organized to submit the target knowledge to be collected, that is, knowledge is collected under the given types of knowledge. For example: knowledge of coal mine safety production regulations, knowledge in the field of coal mine water hazard prevention, etc., and the triple form is used as the basic unit of knowledge. Among them, the submitted triples are stored in the basic form of "head entity - relationship - tail entity", and different knowledge submitters can submit duplicate knowledge for different types of knowledge to be collected and specific facts. Among them, a fact refers to the object described by specific knowledge. For example, for the fact defined as "drain pipe diameter limit", it can be described by specific triple knowledge such as "drain pipe - diameter - less than 0.5m", "drain pipe - diameter - less than 0.2m", etc.

[0046] 2. Storage of coal mine safety knowledge.

[0047] 1) Storage of knowledge.

[0048] In the embodiments of the present invention, the knowledge of each fact is in the form of a triple. A unique ID is assigned to the head entity, relationship, and tail entity in each triple, and each triple is stored in the form of the ID numbers corresponding to its head entity, relationship, and tail entity; a unique ID is assigned to each fact, and all triples are organized according to the corresponding facts and stored in the key-value structure on the distributed ledger of the blockchain.

[0049] Specifically: The triples provided by different knowledge submitters are organized and stored according to the facts they describe. All submitted triples and the facts corresponding to their descriptions are saved on a distributed ledger. For the blockchain framework, a consortium chain can be used as the basic framework, such as the HyperLeger Fabric framework, etc. The ledger information storage input is mainly in the form of key-value pairs. In order to store knowledge and its quality information on the ledger, the triples (i.e., entities and relationships) are converted into three data structures that can support independent operations. These data are saved in the key-value store as part of the ledger state.

[0050] Some examples of IDs are provided below:

[0051] Entity: E_ID - E_NAME

[0052] Relationship: R_ID - R_NAME

[0053] The SOTA triple describing a fact: F_ID - (E_ID, R_ID, E_ID)

[0054] All triples describing a fact: F_ID - {C_ID}

[0055] Among them, SOTA represents the triple with the best current quality. In addition, each entity (including the head entity and the tail entity) and relationship are identified by a unique ID index. For example, in the above example, E_ID and R_ID are the IDs of the entity and relationship, and E_NAME and R_NAME are the specific contents of the entity and relationship. On this basis, each triple is stored in the form of the ordered ID numbers corresponding to its head entity, relationship, and tail entity, that is, (E_ID, R_ID, E_ID) in the above example. For example, a triple (en1, re1, en2) is stored as a list data structure of IDe1, IDr1, IDe2, and the elements in this list correspond to the ID numbers of the entity en1, the relationship re1, and the entity en2 respectively. Since knowledge may be contributed by many knowledge submitters, for the descriptive knowledge uploaded by different submitters for a certain fact, the triple with the best quality is selected to construct the knowledge graph. In addition, all triples are organized according to their corresponding facts, that is, F_ID - (E_ID, R_ID, E_ID) in the above example, and F_ID is the ID of the fact. In this way, the triples from different knowledge submitters can be aggregated into groups describing the same fact, as shown in the key-value store of "All triples describing a fact" above, where C_ID is a unique ID, specifically representing the unique identity of a certain knowledge submitter, and it is also used in the design of the information storage data structure of the knowledge submitter.

[0056] This grouped storage is beneficial to statistically describe the consistency among the knowledge of a fact in the blockchain framework, thus facilitating the calculation of knowledge quality and providing support for evaluating the credibility of knowledge submitters. When submitting new triples to a fact group, the triples therein participate in the evaluation of new knowledge quality. At the same time, multiple knowledge submitters associated with the fact ID (i.e., F_ID) are also organized and recorded together to support the quality evaluation process. For the specific solution, please refer to the following introduction.

[0057] 2) Storage of information of knowledge submitters.

