Process information recommendation method, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention discloses a process information recommendation method, electronic equipment, a storage medium and a program product, and the method comprises the steps: constructing an associated reserve database of a manufacturing process object; obtaining an operation behavior characteristic of querying the associated reserve database by a first target operator; adjusting and updating the recommendation weight of the process object stored in the associated reserve database according to the operation behavior characteristics of the associated reserve database; and determining a matched target process association recommendation result for process query information input by the second target operator according to the adjusted and updated association reserve database. According to the technical scheme provided by the embodiment of the invention, the information recommendation accuracy can be continuously improved based on the interaction data between the user and the process information recommendation platform.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of intelligent recommendation technology, and in particular, to a process information recommendation method, device, electronic device, storage medium and program product. Background Art

[0002] For enterprises, manufacturing products are not completely independent. Even if they are updated and iterated, the structural design and CNC technology of new and old products still have certain similarities and inheritance. Based on this, various process recommendation methods have emerged.

[0003] The most common process recommendation algorithm at present is the nearest neighbor algorithm, which is mainly based on analogical reasoning based on old examples or experience. When making process recommendations, first search for the most similar instance in the instance library, and then choose to directly adopt the instance result or modify the result before adopting it according to certain rules, save the instance, and realize dynamic learning of the instance library. This nearest neighbor algorithm calculates the similarity between the current instance and the instance in the instance library according to certain rules, and then determines the retrieval result according to the similarity. In addition, some technologies build a knowledge graph in the manufacturing field to realize the recommendation of process information through the knowledge graph in the manufacturing field.

[0004] In the process of realizing the present invention, the inventors found that the prior art has the following defects: the process recommendation algorithm based on the nearest neighbor algorithm cannot reasonably evaluate the difference distance between different instances, and as the instance library expands, its retrieval speed and use efficiency will also be significantly reduced. The current method of recommending process information through the manufacturing field knowledge graph does not consider the impact of user retrieval behavior and user experience on the knowledge graph during the use of the knowledge graph, that is, the current manufacturing field knowledge graph system is a static system, which is not conducive to the later user experience and has a low accuracy in process information recommendation. Summary of the invention

[0005] The embodiments of the present invention provide a process information recommendation method, device, electronic device, storage medium and program product, which can continuously improve the accuracy of information recommendation based on the interaction data between users and a process information recommendation platform.

[0006] According to one aspect of the present invention, a process information recommendation method is provided, comprising:

[0007] Build an associated reserve database of manufacturing process objects;

[0008] Acquire the operation behavior characteristics of the first target operator querying the associated reserve database;

[0009] Adjusting and updating the recommended weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database;

[0010] The process query information input by the second target operator is determined according to the adjusted and updated associated reserve database to match the target process associated recommendation result.

[0011] According to another aspect of the present invention, there is provided a process information recommendation device, comprising:

[0012] A related reserve database construction module, used to construct a related reserve database for manufacturing process objects;

[0013] An operation behavior feature acquisition module, used to acquire the operation behavior feature of the first target operator querying the associated reserve database;

[0014] A recommended weight adjustment and update module, used to adjust and update the recommended weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database;

[0015] The target process association recommendation result determination module is used to determine the matching target process association recommendation result for the process query information input by the second target operator according to the adjusted and updated association reserve database.

[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0017] at least one processor; and

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the process information recommendation method described in any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the process information recommendation method described in any embodiment of the present invention when executed.

[0021] According to another aspect of the present invention, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the process information recommendation method according to any embodiment of the present invention is implemented.

[0022] The embodiment of the present invention constructs an associated reserve database for manufacturing process objects and obtains the operational behavior characteristics of the first target operator when querying the associated reserve database, thereby adjusting and updating the recommendation weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database, and then determines the matching target process associated recommendation results for the process query information input by the second target operator according to the adjusted and updated associated reserve database, thereby solving the problem of low recommendation accuracy in the existing process information recommendation methods and being able to continuously improve the accuracy of information recommendation based on the interaction data between users and the process information recommendation platform.

[0023] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 is a flow chart of a process information recommendation method provided in Embodiment 1 of the present invention;

[0026] Figure 2 It is a schematic diagram of a process for recommending part process information based on a part process knowledge adaptive optimization and application verification platform provided by the first embodiment of the present invention;

[0027] Figure 3 is a flow chart of a process information recommendation method provided by Embodiment 2 of the present invention;

[0028] Figure 4 It is a flowchart for constructing a knowledge graph based on manufacturing process provided by the second embodiment of the present invention;

[0029] Figure 5 This is an overall flow chart for constructing a knowledge graph based on manufacturing processes provided in the second embodiment of the present invention;

[0030] Figure 6 is a flow chart of a process information recommendation method provided by Embodiment 3 of the present invention;

[0031] Figure 7 It is a schematic diagram of a part process FO master construction process provided by the third embodiment of the present invention;

[0032] Figure 8is a schematic diagram of a process information recommendation device provided by a fourth embodiment of the present invention;

[0033] Fig. 9 This is a schematic diagram of the structure of an electronic device provided in Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0034] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0035] It should be noted that the terms "first", "second" and "target" in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0036] Embodiment 1

[0037] Figure 1 This is a flowchart of a method for recommending process information provided by the first embodiment of the present invention. This embodiment is applicable to the case where the recommended weights of process objects stored in the associated reserve database are updated by using the operating behavior characteristics of operators to accurately recommend process information. The method can be executed by a process information recommendation device, which can be implemented by software and / or hardware and can generally be integrated in an electronic device. The electronic device can be a terminal device or a server device. As long as it can execute the process information recommendation method, the embodiment of the present invention does not limit the specific device type of the electronic device. Accordingly, Figure 1 As shown, the method includes the following operations:

[0038] S110, constructing an associated reserve database of manufacturing process objects.

