Industrial knowledge graph generation method and device, equipment and storage medium

Abstract

The application belongs to the field of industrial intelligent manufacturing, and discloses an industrial knowledge graph generation method, device, equipment and storage medium. The method comprises the following steps: constructing preset standard data into a first knowledge graph based on an association pattern model, wherein the preset standard data comprises at least one of experience rules, physical and chemical mechanism models, artificial intelligence algorithm models and process configuration diagrams in an actual industrial production process; determining a knowledge source in an industrial production process, performing co-reference resolution and entity disambiguation on data contained in the knowledge source to obtain a second knowledge graph; and generating a target knowledge graph according to the first knowledge graph and the second knowledge graph. Since the target knowledge graph is generated according to the first knowledge graph constructed based on preset standard data and the second knowledge graph constructed based on a knowledge source in an industrial production process, the preset standard data and the knowledge source in the industrial production process can be efficiently integrated, and the integration and condensation of industrial data can be realized.

Description

Technical Field

[0001] This invention relates to the field of industrial intelligent manufacturing technology, and in particular to a method, apparatus, equipment and storage medium for generating industrial knowledge graphs. Background Technology

[0002] With the rapid development of new-generation information technologies such as artificial intelligence, the Internet of Things, cloud computing, and the Industrial Internet, industrial production processes are moving towards intelligence, automation, and greening, driving innovation in future industrial manufacturing production models. The innovation of new models in intelligent manufacturing driven by these new-generation information technologies is mainly reflected in two aspects: first, in industrial data acquisition and integration, including cloud-edge data and market operation decision data from different industrial production levels; second, in the integration and aggregation of industrial data, primarily extracting relevant operational experience, production rules, and physicochemical mechanisms from production data and applying this knowledge to industrial production. These data sources include: configuration diagrams of production process topology, time-series data of actual production process operation, process quality index data, and experience data from skilled workers and experts. From a data modality perspective, industrial process knowledge involves not only structured data such as time-series signals and vibration response data, but also unstructured data such as text, audio, and video, and even semi-structured data such as configuration diagrams, characters, and XML files, exhibiting significant multi-source heterogeneity. Therefore, how to process the complex, multi-source, heterogeneous information in industry and extract knowledge describing the changes in various nodes and equipment in industrial production has become a key infrastructure for my country's industrial development from a production mode of partial adjustment and extensive operation to a production mode of full-process optimization and refined operation and control.

[0003] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main objective of this invention is to provide an industrial knowledge graph generation method, apparatus, device, and storage medium, which aims to solve the technical problem of low efficiency in integrating multi-source heterogeneous data generated in the industrial production field in the prior art.

[0005] To achieve the above objectives, the present invention provides a method for generating an industrial knowledge graph, the method comprising the following steps:

[0006] Based on the association pattern model, a first knowledge graph is constructed from preset standard data, which includes at least one of the following: empirical rules in actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams.

[0007] The knowledge sources in the industrial production process are identified, and the data contained in the knowledge sources are subjected to coreference resolution and entity disambiguation to obtain a second knowledge graph.

[0008] Generate a target knowledge graph based on the first knowledge graph and the second knowledge graph.

[0009] Optionally, the step of determining the knowledge source in the industrial production process, performing coreference resolution and entity disambiguation on the data contained in the knowledge source, and obtaining the second knowledge graph includes:

[0010] Identify the knowledge sources in the industrial production process and determine the data categories of the knowledge sources;

[0011] Based on the data categories, the knowledge sources are extracted and identified to generate a case graph;

[0012] The data in the case graph are subjected to coreference resolution and entity disambiguation to obtain a second knowledge graph.

[0013] Optionally, the step of performing coreference resolution and entity disambiguation on the data in the case graph to obtain the second knowledge graph includes:

[0014] The first entity information and the second entity information are extracted from the case graph using a pre-defined relationship extraction algorithm cluster.

[0015] Determine the entity similarity and entity attribute similarity between the first entity information and the second entity information;

[0016] The case graph is subjected to coreference resolution based on the entity similarity and the entity attribute similarity to obtain the target case graph;

[0017] Based on the first entity information and the second entity information, entity disambiguation is performed on the target case graph to obtain a second knowledge graph.