[0058] In the embodiments of the present invention, the knowledge corresponding to each fact submitted by each knowledge submitter is stored in the distributed ledger of the blockchain in the form of ID using a key-value structure, and each triple includes a corresponding quality score. Specifically, the information of knowledge submitters is maintained in the distributed ledger state to support the quality evaluation of triples and the update of SOTA triples. The knowledge provided by each knowledge submitter about each fact is stored together with the corresponding knowledge quality. The knowledge submitters stored in the ledger are represented in the form of a key-value structure, specifically:

[0059] Knowledge submitter: C_ID - {F_ID - (E_ID, R_ID, E_ID, Q)}

[0060] Among them, Q represents the estimated quality score of the corresponding triple (calculated by the iterative update method introduced later). And when a knowledge submitter submits a new triple, the key-value set corresponding to each knowledge submitter will be updated. The key-value set has a unique ID represented as C_ID. For example, in a specific task, C_ID can be instantiated as "A0001", which can represent a certain submitter. Also, using the above key-value structure, the knowledge set submitted by the knowledge submitter can be found through A0001. Based on this storage method of knowledge submitter information, it is convenient to evaluate the overall credibility of each knowledge submitter through the quality of the knowledge submitted by each knowledge submitter. This overall credibility can in turn be used to evaluate the quality of triples related to the submission of knowledge by knowledge submitters, such as newly submitted knowledge.

[0061] 3. Knowledge quality management for blockchain smart contracts.

[0062] In the solution provided by the embodiments of the present invention, the smart contract of the blockchain is used to verify the validity of knowledge. The following proposed method is integrated into the smart contract of the blockchain by constructing an algorithm module. It includes the evaluation of the credibility of knowledge submitters and the evaluation of knowledge quality. The following details the algorithm logic that needs to be integrated into the smart contract. For the smart contract of the blockchain, the implementation and deployment method of the smart contract of an open-source blockchain framework such as HyperLeger Fabric can be used as the basis.

[0063] 1) Knowledge validity check.

[0064] For the facts already existing in the knowledge graph, that is, the knowledge describing this fact already exists in the current bill, the validity of the newly submitted triples describing the same fact is checked by two processes. First, the current fact and the ID numbers of the entities and relationships already stored should exist in the corresponding key-value store. In addition, whether the submitted triples are relevant to the fact should be verified according to certain rules or check criteria. That is, it is required that the head and tail entities should be the same as the existing triples about this fact. This verification mechanism is established to ensure that knowledge submitters for a certain fact description cannot upload completely irrelevant triples to the existing fact without limit. For facts that do not exist in the knowledge graph, that is, there is no knowledge describing this fact in the current knowledge graph, an invalid fact ID recycling mechanism is proposed.

[0065] It mainly includes the following steps: When the knowledge of a new fact is submitted, a unique ID (F_ID) is assigned to the new fact and associated with the corresponding knowledge. During a subsequent set time period (such as x days), if the number of triples related to this F_ID submitted by other knowledge submitters exceeds the set quantity threshold, then this F_ID will be considered valid, otherwise it will be marked as suspicious. The quantity threshold is represented by the parameter α, which is an adjustable parameter, and the specific value can be set by the user according to the actual situation or experience, such as the average knowledge quantity describing different facts. The suspicious facts and their related triples are delivered to other knowledge submitters for voting. F_IDs without enough votes (that is, the number of votes does not exceed the set value) will be deleted from the key-value store and recycled for other triples on new facts. This processing mechanism is to prevent knowledge submitters from uploading triples unrelated to the knowledge graph to be constructed without limit.

[0066] 2) Knowledge quality evaluation scheme for multiple knowledge submitters.

[0067] The knowledge quality evaluation algorithm for multiple knowledge submitters aims to evaluate the quality of triples from different knowledge submitters and be integrated into the smart contract as a module. To ensure the knowledge quality in the knowledge graph, a knowledge quality evaluation method based on knowledge consistency is used to evaluate the knowledge quality during the dynamic knowledge graph construction process. Its aim is to construct unified evaluation indicators to evaluate the quality of all knowledge describing a certain fact, and select high-quality triples from possibly conflicting fact descriptions as the shared knowledge. This module takes the credibility of each knowledge submitter and all triples of the fact ID associated with it as input, and evaluates the quality of each triple. When the newly submitted knowledge is different from the triples already shared, it re-updates the quality of all triples corresponding to the fact of the new knowledge. At this time, only the quality of non-follower triples will be updated according to the evaluation result. The triple with the highest quality is used as the new SOTA triple describing a certain fact and is shared.