[0039] Among them, the manufacturing process object can be a process object involved in any manufacturing field, for example, it can be a process object involved in various manufacturing fields such as smart grid, automobile manufacturing and aircraft manufacturing. The process object can include but is not limited to abstract entity objects such as parts, molds and processes, and can also include text objects such as outline templates for guiding process flows. As long as it is any process object involved in the manufacturing field, it can be used as a manufacturing process object. The embodiment of the present invention does not limit the specific object type and content of the manufacturing process object. The associated reserve database can be a database for storing process data associated with manufacturing process objects.

[0040] In an embodiment of the present invention, before using the process information recommendation platform to provide intelligent process information recommendation functions to operators, an associated reserve database of manufacturing process objects can be first constructed. Optionally, the associated reserve database can be an integrated knowledge base corresponding to the specific manufacturing field of the manufacturing process object, and the integrated knowledge base can be in the form of a knowledge graph. Alternatively, the associated reserve database can also be a master template library professionally constructed for the manufacturing process object, and the master template library can store the master templates of the manufacturing process object, and each master template can reflect the information of the overall manufacturing process flow of the corresponding manufacturing process object.

[0041] S120: Acquire the operation behavior characteristics of the first target operator inquiring the associated reserve database.

[0042] The first target operator may be an operator who uses the associated reserve database to query relevant process information in the historical data, and may generally be a technician. The operation behavior feature may be used to reflect the behavior feature of the first target operator using the associated reserve database to retrieve process knowledge.

[0043] Exemplarily, the operation behavior of the first target operator to query the associated reserve database may include but is not limited to the behavior of adding, modifying, deleting and searching for process-related data. Exemplarily, the process-related data may include but is not limited to the description basic data of the process object, such as the data recorded in the part manufacturing outline, and the master data corresponding to the process object.

[0044] It is understandable that different types of manufacturing industries have different corresponding process manufacturing links. Taking the aircraft manufacturing industry as an example, the manufacturing of large civil aircraft parts can be divided into specific process areas such as machining, sheet metal, composite materials and assembly. The manufacturing outline (FO, Fabrication Outline) is a process document that guides and controls the manufacturing and rework, repair or assembly of parts by the manufacturing / assembly process department according to the input conditions such as engineering drawings, process standards, authorization documents and related engineering documents for the manufacturing and rework, repair or assembly of parts. Its main purpose is to provide the required information for part manufacturing, repair or rework, provide part transfer control, and serve as a record of manufacturing and inspection.

[0045] In a specific example, taking process parts as an example, the part production outline can be used as the basic data source to build an associated reserve database corresponding to the process parts, and the associated reserve database can also include the content of the FO master. Correspondingly, the first target operator, as a learner, can query and retrieve relevant knowledge of the part process based on the process information recommendation platform and the associated reserve database, such as querying the process content of a process part, or querying the FO master of a process part. Correspondingly, the first target operator can generate four types of behavior data in the process information recommendation platform, as shown in Table 1.

[0046] Table 1 Operation behavior data table

[0047]

[0048] During the operation of the process information recommendation platform by the first target operator, the first target operator can download or upload data through various interfaces shown in Table 2, and perform different log records according to different purposes. Correspondingly, for the interface for downloading data, the relevant number of the downloaded data can be recorded; for the interface for uploading data, the final version and specific time of the uploaded data can be recorded.

[0049] Table 2 Part process data platform interface example table

[0050]

[0051] S130 , adjusting and updating the recommended weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database.

[0052] The storage process object is the related process object stored in the associated reserve database, which may include but is not limited to the various processes involved in the process or the FO master template constructed according to the relevant process outline. The recommendation weight may be a weight value configured for each storage process object in the associated reserve database, which may represent the recommendation degree of each storage process object. It is understandable that the larger the value of the recommendation weight, the greater the recommendation strength of the corresponding storage process object.

[0053] It is understandable that the operational behavior characteristics of the first target operator contain important information and are therefore worth exploring in depth. Accordingly, after obtaining the operational behavior characteristics of the first target operator querying the associated reserve database, the operational behavior characteristics can be statistically analyzed, and the recommended weights of the corresponding storage process objects can be adjusted and updated based on the statistical analysis results of the operational behavior characteristics, so that the recommended weights of the storage process objects with higher query and retrieval frequencies are also dynamically increased.

[0054] It can be seen that the above technical scheme can form process recommendation information that conforms to the operator's usage habits through reinforcement learning of various operator feedback behaviors such as clicks, selections and preferences during the use of the associated reserve database by the first target operator, and form a user-friendly "human-in-the-loop" intelligent interactive framework for the process information recommendation platform, thereby improving the practicality and accuracy of process knowledge recommendation push.

[0055] S140. Determine a matching target process association recommendation result for the process query information input by the second target operator according to the adjusted and updated association reserve database.