[0018] Optionally, the step of performing entity disambiguation on the target case graph based on the first entity information and the second entity information to obtain a second knowledge graph includes:

[0019] The first entity information is isomorphized according to the spatial vector model and the semantic space model to obtain the first isomorphic space;

[0020] The second entity information is isomorphized according to the spatial vector model and the semantic space model to obtain the second isomorphic space;

[0021] Based on the first isomorphic space and the second isomorphic space, entity disambiguation is performed on the target case graph to obtain the second knowledge graph.

[0022] Optionally, the step of extracting and identifying data from the knowledge source based on the data category to generate a case graph includes:

[0023] When the data category is structured data, data extraction is performed on the data contained in the knowledge source, and triples are constructed based on the extracted data;

[0024] A case graph is generated based on the triplet;

[0025] When the data category is unstructured data, data extraction is performed on the data contained in the knowledge source based on the association pattern model, and a case graph is generated based on the extracted data.

[0026] Optionally, the step of performing coreference resolution on the data contained in the knowledge source includes:

[0027] Obtain the association mapping model between entities and alias entities;

[0028] The coreference resolution of the data contained in the knowledge source is performed based on the entity information in the knowledge source and the association mapping model.

[0029] Optionally, after the step of constructing the first knowledge graph from preset standard data based on the association pattern model, the method further includes:

[0030] Determine the target entity and the entity attribute corresponding to the target entity in the preset standard data;

[0031] Construct an association mapping model between the entity and its alias based on the target entity and the entity's attributes.

[0032] Furthermore, to achieve the above objectives, the present invention also provides an industrial knowledge graph generation apparatus, the apparatus comprising:

[0033] A construction module is used to construct a first knowledge graph based on a relational pattern model using preset standard data. The preset standard data includes at least one of the following: empirical rules in actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams.

[0034] The data processing module is used to determine the knowledge sources in the industrial production process, and to perform coreference resolution and entity disambiguation on the data contained in the knowledge sources to obtain a second knowledge graph.

[0035] The generation module is used to generate a target knowledge graph based on the first knowledge graph and the second knowledge graph.

[0036] Furthermore, to achieve the above objectives, the present invention also proposes an industrial knowledge graph generation device, the device comprising: a memory, a processor, and a knowledge graph generation program stored in the memory and executable on the processor, the knowledge graph generation program being configured to implement the steps of the industrial knowledge graph generation method described above.

[0037] Furthermore, to achieve the above objectives, the present invention also proposes a storage medium storing a knowledge graph generation program, which, when executed by a processor, implements the steps of the industrial knowledge graph generation method described above.

[0038] This invention constructs a first knowledge graph from pre-defined standard data based on an association pattern model. The pre-defined standard data includes at least one of the following: empirical rules from actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams. It then identifies knowledge sources within the industrial production process, performs coreference resolution and entity disambiguation on the data contained in these knowledge sources, and obtains a second knowledge graph. Finally, it generates a target knowledge graph based on the first and second knowledge graphs. Because this invention generates the target knowledge graph from the first knowledge graph constructed from pre-defined standard data and the second knowledge graph constructed from knowledge sources within the industrial production process, it can efficiently integrate the pre-defined standard data and knowledge sources within the industrial production process, achieving the integration and aggregation of industrial data. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the structure of the knowledge graph generation device for the hardware operating environment involved in the embodiments of the present invention;

[0040] Figure 2 This is a flowchart illustrating the first embodiment of the industrial knowledge graph generation method of the present invention;

[0041] Figure 3 This is a flowchart illustrating the second embodiment of the industrial knowledge graph generation method of the present invention;

[0042] Figure 4 This is a schematic diagram of the target knowledge graph generation framework in the second embodiment of the industrial knowledge graph generation method of the present invention;

[0043] Figure 5 This is a structural block diagram of the first embodiment of the knowledge graph generation device of the present invention.

[0044] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0045] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0046] Reference Figure 1 , Figure 1 This is a schematic diagram of the knowledge graph generation device structure of the hardware operating environment involved in the embodiments of the present invention.

[0047] like Figure 1 As shown, the knowledge graph generation device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0048] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the knowledge graph generation device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0049] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a knowledge graph generation program.