[0068] The multi-knowledge submitter credible knowledge screening algorithm aims to evaluate the quality of triples from different knowledge submitters and is integrated into the smart contract as a module. To ensure the quality of knowledge in the knowledge graph, a knowledge quality evaluation method based on knowledge consistency is used to evaluate knowledge quality during the dynamic knowledge graph construction process. It aims to construct a unified evaluation index to evaluate the quality of all knowledge describing a certain fact, and select high-quality triples from possibly conflicting fact descriptions as the shared knowledge. This module takes the credibility of each knowledge submitter and all triples of the fact IDs related to it as input, and evaluates the quality of each triple. When the newly submitted knowledge is different from the shared triples, it re-updates the quality of all triples corresponding to the fact of the new knowledge. At this time, only the quality of non-follower triples will be updated according to the evaluation results. The triple with the highest quality is used as the new triple describing a certain fact and is shared.

[0069] For each valid fact screened out, let N be the number of triples describing the fact. These triples are represented as {k 1 , k 2 , …, k N}, which come from different knowledge submitters, and the credibility of the corresponding different knowledge submitters is represented as {p 1 , p 2 , …, p N}; where each k represents a triple, the subscript is the triple number, and each p represents the credibility of the corresponding knowledge submitter, and the subscript corresponds to the triple number. The quality evaluation is carried out in an iterative manner. In each iteration, each triple takes turns as the triple k i to be evaluated, and the other triples in the same fact group are regarded as external triples k e . The credibility of the knowledge submitter corresponding to k i and the consistency with k e are used to evaluate the quality of k i , and the quality of all triples is updated by taking each triple as the object to be evaluated in turn. The estimated quality of this triple to be evaluated and a certain external triple in the l-th iteration is expressed as and

[0070] The specific process is as follows: For the triple k i to be evaluated and the external triple k e , its initial quality in the iteration is initialized with the credibility of the corresponding knowledge submitter (at the initial moment, an initial credibility is set for each knowledge submitter), and is expressed as:

[0071]

[0072] For the update process, it is based on the consistency between the knowledge to be evaluated (triple) and external knowledge. When the knowledge to be evaluated is consistent or inconsistent with other knowledge, its quality will increase or decrease accordingly. Specifically, the quality of k i in the iteration update in the l-th iteration is expressed as:

[0073]

[0074] where is the quality score of the triple k obtained in the l-th iteration i ; is the quality score of the triple k obtained in the (l - 1)-th iteration i ; is the upper bound of quality update, which is calculated using and the consistency with k e ; is the update degree, which is calculated using and , the quality score of the triple k obtained in the (l - 1)-th iteration e ; when l = 1,

[0075] Iterate continuously until the maximum number of iterations L is reached (i.e., l = L), and the obtained at this time is used as the quality score of the triple k i , i ∈ {1, 2,..., N}.

[0076] The upper bound of quality update is the extreme value that the change in knowledge quality can reach in one iteration, reflecting the situation when knowledge is extremely supported or negated. In addition, it is also related to the number of iterations, indicating that after a certain number of iterations, the knowledge quality tends to be stable. is defined as:

[0077]

[0078] where k e = k i means that k i and k e have the same head entity, relation, and tail entity, that is, the two triples are consistent. If k i is consistent with the external triple k e , then the quality of k i will increase, and if it is inconsistent, the quality of k i will decrease. is a convergence factor that balances the impact of external comparison and the credibility of knowledge submitters on quality assessment.

[0079] For the update range depends on k i and k e 's current quality. When the external knowledge is authoritative and the gap between the two is large, the quality of the knowledge to be evaluated can be greatly updated. Specifically, is expressed as:

[0080]

[0081] is jointly determined by the quality of external knowledge and the knowledge to be evaluated, and it consists of two factors, namely and For the quality of external knowledge it is separately used as a factor for the degree of update because external knowledge k with higher credibility e should be more persuasive. For the other factor, the difference between the credibility of k i itself and the credibility of external knowledge will also affect the update range. If is low, then external knowledge with high prior quality may greatly affect the quality update of k i because it is more credible knowledge than k i and the greater the difference, the greater the impact. On the contrary, if the prior of k i is very high, then the impact of k e on the quality update of k i is not as large as in the previous case because both pieces of knowledge are empirically credible and there is more confidence in their statements.

[0082] 3. Knowledge graph construction and update.

[0083] In the embodiments of the present invention, for each valid fact, the triple with the highest quality score is selected, and the triples with the highest quality scores of all valid facts are integrated to construct and store a coal mine safety knowledge graph; at the same time, after new knowledge is submitted, the above scheme is used to estimate the quality of new knowledge based on the consistency between the knowledge describing a fact, and update the quality of all current knowledge describing the same fact, and then update the knowledge graph.