[0056] Among them, the second target operator may be an operator who currently uses the associated reserve database to query and retrieve process information. The process query information, that is, the query information currently input by the second target operator to the process information recommendation platform, can be used to query process-related information. Exemplarily, the process query information may include but is not limited to the process description content or the module description content of the FO master copy, etc. The target process association recommendation result may be the recommendation information provided by the process information recommendation platform to the second target operator based on the process query information, and the recommendation information may meet the query needs of the second target operator for process query information.

[0057] In an embodiment of the present invention, after completing the adjustment and update operation of the recommended weights of the process objects stored in the associated reserve database, the process information recommendation platform can adjust the updated associated reserve database according to the recommended weights, and determine the matching target process associated recommendation results for the process query information input by the second target operator.

[0058] After the process information recommendation platform detects the process query information input by the second target operator, it can calculate the similarity between each stored process object with dynamically adjusted recommendation weight and the process query information, and recommend the corresponding target process-related recommendation results to the second target operator in order of similarity from large to small.

[0059] It is understandable that enterprises will store a large number of manufacturing outline resources on the data platform related to manufacturing process knowledge. For newly joined technicians, there will be resource overload and difficulty in selection, which cannot meet the personalized and adaptive learning needs of new technicians. In order to solve the problem of "consistency" of new technicians' training and learning based on process knowledge, the embodiment of the present invention analyzes the various process deletions, process sequence arrangements, and related precautions and notes of technicians, establishes a personalized learner model as a premise, and constructs an associated reserve database of manufacturing process objects as data support. The learner model is combined with the associated reserve database of manufacturing process objects to realize the intelligent recommendation function of process information. On the basis of using the associated reserve database of manufacturing process objects to express the manufacturing process knowledge in a structured manner, a method for recommending process knowledge learning paths that are adaptively adjusted is given to technicians in combination with their individual operation behavior characteristics, which is conducive to improving the training efficiency of technicians on process knowledge.

[0060] In the above scheme, the process of adjusting and updating the recommended weights of process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database belongs to the adaptive feedback process instruction optimization process in which the process information recommendation platform analyzes and integrates the data obtained in the previous order.

[0061] At present, in order to solve the lack of effective knowledge recommendation and integration in the process design and manufacturing of manufacturing process objects, especially long-cycle and long-process process objects (such as parts, etc.), process personnel rely on personal experience to carry out a lot of repetitive design work, which easily leads to the problem of low efficiency of the above work. At the same time, due to the difficulty in extracting effective knowledge in the process design process, low knowledge reuse rate and sharing ability, and poor supply chain collaborative design and manufacturing capabilities, there is a lack of a full-featured platform for system design and knowledge sharing.

[0062] Figure 2 is a flow chart of recommending part process information based on a part process knowledge adaptive optimization and application verification platform provided by the first embodiment of the present invention. In a specific example, Figure 2As shown, taking the part process as an example, the associated reserve database may include but is not limited to the part process knowledge base and the outline master template library, wherein the part process knowledge base is an integrated knowledge base. The part process knowledge adaptive optimization and application verification platform can first construct the part process knowledge base and the outline master template library based on multi-source data, and the outline master template library can be used as a partial database in the part process knowledge base ( Figure 2 At the same time, the part process knowledge adaptive optimization and application verification platform can obtain the various operational behavior characteristics of the technician using the part process knowledge base to query or build the FO master template, and then adjust the relevant weights of the FO master template library and the part process knowledge base based on the technician's behavior characteristics, and finally dynamically adjust the FO master template recommendation order or the related process recommendation order, etc., to form an adaptive part process knowledge recommendation closed loop. Through the research on the knowledge-driven part process knowledge adaptive optimization retrieval method and device, it can help to build a process design and model management integration platform for network collaboration, break through the core technology of digital collaboration jointly developed by the supply chain, and continuously improve the recommendation accuracy in the process of users using the information recommendation platform, and promote the transformation of the new generation of aircraft to a digital design and manufacturing system model.

[0063] The embodiment of the present invention constructs an associated reserve database for manufacturing process objects and obtains the operational behavior characteristics of the first target operator when querying the associated reserve database, thereby adjusting and updating the recommendation weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database, and then determines the matching target process associated recommendation results for the process query information input by the second target operator according to the adjusted and updated associated reserve database, thereby solving the problem of low recommendation accuracy in the existing process information recommendation methods and being able to continuously improve the accuracy of information recommendation based on the interaction data between users and the process information recommendation platform.

[0064] Embodiment 2

[0065] Figure 3 is a flowchart of a method for recommending process information provided by the second embodiment of the present invention. This embodiment is specific based on the above embodiment. In this embodiment, the integrated knowledge base is used as an associated reserve database, and the associated reserve database for constructing manufacturing process objects is provided, and the recommended weights of the process objects stored in the associated reserve database are adjusted and updated, and the process query information input by the second target operator is determined according to the adjusted and updated associated reserve database. Correspondingly, Figure 3 As shown, the method of this embodiment may include:

[0066] S210 , acquiring multi-source heterogeneous data provided by each associated data platform of the manufacturing process object, and formatting the multi-source heterogeneous data to obtain multi-tuple structure data.