[0050] exist Figure 1 In the knowledge graph generation device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the knowledge graph generation device of the present invention can be set in the knowledge graph generation device, and the knowledge graph generation device calls the knowledge graph generation program stored in the memory 1005 through the processor 1001 and executes the industrial knowledge graph generation method provided in the embodiment of the present invention.

[0051] Based on the aforementioned knowledge graph generation device, this invention provides a method for generating an industrial knowledge graph, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the industrial knowledge graph generation method of the present invention.

[0052] In this embodiment, the industrial knowledge graph generation method includes the following steps:

[0053] Step S10: Construct a first knowledge graph based on the pre-set standard data using the association pattern model. The pre-set standard data includes at least one of the following: empirical rules in actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams.

[0054] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a mobile phone, tablet computer, or personal computer, or an electronic device or knowledge graph generation device capable of performing the above functions. The following description uses the knowledge graph generation device as an example to illustrate this embodiment and the subsequent embodiments.

[0055] It should be noted that the association pattern model may include multiple relational pattern models such as association rules and association templates. The association rules may be appropriate association rules extracted based on the analysis of the commonalities and differences between certain existing knowledge based on the experience of domain experts. The association templates may be analyzed from the perspective of data, using methods such as normalized variable analysis and cross-correlation analysis to analyze the association patterns between variable entities.

[0056] It should be understood that knowledge entities reflect static information about equipment, processes, and products, making it difficult to represent them uniformly using a single relational model. Therefore, it is necessary to process the data information in different types of production process knowledge entities based on expert experience and association pattern analysis, and to construct multiple different association pattern models to build the first knowledge graph from the preset standard data.

[0057] Step S20: Determine the knowledge sources in the industrial production process, perform coreference resolution and entity disambiguation on the data contained in the knowledge sources, and obtain the second knowledge graph.

[0058] It should be noted that the knowledge sources in the industrial production process can be equipment operation data, equipment operation reports, Internet corpus data, etc. generated during the industrial production process.

[0059] It should be understood that, to avoid repetition, people habitually use pronouns, appellations, and abbreviations to refer to the full name of previously mentioned entities. For example, an article might begin with "Harbin Institute of Technology," and later might use "HIT," "Gongda," etc., and might also mention "this university," "she," etc. This phenomenon is called coreference. Therefore, coreference resolution can be the process of merging different descriptions of the same knowledge entity from the knowledge source. The essence of entity disambiguation is that a word may have multiple meanings, that is, its meaning may differ in different contexts. Therefore, entity disambiguation can be used to determine whether two entities refer to the same entity.

[0060] Furthermore, in order to improve the efficiency of coreference resolution of knowledge sources in industrial production processes, the step of performing coreference resolution on the data contained in the knowledge source includes:

[0061] Obtain the association mapping model between entities and alias entities;

[0062] The coreference resolution of the data contained in the knowledge source is performed based on the entity information in the knowledge source and the association mapping model.

[0063] It should be noted that the association mapping model may include the mapping relationship between entity names and their corresponding alias entities. For example, multiple alias entities may point to the same entity. The coreference resolution of the data contained in the knowledge source based on the entity information in the knowledge source and the association mapping model may involve determining whether the entity name corresponding to the entity information is an alias entity of a certain entity based on the entity information in the knowledge source; if so, coreference resolution is performed on the entities in the entity information based on the entity corresponding to the alias entity. The construction of the association mapping model between entities and alias entities may involve determining the target entity in the preset standard data and the entity attributes corresponding to the target entity.

[0064] A mapping model between the target entity and its alias entities is constructed based on the target entity and its attributes. Specifically, this can involve analyzing the entity attributes of each target entity in the preset standard data, clustering entity information pointing to the same entity based on these attributes, and obtaining multiple alias entities with different names pointing to the same entity. Each target entity in the preset standard data can be an entity contained within the preset standard data itself.

[0065] Step S30: Generate a target knowledge graph based on the first knowledge graph and the second knowledge graph.

[0066] It should be noted that generating the target knowledge graph based on the first knowledge graph and the second knowledge graph can be achieved by integrating the data in the first knowledge graph and the second knowledge graph, and performing coreference resolution and entity disambiguation on the integrated data to obtain the target knowledge graph.