[0084] In the embodiments of the present invention, through the knowledge graph, the knowledge in each information source can be integrated and graphically connected to form a large, structured knowledge base, which is convenient for users to find the required knowledge and also convenient for subsequent knowledge updates.

[0085] II. Example introduction.

[0086] The present invention takes the upload and quality evaluation of triples submitted by multiple knowledge submitters for the same fact as a detailed example, mainly elaborates on how to process the knowledge submitted by multiple knowledge submitters to form a key-value storage form acceptable to the blockchain, as well as processes such as quality evaluation. For the architecture of the blockchain framework, an open-source blockchain framework such as HyperLeger Fabric can be adopted.

[0087] Step a1: Assume that there are currently three coal mine safety knowledge submitters, denoted as {A1, A2, A3} respectively. These three people submit knowledge about the description of a specific fact. For example, knowledge submitters {A1, A2, A3} respectively submit knowledge statements about the fact of "the mining depth limit of the first production level in outburst-prone mines". The triples submitted are {A1: "For the first production level mining in outburst-prone mines, the depth is less than or equal to 800m", A2: "For the first production level mining in outburst-prone mines, the depth is less than or equal to 800m", A3: "For the first production level mining in outburst-prone mines, the depth is less than or equal to 300m"}. The knowledge submitted by the above multiple knowledge submitters is saved in the key-value storage as part of the blockchain ledger state. Taking the knowledge submitted by A1 as an example:

[0088] Entities: {E0001 - The first production level mining in outburst-prone mines; E0002 - Less than or equal to 800m; E0003 - Less than or equal to 300m};

[0089] Relationship: R0001 - Depth;

[0090] The triple representation describing this fact: F_0001 - (E0001, R0001, E0002);

[0091] All triples describing a certain fact: F_0001 - {A0001};

[0092] Among them, E0001, E0002, F_0001, A0001, etc. are unique index ID identifiers assigned by the system, where A0001 is an instantiation of the identity identifier C_ID. Each entity, relationship, the described fact, and the knowledge submitter are all identified by unique ID indexes.

[0093] Step a2: After assigning storage IDs to the triples and the described fact, multiple knowledge submitters related to the current fact, that is, the fact numbered "F_0001" in this example, are also organized and recorded together to support the quality evaluation process. For example, in this example, the three knowledge submitters A1, A2, and A3 respectively contributed three triples related to the fact F_0001. The key-value items corresponding to all triple knowledge submitters describing the fact F_0001 are: F_0001 - {A0001, A0002, A0003}.

[0094] Step a3: In the maintenance of blockchain ledger information, the information of the knowledge submitter is also maintained in the ledger state. Specifically, for the storage of the knowledge submitter information, a new data structure is constructed to store the knowledge provided by each knowledge submitter and its related information. The knowledge submitters stored on the ledger are represented in a key-value structure. For example, in this example, the storage information of knowledge submitter A1 can be represented as:

[0095] A0001-{F_0001-(E0001,R0001,E0002,Q)}

[0096] Among them, "Q" represents the estimated quality of the corresponding triple, and the value range is (0, 1). When the knowledge submitter submits a triple, the corresponding key-value set of the knowledge submitter will be updated. Based on this storage method of knowledge submitter information, it is convenient to evaluate the overall credibility of each knowledge submitter by the quality of the knowledge submitted by each knowledge submitter. This overall credibility can in turn be used to evaluate the quality of the triples related to the submission of the knowledge submitter, such as newly submitted knowledge, etc. The specific credibility calculation method is shown in step b3.

[0097] Step b1: After storing the uploaded triples logically, perform validity verification on the newly added triples. In this example, when knowledge submitter A1 submits the knowledge (E0001, R0001, E0002), a new fact ID, that is, F_0001, is assigned to the newly submitted triple. Within a certain time range subsequently, if the number of triples related to this fact ID, that is, F_0001, submitted by other knowledge submitters exceeds a certain threshold, then this fact ID will be regarded as valid, otherwise it will be marked as suspicious.