[0067] Optionally, the multi-tuple structure data may be structure data of a type such as a triplet or a quintuple. The embodiment of the present invention does not limit the structure type of the multi-tuple structure data.

[0068] In an embodiment of the present invention, optionally, the associated reserve database of the manufacturing process object can be an integrated knowledge base, and the integrated knowledge base can be a knowledge graph database in the field of process manufacturing. The knowledge graph database in the field of process manufacturing is used to express the attributes of data and knowledge in the field of process manufacturing and the inherent relationship between data and knowledge. Compared with the general field knowledge graph, the intelligent manufacturing knowledge graph has not changed much in the basic architecture, but still has four different characteristics: the stored manufacturing field data and knowledge include product design process, process, processing process and other product life cycle data and knowledge; the application scenarios and problems in the manufacturing field mainly focus on the efficiency and quality in the product design and production process; high interpretability and accuracy requirements: interpretability requirements are convenient for users such as technicians to verify and understand their reasoning results, and its accuracy affects the efficiency and quality of each manufacturing task of the enterprise; few resources and strong specificity: it is difficult to obtain a large amount of data and knowledge from the Internet, and it is mainly based on the characteristics of the manufacturing enterprise and the data and knowledge accumulated by the enterprise itself to build a unique knowledge graph.

[0069] The research on process manufacturing knowledge graph first needs to realize the formal expression of knowledge in this field, that is, to realize the knowledge expression method that can be understood by machines under the requirements of multimodal data. The key technical means involved in the knowledge expression in the process manufacturing field include predicate logic knowledge representation method, framework knowledge representation method, knowledge representation method based on semantic network, knowledge representation method based on ontology, etc. Secondly, the process manufacturing knowledge graph also needs to realize the automatic construction of knowledge in this field, and the key technologies involved are named entity recognition, relationship extraction, entity alignment, etc. In terms of named entity technology, transfer learning technology can be applied to the process of named entity recognition in professional literature in the field of numerical control systems. In terms of relationship extraction, relationship extraction is performed based on the social interaction context of a large number of manufacturing relationships to help decision makers obtain decision support information that matches manufacturing demand capabilities, laying a good foundation for subsequent user intention analysis. However, due to the high accuracy requirements for relationship extraction in the field of intelligent manufacturing and the high cost of related data annotation, there are still various data noise and long-tail relationship problems that need to be solved. In terms of entity alignment, the construction and fusion process of process manufacturing knowledge graphs often faces multi-domain environments, and the distribution of entities is more variable. Graph convolutional networks and attention mechanisms can be introduced to characterize the importance weights of associated entities in the entity distribution in the knowledge graph of this field.

[0070] In an embodiment of the present invention, the manufacturing outline of the manufacturing process object can be used as a basic data source to build an associated reserve database. The manufacturing outline of the manufacturing process object is an understanding of the regularity of manufacturing activities formed by processing and refining various information in the manufacturing process of the manufacturing process object. From a structural point of view, the industrial knowledge involved in the manufacturing outline of the manufacturing process object is a combination of product, process and resource information in the manufacturing public welfare field that are linked together through certain rules. Different from single problem-solving knowledge, the process field knowledge for various products, processes and manufacturing activities is diverse, and the smallest granularity descriptive knowledge is decomposed into information. From a process point of view, the essence of the process of refining information into knowledge lies in the solution of process problems. Only on the basis of process mechanism research can process knowledge be accurately obtained. Since the knowledge source types of the manufacturing outline of the manufacturing process object are many and the composition information relationship is complex, it requires systematic organization and management to exert its effectiveness. With the continuous development and widespread application of CAD (Computer Aided Design) / CAM (Computer Aided Manufacturing) systems, a large number of digital three-dimensional CAD models and associated process data / knowledge are constantly generated and stored in the data / knowledge base of the enterprise.

[0071] Figure 4 is a flowchart for constructing a knowledge graph based on manufacturing process provided by the second embodiment of the present invention. In a specific example, Figure 4 As shown, when constructing a knowledge graph based on manufacturing process, multi-source heterogeneous data can be obtained from various associated data platforms of manufacturing process objects, and the obtained multi-source heterogeneous data can be formatted to obtain multi-tuple structure data, and then the knowledge graph is constructed and the system is expanded based on the multi-tuple structure data, thereby forming the results of the process manufacturing process knowledge graph application software platform.

[0072] In the process of constructing the knowledge graph of the entire process manufacturing outline, the following points need to be noted: the process manufacturing outline is a web page data format stored in the form of HTML (HyperText Markup Language), which is essentially a semi-structured data input; each data part within the process manufacturing outline needs to have clear relevant definitions and associated attributes in order to effectively integrate data and extract knowledge, thereby forming a preliminary knowledge representation; it is necessary to discuss with the process engineer how to do entity disambiguation and coreference resolution to form a standard knowledge representation; in the process part manufacturing outline, there are various process engineers' modification records and instructions that can form user log data records, which are an important source of knowledge reasoning and knowledge discovery. After the quality evaluation of the process engineer, the knowledge system is finally expanded; by forming a dynamically updated part process knowledge graph with a "human in the loop" working mode, and finally deploying it in a network collaborative platform, it can be used as a query reference for process engineers in the daily preparation of manufacturing outlines, and can also be used as a platform tool for process engineers' training and learning.