[0067] This embodiment constructs a first knowledge graph based on a relational pattern model using preset standard data. The preset standard data includes at least one of the following: empirical rules from actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams. Knowledge sources within the industrial production process are identified, and the data contained in these knowledge sources undergo coreference resolution and entity disambiguation to obtain a second knowledge graph. A target knowledge graph is then generated based on the first and second knowledge graphs. Because this embodiment generates the target knowledge graph from the first knowledge graph constructed using preset standard data and the second knowledge graph constructed from knowledge sources within the industrial production process, it can efficiently integrate the preset standard data and knowledge sources within the industrial production process, achieving the integration and aggregation of industrial data.

[0068] refer to Figure 3 , Figure 3 This is a flowchart illustrating the second embodiment of the industrial knowledge graph generation method of the present invention.

[0069] Based on the first embodiment described above, in this embodiment, step S20 includes:

[0070] Step S201: Determine the knowledge sources in the industrial production process and determine the data categories of the knowledge sources.

[0071] It should be noted that the data categories may include structured data, semi-structured data, and / or unstructured data.

[0072] Step S202: Extract and identify data from the knowledge source according to the data category to generate a case map.

[0073] It should be noted that the step of extracting and identifying data from the knowledge source based on the data category to generate a case graph can involve extracting and identifying data with different structures from the knowledge source using different methods to generate the case graph. Specifically, when the data category is structured data, data extraction can be performed on the data contained in the knowledge source, and triples can be constructed based on the extracted data; a case graph can then be generated based on the triples. When the data category is unstructured data, data extraction can be performed on the data contained in the knowledge source based on an association pattern model, and a case graph can then be generated based on the extracted data.

[0074] It should be noted that the unstructured data may include semi-structured data and / or unstructured data. When the data category is structured data, data extraction is performed on the data contained in the knowledge source, and triples are constructed based on the extracted data. Generating a case graph based on the triples can be achieved when the data category of the data contained in the data source is structured data. This involves extracting data such as the actual data source, verifiable scientific questions, and transferable application scenarios corresponding to the structured data, constructing triples in the format of "entity-attribute-attribute value" based on the extracted data, and abstracting the triples and corresponding data into a case graph using a knowledge graph.

[0075] When the data category is unstructured data, data extraction is performed on the data contained in the knowledge source based on the association pattern model, and a case graph is generated based on the extracted data. This can be done when the data category of the data contained in the data source is unstructured or semi-structured data, by identifying and extracting data from the unstructured or semi-structured data using named entity recognition, and constructing relevant relationships on the extracted data based on the association rules and association templates in the association pattern model, thereby generating a case graph.

[0076] In practical implementation, for sample cases of structured data (e.g., binary tables), knowledge entity attributes are constructed based on their actual data sources, verifiable scientific questions, and transferable application scenarios to complete the construction of "entity-attribute-attribute value (sample individual)" triples. The sample cases are then abstracted into a case graph in the form of a knowledge graph. For semi-structured and unstructured data sample cases, rule-based, statistical, machine learning, and deep learning named entity recognition methods are used to identify and extract industrial knowledge entities. Relationship rules or templates are constructed based on expert experience and association pattern analysis to realize the construction of various different relationship patterns, mine the correlation relationships of multi-source heterogeneous knowledge, and generate a case graph.

[0077] Step S203: Perform coreference resolution and entity disambiguation on the data in the case graph to obtain the second knowledge graph.

[0078] It should be noted that the coreference resolution of the data in the case graph can be achieved by calculating the attribute similarity between two entities in the case graph. When the attribute similarity of the entities is greater than a first preset similarity threshold, it is determined that the two entities point to the same entity, thus achieving coreference resolution. The entity disambiguation of the data in the case graph can be achieved by calculating the similarity of the corresponding attributes of the entities based on the semantic analysis model and the spatial vector model when the entity names of two entities in the case graph are the same or the similarity is higher than the first preset similarity threshold. When the attribute similarity is less than a second preset similarity threshold, it is determined that the two entities with the same entity name are different entities, thus achieving entity disambiguation. A second knowledge graph is obtained from the case graph after coreference resolution and entity disambiguation. The first and second preset thresholds mentioned above can both be thresholds preset according to the actual usage scenario, and this embodiment does not impose any limitations on them.