[0098] For example, when knowledge submitters A2 and A3 submit knowledge related to fact F_0001 within a certain time threshold α, this ID: F_0001 is marked as valid. The suspicious facts and their related triples are delivered to other knowledge submitters for voting. If the number of votes is less than half of the number of voters, the fact ID: F_0001 will be recycled, deleted from the key-value storage, and its related knowledge will be regarded as invalid knowledge. This processing mechanism is to prevent knowledge submitters from uploading triples unrelated to the graph to be constructed without limit.

[0099] Step b2: For the submission of valid knowledge, next, perform quality evaluation. In this example, take the triple (E0001, R0001, E0002) uploaded by knowledge submitter A1 as an example. The quality of the triple to be evaluated is initialized using the credibility of the knowledge submitter. Assume the initial credibility of knowledge submitter A1 is p 1 , and the initial quality of the current triple is

[0100]

[0101] Next, use the knowledge submitted by other knowledge submitters to update the quality of the knowledge according to the consistency with the current triple to be evaluated. Compare the knowledge to be evaluated with the knowledge submitted by different other knowledge submitters in turn. The update formula is as follows:

[0102]

[0103] where is the upper bound of quality update, is the update degree, l is the iteration round. That is, when all external knowledge is calculated, l = l + 1, and the range of l is (0, L), where L is an adjustable parameter, namely the maximum iteration round. represents the extreme value that the change in knowledge quality can reach in one iteration. Among them is defined as:

[0104]

[0105] where k e = k i means that k i and k e have the same head entity, relationship and tail entity, that is, the two triples are consistent. In this example, the knowledge submitted by three knowledge submitters for the current fact is: {A1: "For the first production level mining in outburst-prone mines, the depth is less than or equal to 800m", A2: "For the first production level mining in outburst-prone mines, the depth is less than or equal to 800m", A3: "For the first production level mining in outburst-prone mines, the depth is less than or equal to 300m"}. According to the formula, when using the knowledge submitted by A2 to update the quality of the current knowledge, since the knowledge submitted by knowledge submitters A1 and A2 is the same, that is, use the formula to calculate and then use and to calculate

[0106]

[0107] is jointly determined by the quality of the external knowledge in the previous iteration and the quality of the knowledge to be evaluated. At this time, when the quality of the knowledge submitted by A1 for the fact F_0001 is calculated, the quality of the knowledge submitted by A2 and A3 is calculated in turn. When is calculated, let l = l + 1 and perform the next round of calculation. When l is equal to the iteration round upper limit L, the iteration is completed. Obtain the quality Among them, the triple with the highest knowledge quality in the set is taken as the triple describing fact F1. For example, in this example, if the triple submitted by A1 has the highest quality, its key-value pair is updated, such as: F_0001-(E0001,R0001,E0002). That is, the triple describing the fact that "the mining depth limit of the first production level of outburst mines" is confirmed as "for the first production level of outburst mines, the mining depth is less than or equal to 800m", and is saved in the blockchain ledger in the form of the key-value pair {F_0001-(E0001,R0001,E0002)}.

[0108] Step b3: Further, after obtaining the knowledge quality of all submissions of each knowledge submitter, update the credibility of the current knowledge submitter. Calculate the average value and update its credibility p i Suppose the set of the quality of all knowledge uploaded by A1 is {Q 1 ,Q 2 ,…,Q n}, then the credibility of the knowledge submitter A1 is updated to:

[0109] The above scheme provided by the embodiments of the present invention mainly has the following advantages:

[0110] 1) It can solve the knowledge quality problem and security problem faced by multiple contributors in constructing the knowledge graph during the construction of the coal mine safety knowledge base.

[0111] 2) The adopted knowledge quality assessment scheme can utilize the consistency between knowledge to estimate the credibility of knowledge and ensure the impact of the wrong knowledge uploaded during the collaborative construction process on the graph quality.

[0112] 3) It can be adapted to the construction of various coal mine safety domain knowledge graphs, and it does not depend on specific domain information. It is a collaborative construction framework for knowledge graphs and can provide a low-cost collaborative construction implementation method for complex and large-scale coal mine knowledge integration and knowledge base generation.

[0113] Through the description of the above implementation manners, those skilled in the art can clearly understand that the above embodiments can be implemented by software or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions of the above embodiments can be embodied in the form of a software product, and the software product can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments of the present invention.