[0073] Optionally, the process manufacturing outline knowledge graph construction strategy can have the following technical paths:

[0074] (1) Adaptive path recommendation technology based on knowledge graph: The adaptive learning path recommendation technology based on knowledge graph is mainly divided into three modules. The first is the establishment of learner model; the second is the construction of domain knowledge graph and system expansion; the third is the application of learner model on the knowledge graph with expanded system, that is, recommendation strategy.

[0075] (2) Technical approaches for knowledge graph construction and representation for unstructured data: The knowledge graph construction and representation technology for unstructured data is mainly divided into five parts: knowledge graph framework design, title-based unstructured information extraction, structured knowledge extraction based on migration sorting model, entity linking model for knowledge fusion, and large-scale knowledge graph representation technology based on entity feature fusion.

[0076] In the process of building a knowledge graph, the knowledge base needs to be used to store the knowledge graph. In the process of extracting knowledge, it is necessary to build a knowledge ontology to extract structured, semi-structured and unstructured information. Figure 5 is a general flow chart of constructing a knowledge graph based on manufacturing process provided by the second embodiment of the present invention. In a specific example, Figure 5As shown in the figure, for structured knowledge, data mapping can be used to convert structured data into triples; for semi-structured and unstructured data, deep learning technology can be used to extract entities, relationships, and entity attributes. For the process knowledge base, for example, a large amount of data in the parts manufacturing outline is unstructured data such as text and pictures, so it is necessary to focus on knowledge extraction, linking, and representation of this type of unstructured information, and ultimately achieve high-precision knowledge extraction and representation of unstructured text.

[0077] In the embodiment of the present invention, the data of the process assembly process knowledge base is derived from the process manufacturing outline provided by the relevant process data platform. Such outline data is generally multiple heterogeneous data. For example, taking parts as manufacturing process objects, the data of the parts manufacturing outline mainly includes the following five parts.

[0078] (1) Basic information:

[0079] The basic information part is part of the content automatically generated by the compilation software during the FO compilation process, but it needs to be verified and modified by the technician. The "Part Number" in the basic information is the CATIA digital model number PART_NO of the part, which is unique and can be used as the unique identifier and index key of the knowledge base, such as the left skin "5341C03003G74-901".

[0080] (2) Historical record information:

[0081] The historical record information records the entire process of FO from preparation, proofreading, review to approval. The historical record information includes modification information, review information and tracking information. Corresponding associations can be established here to form the triple content of the knowledge graph. For example: the modification information is divided into the triple form of "work number"-"review result"-"part defect".

[0082] The modification information may include the work number of the modification content, the modification operation, the modification reason, and the deviation order. The modification operation, the modification reason, and the deviation order may be part of the "review result" in the modification information triple.

[0083] (3) Manufacturing basis:

[0084] The manufacturing basis is a programmatic document of the entire manufacturing outline, which describes the name of the digital model for processing and manufacturing, and the manufacturing standard CDS (Component Design Specification) file. The representative manufacturing basis triple data form is the "process digital model-part-part attribute" triple data;

[0085] (4) Parts information:

[0086] Part information represents the name and reference sketch of the part, and describes the overall properties of the part. A typical part information triple data form is a "part name-part number-part sketch" triple data.

[0087] (5) Process information:

[0088] The process information is the core content of the entire part outline. The data contained in its fields corresponds to specific process attribute information. In particular, the attributes of the work content field are relatively complex, and multiple layers of attribute information are nested inside. The process information is a typical nested table structure and belongs to structured data. The triple form here can be the triple form of "process-work content-work content of this process", such as "010"-"work content"-"FURNISH METAL furnace batch number: 691085".

[0089] S220. Configuring attribute information for the multi-tuple structure data; wherein the attribute information of the multi-tuple structure data includes quantitative attributes and qualitative attributes.

[0090] It is understandable that for the multi-tuple data content in the integrated knowledge base construction phase, both the entity part and the relationship part of the multi-tuple need to be defined. The attributes of the multi-tuple data can be divided into quantitative attributes and qualitative attributes.

[0091] Among them, quantitative attributes refer to attributes that can be described by value expressions, such as but not limited to surface roughness, corner tolerance, dimensional tolerance, and net edge requirements. Specifically, the similarity of quantitative attributes can be calculated using the following formula:

[0092]

[0093] Among them, X and Y represent two manufacturing process objects, such as two parts. represents the i-th characteristic attribute of X, represents the i-th characteristic attribute of Y, max(C i ) represents the maximum value of the i-th characteristic attribute in all manufacturing process objects, min(C i ) represents the minimum value of the i-th characteristic attribute among all manufacturing process objects.

[0094] Qualitative attributes can be other attributes that cannot be quantified but have semantic characteristics. For example, the similarity between "making holes" and "drilling holes" is much greater than the similarity between "making holes" and "deburring holes". The similarity of qualitative attributes can be described by semantic similarity based on word vectors. The calculation process is shown in the following formula:

[0095]

[0096] Among them, wv represents the word vector representation of qualitative attributes, which can be obtained by training a deep neural network with a self-attention mechanism. X and Y represent two manufacturing process objects, such as two parts. represents the i-th characteristic attribute of X, Represents the i-th feature attribute of Y.