[0079] Furthermore, in order to improve the integration efficiency of industrial data, step S203 may include: extracting first entity information and second entity information from the case graph using a preset relationship extraction algorithm cluster;

[0080] Determine the entity similarity and entity attribute similarity between the first entity information and the second entity information;

[0081] The case graph is subjected to coreference resolution based on the entity similarity and the entity attribute similarity to obtain the target case graph;

[0082] Based on the first entity information and the second entity information, entity disambiguation is performed on the target case graph to obtain a second knowledge graph.

[0083] It should be noted that the purpose of relation extraction is to extract the relationship between two knowledge entities from data information, thereby realizing semantic connections between the two entities. The proposed pre-defined relation extraction algorithm cluster is constructed from the perspectives of expert experience, machine learning, and deep learning. Expert experience involves manually constructing relationships from existing experiential knowledge, such as hierarchical relationships or valve start / stop order relationships. Machine learning algorithms transform relation extraction into clustering and classification problems, such as methods based on feature knowledge. Deep learning methods learn relationships from a large amount of historical data to form an evaluative network method that can quantify the relationships between entities. This embodiment constructs the pre-defined relation extraction algorithm cluster from the above perspectives. It is used to extract the relationship between two knowledge entities, namely, the first entity information and the second entity information, from a case graph.

[0084] It should be noted that determining the entity similarity and entity attribute similarity between the first entity information and the second entity information can be done by calculating the entity similarity and entity attribute similarity between the first entity information and the second entity information based on a constructed similarity / relevance / association algorithm cluster. The entity similarity can be the similarity of entity vectors in the first entity information and the second entity information. The entity attribute similarity can include the similarity of attribute vectors of the first and last entities and attribute vectors of the first and last relationships between the entities in the first entity information and the second entity information. The similarity / relevance / association algorithm cluster can be used to calculate the similarity of concepts, relationships, and attributes of knowledge entities, using metrics such as Euclidean distance, probabilistic distance, divergence distance, and correlation to quantify the similarity between two entities with different names but the same content.

[0085] The process of dissolving coreference in the case graph based on entity similarity and entity attribute similarity to obtain the target case graph can be achieved by constructing standardized knowledge entities using an entity standardization evaluation mechanism. This involves using standardized actual physical entities as the core (entity similarity) and measuring multi-granularity information (entity attribute similarity) of uncertain entities, attributes, and relationships to form a series of evidence chains. Evidence weight factors are assigned based on historical information to achieve coreference dissolution in the purification process, thus obtaining the target case graph after coreference dissolution. Assigning evidence weight factors based on historical information can be achieved by constructing credibility allocation factors. These credibility allocation factors can be constructed based on the actual physical meaning reflected by the entities and the information coverage of the entity knowledge, etc., to build credibility allocation factor evaluation indicators. Different entity characteristics and different similarity measurement algorithms are used to measure completely different system operating states, and weight factors are assigned to the different pieces of evidence measured.

[0086] It should be understood that the construction of an entity specification evaluation mechanism can be based on industry standards and domain knowledge, and from the perspectives of name standardization, actual physical meaning and interpretability, a knowledge entity specification evaluation mechanism can be constructed to establish the simplest set of knowledge entities.

[0087] Information in industrial production processes is multi-source and heterogeneous, leading to numerous issues such as representational ambiguity, quantification ambiguity, and other instances of names being different from each other after extraction. Therefore, entity disambiguation is necessary for the data in the data sources.

[0088] The step of performing entity disambiguation on the target case graph based on the first entity information and the second entity information to obtain the second knowledge graph can be achieved by calculating the entity similarity and entity attribute similarity between the first entity information and the second entity information, and determining whether the first entity information and the second entity information point to different entities based on the entity similarity and entity attribute similarity, thereby achieving entity disambiguation of the target case graph. For example, if the entity similarity between the first entity information and the second entity information is greater than a first preset threshold, but the entity attribute similarity is less than a second preset threshold, it is determined that the first entity information and the second entity information point to different entities, thus achieving entity disambiguation of the target case graph. Specifically, the principle can be to measure the similarity, relevance, and association between knowledge entity objects and physical references, attribute items, etc., and to construct isomorphic spaces of different granularities and modalities using methods such as spatial vector models and semantic space models, thereby achieving entity disambiguation of knowledge in the purification process.