[0114] Embodiment 2

[0115] The present invention also provides a collaborative construction system for a trusted coal mine safety knowledge graph oriented to a blockchain, which is mainly used to implement the method provided in the foregoing embodiments, such as Figure 3 As shown, the system mainly includes:

[0116] A knowledge collection unit, configured to collect knowledge of each fact in the field of coal mine safety submitted by a knowledge submitter;

[0117] A knowledge storage and quality management unit, configured to assign a unique ID to the knowledge of each fact, organize the corresponding knowledge according to the fact, and store it in the blockchain in the form of an ID; and store the knowledge corresponding to each fact submitted by each knowledge submitter in the blockchain in the form of an ID, and each piece of knowledge corresponding to each stored fact includes a corresponding quality score; wherein, for the knowledge of each newly submitted fact, perform a validity check to determine whether the corresponding fact is valid; for each piece of knowledge of each valid fact, use the credibility of the corresponding knowledge submitter and the consistency with other knowledge under the same valid fact to perform a quality evaluation to obtain a quality score; the credibility of the knowledge submitter is calculated using the quality scores of the knowledge corresponding to each fact submitted by it;

[0118] A coal mine safety knowledge graph construction unit, configured to, for each valid fact, select the triple with the highest quality score, and construct a coal mine safety knowledge graph by integrating the triples with the highest quality scores of all valid facts.

[0119] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above division of each functional module is used as an example. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the system is divided into different functional modules to complete all or part of the functions described above.

[0120] Embodiment III

[0121] The present invention also provides a processing device, such as Figure 4 As shown, it mainly includes: one or more processors; a memory, configured to store one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the method provided in the foregoing embodiments.

[0122] Further, the processing device further includes at least one input device and at least one output device; in the processing device, the processor, the memory, the input device, and the output device are connected through a bus.

[0123] In the embodiments of the present invention, the specific types of the memory, the input device, and the output device are not limited; for example:

[0124] The input device can be a touch screen, an image acquisition device, a physical button, a mouse, etc.;

[0125] The output device can be a display terminal;

[0126] The memory can be a Random Access Memory (RAM), or a non-volatile memory, such as a disk memory.

[0127] Embodiment 4

[0128] The present invention also provides a readable storage medium storing a computer program, which when executed by a processor implements the method provided in the foregoing embodiments.

[0129] In the embodiments of the present invention, the readable storage medium, as a computer-readable storage medium, can be disposed in the foregoing processing device, for example, as the memory in the processing device. In addition, the readable storage medium can also be various media that can store program codes, such as a USB flash drive, a mobile hard disk, a Read-Only Memory (ROM), a magnetic disk, or an optical disc.

[0130] As described above, the foregoing are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A blockchain-oriented method for collaboratively constructing a trusted coal mine safety knowledge graph, characterized in that: include: Collect the knowledge of various facts in the field of coal mine safety submitted by knowledge submitters; Assign a unique ID to each factual knowledge, organize the corresponding knowledge according to the facts, and store it in the blockchain in the form of ID; The knowledge corresponding to each fact submitted by each knowledge submitter is stored in the blockchain in the form of an ID, and the knowledge corresponding to each fact stored contains a corresponding quality score; wherein, the validity of the knowledge of each newly submitted fact is checked to determine whether the corresponding fact is valid; for each knowledge of each valid fact, the credibility of the corresponding knowledge submitter and the consistency with other knowledge under the same valid fact are used to conduct a quality evaluation and obtain a quality score; The credibility of the knowledge submitter is calculated using the quality score of the knowledge corresponding to each fact submitted by the submitter; For each valid fact, the triple with the highest quality score is selected, and the triples with the highest quality scores of all valid facts are combined to construct a coal mine safety knowledge graph.

2. According to a blockchain-oriented trusted coal mine safety knowledge graph collaborative construction method according to claim 1, it is characterized in that: The collecting of knowledge of various facts in the field of coal mine safety submitted by knowledge submitters includes: oriented to the field of coal mine safety, collecting knowledge of various facts of various types submitted by knowledge submitters in the form of triples.