[0097] Under the specified FO master template, during the process of compiling relevant content, the operator will input relevant attribute feature content. For example, when the operator inputs "part corner tolerance, root R angle tolerance", the process information recommendation platform will recommend the value "±0.75mm" corresponding to the quantitative attribute as a parameter; if the operator enters "set the process ear hole according to the digital model", the process information recommendation platform will automatically obtain the recommended prompts "hole making" and "diameter" related keywords as input through semantic similarity.

[0098] S230 , determining a qualitative attribute set of the multi-tuple structure data in the integrated knowledge base, and assigning a first initial recommendation weight to the qualitative attribute set of the multi-tuple structure data.

[0099] The first initial recommendation weight may be a weight value assigned to a qualitative attribute set of the multi-tuple structure data in an initialization phase.

[0100] For the process, the qualitative attribute content is easily affected by the operator's feedback, while the quantitative attribute content is objective and is not affected by the operator. Therefore, when assigning weights, the corresponding weight values ​​can be assigned to the qualitative attributes. Specifically, the qualitative attribute set {C1, C2, ..., C M}, where C M The Mth qualitative attribute is represented, and an initial first initial recommendation weight is assigned to the qualitative attribute set of the multi-tuple structure data.

[0101] Optionally, the first initial recommendation weight may be an average weight: in, represents the first initial recommendation weight of the Mth qualitative attribute, where M represents the number of qualitative attributes in a qualitative attribute set of a certain multi-tuple structure data.

[0102] S240: Determine the usage count of each characteristic attribute in the qualitative attribute set according to the operation behavior characteristics of the integrated knowledge base, and determine the update recommendation weight of each characteristic attribute according to the usage count of each characteristic attribute.

[0103] In an embodiment of the present invention, the update recommendation weight of each characteristic attribute can be dynamically adjusted based on the number of times each characteristic attribute in the qualitative attribute set is used, so that the value of the update recommendation weight can increase as the number of times the characteristic attribute is used increases, or decrease as the number of times the characteristic attribute is used decreases.

[0104] In a specific example, assuming that the process is taken as the manufacturing process object, the first target operator continuously uses d m times (m=1,2,3...,M), then the update recommendation weight after the process is used can be redistributed as follows: in, represents the updated recommendation weight of the Mth qualitative attribute.

[0105] It is understandable that the first target operator will also operate on the various keywords in the process when adding, deleting, modifying and checking the FO process, which will also affect the weight of the keywords. Keywords with a higher frequency of additions are more likely to be recommended for use in the subsequent FO compilation process.

[0106] S250: adjusting and updating the first initial recommendation weight of each of the characteristic attributes according to the update recommendation weight of each of the characteristic attributes.

[0107] Accordingly, after the update recommendation weight of each characteristic attribute is determined according to the number of times each characteristic attribute is used, the first initial recommendation weight of the corresponding characteristic attribute can be dynamically adjusted and updated using the determined update recommendation weight.

[0108] S260, scanning target keywords of all stored process objects in the integrated knowledge base, and calculating similarity between the target keyword and process query information according to updated recommendation weights of feature attributes of the target keyword.

[0109] When the second target operator inputs process query information to the process information recommendation platform, the process information recommendation platform can scan and integrate the target keywords of all stored process objects in the knowledge base, and calculate the similarity between the target keywords and the process query information based on the updated recommendation weights of the characteristic attributes of the target keywords.

[0110] In a specific example, taking the process as a manufacturing process object, after the second target operator fills in the relevant text information C in the process according to the process information recommendation platform, the process information recommendation platform can scan all process keywords i of all parts Z and calculate the similarity between the filled information C and each keyword as follows:

[0111]

[0112] in, represents the similarity between text information C and process keyword i of part Z, Indicates the updated recommendation weight of process keyword i for part Z.

[0113] Optionally, if the process query information input by the second target operator also includes quantitative attribute information, after calculating the similarity between the target keyword and the process query information based on the updated recommendation weight of the qualitative feature attribute of the target keyword, the similarity of the quantitative attribute information between each target keyword and the process query information can also be calculated.

[0114] S270 , sorting the stored process objects corresponding to each of the target keywords according to the similarity between each of the target keywords and the process query information.

[0115] S280 , determining the target process-related recommendation result according to the sorting result of the stored process objects corresponding to each of the target keywords.

[0116] Correspondingly, after calculating the similarity between the target keyword and the process query information based on the updated recommendation weight of the characteristic attribute of the target keyword, the calculated similarity can be used to sort the stored process objects corresponding to each target keyword in reverse order from large to small, and the target process-related recommendation results can be determined based on the sorting results of the stored process objects corresponding to each target keyword.

[0117] Optionally, if only the similarity of qualitative attributes is calculated, the stored process objects corresponding to each target keyword can be sorted based on the similarity of qualitative attributes alone, and the target process association recommendation result can be determined based on the sorting result of the stored process objects corresponding to each target keyword. In a specific example, assuming that the similarity of process keyword 1 is 90%, the similarity of process keyword 2 is 80%, and the similarity of process keyword 1 is 70%, the sorting result of each process keyword can be: process keyword 1-process keyword 2-process keyword 3.