[0089] Furthermore, in order to achieve entity disambiguation of industrial data, the step of performing entity disambiguation on the target case graph based on the first entity information and the second entity information to obtain a second knowledge graph includes:

[0090] The first entity information is isomorphized according to the spatial vector model and the semantic space model to obtain the first isomorphic space;

[0091] The second entity information is isomorphized according to the spatial vector model and the semantic space model to obtain the second isomorphic space;

[0092] Based on the first isomorphic space and the second isomorphic space, entity disambiguation is performed on the target case graph to obtain the second knowledge graph.

[0093] It should be noted that the isomorphism of the first entity information based on the spatial vector model and the semantic space model to obtain the first isomorphic space can be based on the relationship between knowledge entity objects and physical references and attribute items in the first entity information, using the spatial vector model and the semantic space model to construct isomorphic spaces of different granularities. Similarly, the isomorphism of the second entity information based on the spatial vector model and the semantic space model yields the second isomorphic space.

[0094] The step of performing entity disambiguation on the target case graph based on the first isomorphic space and the second isomorphic space to obtain the second knowledge graph can be achieved by representing the simplified purification process knowledge as a hypergraph based on the first isomorphic space and the second isomorphic space, where nodes represent the simplified entity classes and edges represent the simplified relationship classes between entity nodes, thereby constructing a mid-level high-dimensional purification process knowledge space, and performing entity disambiguation on homonymous entities in the first entity information and the second entity information based on the mid-level high-dimensional purification process knowledge space.

[0095] In specific implementation, it can be referred to Figure 4 , Figure 4 This is a schematic diagram of the target knowledge graph generation framework of the second embodiment of the industrial knowledge graph generation method of the present invention, with reference to... Figure 4 , Figure 4 The system constructs standardized knowledge entities, i.e., the first knowledge graph, from pre-defined standard data through an entity standardization evaluation mechanism (including an association pattern model). For duplicate and inconsistent entities with the same name in the knowledge source, coreference resolution is performed based on entity and attribute similarity measures, historical data, and credibility factor calculations. For conflicting and ambiguous entities with the same name in the knowledge source, entity disambiguation is performed based on entity and attribute similarity measures, spatial vector models, and semantic space models. Knowledge is solidified based on the standardized knowledge entities constructed from the entity-disambiguated data, the coreference-resolved data, and the pre-defined standard data, generating the target knowledge graph. During knowledge solidification using the standardized knowledge entities constructed from the entity-disambiguated data, the coreference-resolved data, and the pre-defined standard data, entity disambiguation or coreference resolution can also be performed on homonymous or heteronymous data within the above three types of data.

[0096] This embodiment identifies knowledge sources in the industrial production process and determines the data categories of these knowledge sources. Based on the data categories, it extracts and identifies data from the knowledge sources to generate a case graph. It then performs coreference resolution and entity disambiguation on the data in the case graph to obtain a second knowledge graph. This embodiment uses a preset relation extraction algorithm cluster to extract first and second entity information from the case graph, and then performs coreference resolution and entity disambiguation on the extracted information. This solves the problem of knowledge extraction and modeling from multi-source heterogeneous information, providing theoretical and technical support for ensuring the stable and safe operation of intelligent manufacturing and improving product quality.

[0097] Reference Figure 5 , Figure 5 This is a structural block diagram of the first embodiment of the knowledge graph generation device of the present invention.

[0098] like Figure 5 As shown, the knowledge graph generation device proposed in this embodiment of the invention includes:

[0099] Module 10 is used to construct a first knowledge graph based on a relational pattern model using preset standard data. The preset standard data includes at least one of the following: empirical rules in actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams.

[0100] Data processing module 20 is used to determine the knowledge source in the industrial production process, perform coreference resolution and entity disambiguation on the data contained in the knowledge source, and obtain a second knowledge graph;

[0101] The generation module 30 is used to generate a target knowledge graph based on the first knowledge graph and the second knowledge graph.