3. According to a blockchain-oriented trusted coal mine safety knowledge graph collaborative construction method according to claim 1, it is characterized in that: The above-mentioned assigns a unique ID to the knowledge of each fact, organizes the corresponding knowledge according to the fact, and stores it in the blockchain in the form of an ID; The knowledge corresponding to each fact submitted by each knowledge submitter is stored in the blockchain in the form of an ID. The knowledge corresponding to each fact stored contains the corresponding quality score including: The knowledge of each fact is in the form of triples. Unique IDs are assigned to the head entity, relationship, and tail entity in each triple. Each triple is stored in the form of the ID numbers corresponding to its head entity, relationship, and tail entity. Each fact is assigned a unique ID, and all triples are organized according to the corresponding facts and stored in the blockchain's distributed ledger in a key-value structure; The knowledge corresponding to each fact submitted by each knowledge submitter is stored in the distributed ledger of the blockchain in the form of an ID using a key-value structure. Each triple contains a corresponding quality score.

4. According to a blockchain-oriented trusted coal mine safety knowledge graph collaborative construction method according to claim 1, it is characterized in that: The validity check of the knowledge of each newly submitted fact to determine whether the corresponding fact is valid includes: When the knowledge of a new fact is submitted, a unique ID is assigned to the new fact and associated with the corresponding knowledge. Within a set time range, if the number of knowledge associated with the ID of the new fact exceeds a set threshold, the new fact is a valid fact; otherwise, it is marked as a suspicious fact. The suspicious fact and its related knowledge are delivered to other knowledge submitters for voting. If the number of votes does not exceed the set value, the ID of the new fact is recycled.

5. According to a blockchain-oriented trusted coal mine safety knowledge graph collaborative construction method according to claim 1, it is characterized in that: For each knowledge of each valid fact, the quality evaluation is performed using the credibility of the corresponding knowledge submitter and the consistency with other knowledge under the same valid fact, and the quality score obtained includes: Each piece of knowledge of each valid fact is in the form of a triple. The number is N, and the N triples are represented by {k1, k2, …, k N }, each k represents a triple, the subscript is the triple number, and the credibility of the knowledge submitter corresponding to N triples is expressed as {p1, p2, …, p N }, each p represents the credibility of the corresponding knowledge submitter, and the subscript corresponds to the triple number; The quality evaluation is performed in an iterative manner. In each iteration, each triplet is taken as the triplet k to be evaluated. i , the other triples in the same fact group are considered as external triples k e , using k i The credibility of the corresponding knowledge submitter and k w The consistency of k i Quality evaluation is performed, and the quality of all triples is updated by taking each triple as the object to be evaluated in turn, and finally the quality score of each triple is obtained; where i,e∈{1,2,…,N},e≠i.

6. A blockchain-oriented trusted coal mine safety knowledge graph collaborative construction method according to claim 5, characterized in that: Each iteration process includes: For the lth iteration, k i The quality evaluation formula is: in, is the triple k obtained in the lth iteration i The mass fraction of, when l = L, As triple k i The quality score of , L is the maximum iteration round; is the triple k obtained in the l-1th iteration i The quality score of is the upper bound of the quality update, which uses and k e Calculated by consistency; To update the level, use and To calculate, The triple k obtained in the l-1th iteration e The mass fraction of 7. A blockchain-oriented trusted coal mine safety knowledge graph collaborative construction method according to claim 6, characterized in that: Upper bound on quality updates With the update level The calculation formula is expressed as:

8. A blockchain-oriented trusted coal mine safety knowledge graph collaborative construction system, characterized by: The method for implementing any one of claims 1 to 7 comprises: A knowledge collection unit, used to collect knowledge of various facts in the field of coal mine safety submitted by knowledge submitters; The knowledge storage and quality management unit is used to assign a unique ID to the knowledge of each fact, organize the corresponding knowledge according to the fact, and store it in the blockchain in the form of an ID; and store the knowledge corresponding to each fact submitted by each knowledge submitter in the blockchain in the form of an ID, and the knowledge corresponding to each fact stored contains a corresponding quality score; wherein, the validity check is performed on the knowledge of each newly submitted fact to determine whether the corresponding fact is valid; for each knowledge of each valid fact, the credibility of the corresponding knowledge submitter and the consistency with other knowledge under the same valid fact are used to perform quality evaluation and obtain a quality score; the credibility of the knowledge submitter is calculated using the quality score of the knowledge corresponding to each fact submitted by the knowledge submitter; The coal mine safety knowledge graph construction unit is used to select the triple with the highest quality score for each valid fact, and to construct the coal mine safety knowledge graph by integrating the triples with the highest quality scores of all valid facts.

9. A processing device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images