[0118] Optionally, if the similarity of qualitative attributes and quantitative attributes is calculated at the same time, the stored process objects corresponding to each target keyword can be sorted based on the similarity of qualitative attributes and quantitative attributes at the same time, and the target process-related recommendation results can be determined based on the sorting results of the stored process objects corresponding to each target keyword. In a specific example, assuming that the process query information includes the quantitative attribute part angle tolerance: 0.75, the process information recommendation platform calculates that the similarity of process keyword 1 is 90%, and the part angle tolerance is 0.9; the similarity of process keyword 2 is 85%, and the part angle tolerance is 0.75; the similarity of process keyword 1 is 70%, and the part angle tolerance is 0.95; then the sorting results of each process keyword can be:

[0119] 1. Process keyword 2 Part corner tolerance: ±0.75;

[0120] 2. Process keywords 1 Part corner tolerance: ±0.9;

[0121] 3. Process keyword 3: Part corner tolerance: ±0.95.

[0122] It should be noted that if the similarity of qualitative attributes and quantitative attributes is calculated at the same time, similarity weights can be set for the similarity of qualitative attributes and the similarity of quantitative attributes respectively to characterize the importance of similarity in a certain dimension, and the comprehensive similarity of the target keyword can be calculated based on the similarity weights set for the similarity of qualitative attributes and the similarity of quantitative attributes. For example, the similarity weight of the similarity of qualitative attributes can be 0.5, the similarity weight of the similarity of quantitative attributes can be 0.5, and so on. If the similarity of the qualitative attributes of a target keyword is 90% and the similarity of the quantitative attributes is 80%, then the comprehensive similarity of the target keyword is: 90%*0.5+80%*0.5=85%.

[0123] By reverse sorting, the target keyword with the highest similarity score is the keyword that is most similar to the newly added process query information of the second target operator. Its score will also increase as the recommendation weight increases, and the recommendation ranking will be higher.

[0124] The technical solution of the embodiment of the present invention starts from the perspective of keywords in the process manufacturing outline, integrates the operator's usage behavior feature information, proposes a recommendation algorithm for the process manufacturing outline keywords, models the operator's click behavior, and dynamically adjusts the similarity between the operator's new information and the selected keywords through weight analysis. It can continuously improve the accuracy of recommendations during the operator's use and greatly improve the practicality of the manufacturing process knowledge graph. Through adaptive keyword push based on operator feedback, it has strong scalability, can access more knowledge base content, form a more personalized push function, simplify the process of technician training, and solidify the technician's production process experience through weight distribution, which can be used for subsequent technician training.

[0125] Embodiment 3

[0126] Figure 6 is a flowchart of a method for recommending process information provided by the third embodiment of the present invention. This embodiment is specific based on the above embodiment. In this embodiment, the outline master template library is used as an associated reserve database, and the associated reserve database for constructing manufacturing process objects is provided, and the recommended weights of the process objects stored in the associated reserve database are adjusted and updated, and the process query information input by the second target operator is determined according to the adjusted and updated associated reserve database. Correspondingly, Figure 6 As shown, the method of this embodiment may include:

[0127] S310, constructing an outline master of the manufacturing process object, and constructing an outline master template library of the manufacturing process object according to the outline master of the manufacturing process object; wherein the outline master includes a plurality of link processing modules.

[0128] Among them, the outline master can be used for the overall process flow of the standard manufacturing process object. The outline master template library can be used to manage all outline masters of the manufacturing process object.

[0129] In an embodiment of the present invention, the outline master of each manufacturing process object may include multiple link processing modules, and each link processing module may abstract a certain process link of the manufacturing process object. Exemplarily, when the manufacturing process object is a sheet metal part product put into production, its outline master may include four link processing modules, namely, a feeding module, a shape milling module, a heat treatment module, and a painting module. It is understandable that the type and number of link processing modules included in its outline master may be different depending on the type of manufacturing process object.

[0130] Figure 7 Schematic diagram of a part process FO master construction process provided by the third embodiment of the present invention. In a specific example, Figure 7 As shown, taking a part as an example of a manufacturing process object, the link processing modules of the part process outline master can include a feeding module, a shape milling module, a heat treatment module, and a painting module. Accordingly, the construction process of the part process outline master can include the following operations:

[0131] (1) Extract relevant information from the digital model structure tree and select the appropriate material series. The digital model structure tree can include material information involved in the part process, which comes from multi-source heterogeneous data in the integrated knowledge base.

[0132] (2) Select the feeding module.

[0133] (3) Select the shape milling module according to the shape requirements. It is divided into multiple-choice one and multiple-choice multiple modes. In the multiple-choice one shape milling mode, in addition to the standard module, other types of optional modules can be appropriately added.

[0134] (4) After determining the milling module, you can also choose the molding method. Optionally, you can choose a one-shot molding or a two-shot molding method based on the complexity of the digital model.

[0135] (5) Select the heat treatment method. For the heat treatment module, relevant information can be extracted from the digital model structure tree to select whether to perform a one-step heat treatment or a two-step heat treatment method. The heat treatment process may include but is not limited to reaming, reshaping, weighing, conductivity testing, fluorescent penetrant testing, and selection of conversion molds. It is understandable that problems found in the part process processing flow may be found in the heat treatment link. For this reason, corresponding modules can be added in the heat treatment link to adapt and solve the problems found in the part process processing flow of the above process.

[0136] (6) Select the painting method. Optionally, you can choose whether to perform multi-step painting and whether to apply topcoat based on the primer in the digital model structure tree.