[0102] This embodiment constructs a first knowledge graph based on a relational pattern model using preset standard data. The preset standard data includes at least one of the following: empirical rules from actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams. Knowledge sources within the industrial production process are identified, and the data contained in these knowledge sources undergo coreference resolution and entity disambiguation to obtain a second knowledge graph. A target knowledge graph is then generated based on the first and second knowledge graphs. Because this embodiment generates the target knowledge graph from the first knowledge graph constructed using preset standard data and the second knowledge graph constructed from knowledge sources within the industrial production process, it can efficiently integrate the preset standard data and knowledge sources within the industrial production process, achieving the integration and aggregation of industrial data.

[0103] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0104] In addition, for technical details not described in detail in this embodiment, please refer to the industrial knowledge graph generation method provided in any embodiment of the present invention, which will not be repeated here.

[0105] Based on the first embodiment of the knowledge graph generation device of the present invention described above, a second embodiment of the knowledge graph generation device of the present invention is proposed.

[0106] In this embodiment, the data processing module 20 is further configured to determine the knowledge source in the industrial production process and determine the data category of the knowledge source;

[0107] Based on the data categories, the knowledge sources are extracted and identified to generate a case graph;

[0108] The data in the case graph are subjected to coreference resolution and entity disambiguation to obtain a second knowledge graph.

[0109] Furthermore, the data processing module 20 is also used to extract first entity information and second entity information from the case graph using a preset relationship extraction algorithm cluster;

[0110] Determine the entity similarity and entity attribute similarity between the first entity information and the second entity information;

[0111] The case graph is subjected to coreference resolution based on the entity similarity and the entity attribute similarity to obtain the target case graph;

[0112] Based on the first entity information and the second entity information, entity disambiguation is performed on the target case graph to obtain a second knowledge graph.

[0113] Furthermore, the data processing module 20 is also used to isomorphize the first entity information according to the spatial vector model and the semantic space model to obtain a first isomorphic space;

[0114] The second entity information is isomorphized according to the spatial vector model and the semantic space model to obtain the second isomorphic space;

[0115] Based on the first isomorphic space and the second isomorphic space, entity disambiguation is performed on the target case graph to obtain the second knowledge graph.

[0116] Furthermore, the data processing module 20 is also used to extract data from the data contained in the knowledge source when the data category is structured data, and to construct triples based on the extracted data;

[0117] A case graph is generated based on the triplet;

[0118] When the data category is unstructured data, data extraction is performed on the data contained in the knowledge source based on the association pattern model, and a case graph is generated based on the extracted data.

[0119] Furthermore, the data processing module 20 is also used to obtain the association mapping model between entities and alias entities;

[0120] The coreference resolution of the data contained in the knowledge source is performed based on the entity information in the knowledge source and the association mapping model.

[0121] Furthermore, the data processing module 20 is also used to determine the target entity in the preset standard data and the entity attribute corresponding to the target entity;

[0122] Construct an association mapping model between the entity and its alias based on the target entity and the entity's attributes.

[0123] Other embodiments or specific implementations of the knowledge graph generation device of the present invention can be found in the above-described method embodiments, and will not be repeated here.

[0124] Furthermore, this embodiment of the invention also proposes a storage medium storing a knowledge graph generation program, which, when executed by a processor, implements the steps of the industrial knowledge graph generation method described above.

[0125] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0126] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0127] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0128] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for generating an industrial knowledge graph, characterized in that, The knowledge graph generation method includes the following steps: Based on the association pattern model, a first knowledge graph is constructed from preset standard data, which includes at least one of the following: empirical rules in actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams. The knowledge sources in the industrial production process are identified, and the data contained in the knowledge sources are subjected to coreference resolution and entity disambiguation to obtain a second knowledge graph. Generate a target knowledge graph based on the first knowledge graph and the second knowledge graph; The process of determining the knowledge sources in the industrial production process, performing coreference resolution and entity disambiguation on the data contained in the knowledge sources to obtain a second knowledge graph includes: Identify the knowledge sources in the industrial production process and determine the data categories of the knowledge sources; Based on the data categories, the knowledge sources are extracted and identified to generate a case graph; The data in the case graph are subjected to coreference resolution and entity disambiguation to obtain a second knowledge graph; The step of extracting and identifying data from the knowledge source based on the data category to generate a case graph includes: When the data category is structured data, data extraction is performed on the data contained in the knowledge source, triples are constructed based on the extracted data, and a case graph is generated based on the triples. When the data category is unstructured or semi-structured data, data in the unstructured or semi-structured data is identified and extracted according to the named entity recognition method, and the relevant relationships of the extracted data are constructed according to the association rules and association templates in the association pattern model to generate a case graph.

2. The industrial knowledge graph generation method as described in claim 1, characterized in that, The step of performing coreference resolution and entity disambiguation on the data in the case graph to obtain the second knowledge graph includes: The first entity information and the second entity information are extracted from the case graph using a pre-defined relationship extraction algorithm cluster. Determine the entity similarity and entity attribute similarity between the first entity information and the second entity information; The case graph is subjected to coreference resolution based on the entity similarity and the entity attribute similarity to obtain the target case graph; Based on the first entity information and the second entity information, entity disambiguation is performed on the target case graph to obtain a second knowledge graph.

3. The industrial knowledge graph generation method as described in claim 2, characterized in that, The step of performing entity disambiguation on the target case graph based on the first entity information and the second entity information to obtain a second knowledge graph includes: The first entity information is isomorphized according to the spatial vector model and the semantic space model to obtain the first isomorphic space; The second entity information is isomorphized according to the spatial vector model and the semantic space model to obtain the second isomorphic space; Based on the first isomorphic space and the second isomorphic space, entity disambiguation is performed on the target case graph to obtain the second knowledge graph.

4. The industrial knowledge graph generation method as described in claim 1, characterized in that, The step of extracting and identifying data from the knowledge source based on the data category to generate a case graph includes: When the data category is structured data, data extraction is performed on the data contained in the knowledge source, and triples are constructed based on the extracted data; A case graph is generated based on the triplet; When the data category is unstructured data, data extraction is performed on the data contained in the knowledge source based on the association pattern model, and a case graph is generated based on the extracted data.

5. The industrial knowledge graph generation method as described in claim 1, characterized in that, The step of performing coreference resolution on the data contained in the knowledge source includes: Obtain the association mapping model between entities and alias entities; The coreference resolution of the data contained in the knowledge source is performed based on the entity information in the knowledge source and the association mapping model.

6. The industrial knowledge graph generation method as described in claim 5, characterized in that, After the step of constructing the first knowledge graph from the preset standard data based on the association pattern model, the method further includes: Determine the target entity and the entity attribute corresponding to the target entity in the preset standard data; Construct an association mapping model between the entity and its alias based on the target entity and the entity's attributes.

7. An industrial knowledge graph generation device, characterized in that, The knowledge graph generation device includes: A construction module is used to construct a first knowledge graph based on a relational pattern model using preset standard data. The preset standard data includes at least one of the following: empirical rules in actual industrial production processes, physicochemical mechanism models, artificial intelligence algorithm models, and process configuration diagrams. The data processing module is used to determine the knowledge sources in the industrial production process, and to perform coreference resolution and entity disambiguation on the data contained in the knowledge sources to obtain a second knowledge graph. The generation module is used to generate a target knowledge graph based on the first knowledge graph and the second knowledge graph; The process of determining the knowledge sources in the industrial production process, performing coreference resolution and entity disambiguation on the data contained in the knowledge sources to obtain a second knowledge graph includes: Identify the knowledge sources in the industrial production process and determine the data categories of the knowledge sources; Based on the data categories, the knowledge sources are extracted and identified to generate a case graph; The data in the case graph are subjected to coreference resolution and entity disambiguation to obtain a second knowledge graph; The step of extracting and identifying data from the knowledge source based on the data category to generate a case graph includes: When the data category is structured data, data extraction is performed on the data contained in the knowledge source, triples are constructed based on the extracted data, and a case graph is generated based on the triples. When the data category is unstructured or semi-structured data, data in the unstructured or semi-structured data is identified and extracted according to the named entity recognition method, and the relevant relationships of the extracted data are constructed according to the association rules and association templates in the association pattern model to generate a case graph.

8. An industrial knowledge graph generation device, characterized in that, The device includes: a memory, a processor, and a knowledge graph generation program stored in the memory and executable on the processor, the knowledge graph generation program being configured to implement the steps of the industrial knowledge graph generation method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that, The storage medium stores a knowledge graph generation program, which, when executed by a processor, implements the steps of the industrial knowledge graph generation method as described in any one of claims 1 to 6.

Images