[0137] In an optional embodiment of the present invention, after constructing the associated reserve database of manufacturing process objects, it may also include: constructing text blocks for each link processing module of the outline master; and calculating the topic distribution of the text blocks using a topic model construction method.

[0138] In some cases, FOs for multiple manufacturing process objects may be compiled using similar master copies. For such scenarios, unsupervised machine learning techniques can be used to treat each FO document as a word frequency vector, thereby converting text information into digital information that is easy to model. Each FO document represents the probability distribution of some topic composition, and these topics represent the probability distribution of many words. The FOs of manufacturing process objects that use the same FO master copy are similar in probability distribution and should have the same topic distribution.

[0139] In an embodiment of the present invention, optionally, a topic model construction method can be used to calculate the topic distribution of the outline master. Optionally, the topic model construction method can use an LDA (Latent Dirichlet Allocation) algorithm. LDA is a topic-based probability generation model. Based on the LDA algorithm, the potential topics in the manufacturing process object FO document can be realized, and the FO master template is recommended based on the similarity with the potential topics in each FO document. The process information recommendation platform based on LDA can effectively solve the data sparsity and cold start problems, and can improve the accuracy and diversity of recommendations.

[0140] Since the outline master includes multiple link processing modules, corresponding weights can be set for each link processing module to characterize the topic characteristics of the FO master template, and then the text involved in each link processing module is composed into a text block, and the topic distribution of this text block is calculated. Optionally, one link processing module can create a corresponding text block.

[0141] In a specific example, assuming that the link processing modules of the FO master template include four modules: feeding module, shape milling module, heat treatment module and painting module, then a corresponding weight W can be set for each link processing module. k , k=1,2,3,4 to characterize the theme characteristics of the FO master template. Then the text involved in each link processing module is composed of text blocks T k ={T kl ,l=1,2,3,...,L k}, k = 1, 2, 3, 4, where T kl Represents the lth word in the kth module and calculates this text block T k The topic distribution θ k .

[0142] Optionally, the steps to calculate the topic distribution of a text block are as follows:

[0143] (1) Generate text block T by sampling according to Dirichlet distribution with α as the prior parameter k The topic distribution θ k .

[0144] (2) From the text block topic distribution θ k Sampling generated text block T k The topic z of the jth word kj .

[0145] (3) Generate topic z by sampling according to the Dirichlet distribution with β as the prior parameter kj The corresponding word distribution φ kj .

[0146] (4) From the word topic distribution φ kj Sampling generated words w kj .

[0147] (5) Rescan the entire FO master template library, and for each w kj Use Gibbs Sampling to sample, find its theme, and update it in the corpus.

[0148] (6) Return to step (1) and continue to step (5), repeating the loop until Gibbs Sampling converges, and finally obtaining each text block T k The topic distribution

[0149] Correspondingly, after calculating the topic distribution of the text blocks corresponding to the link processing modules of each major outline master, the similarity between different FO masters can be calculated based on the topic distribution, which can be represented by the following formula.

[0150]

[0151] Among them, P and Q represent two FO mother copies, E P is the expectation of FO parent P, W k is a hyperparameter adjusted based on operator experience, Represents the text block T in the FO parent text P k The topic distribution of Indicates the text block T in the FO parent Q k The topic distribution.

[0152] S320. Allocate a second initial recommendation weight to each major outline master template in the outline master template library.

[0153] Among them, the second initial recommendation weight can be the weight value assigned to each major outline master in the initialization stage.

[0154] Specifically, the existing FO master template library is {F1, F2, ..., F N}, F N It represents the Nth FO master, and can assign the second initial recommendation weight to each major outline master in the FO master template library.

[0155] Optionally, the second initial recommendation weight may be an average weight. in, represents the second initial recommendation weight of the Nth FO parent, where N represents the number of FO parents.

[0156] S330. Determine the number of clicks on each of the outline masters according to the operational behavior characteristics of the outline master template library, and determine the update recommendation weight of each of the outline masters according to the number of clicks on each of the outline masters.

[0157] In an embodiment of the present invention, the update recommendation weights of each outline master can be dynamically adjusted based on the number of clicks on each outline master in the qualitative attribute set, so that the value of the update recommendation weight can increase with the increase of the number of clicks on the outline master, or decrease with the decrease of the number of clicks on the outline master.

[0158] In a specific example, assuming that the process is taken as the manufacturing process object, the first target operator performs the operation on the outline master F m Click on c m times (m=1,2,3...,N), then the update recommendation weight after clicking on the outline master can be redistributed as follows: in, Represents the update recommendation weight of the mth outline master.

[0159] During the logging stage of the first target operator's operation behavior, the process information recommendation platform can record the first target operator's FO new data behavior, FO version upgrade / ink change behavior, search for the minimum change FO master template behavior, and FO process addition, deletion, modification, and query behavior. Among them, when the first target operator compiles a new FO file due to the process update of the manufacturing process object, he often searches and selects the relevant FO master template in the search for the minimum change FO master template behavior. The FO master template selected by the first target operator will automatically adjust the size of the recommendation weight so that it is easier to be detected in subsequent searches.

[0160] S340, adjusting and updating the second initial recommendation weight of each of the outline masters according to the update recommendation weight of each of the outline masters.

[0161] Correspondingly, after the update recommendation weights of each major outline master are determined according to the number of clicks of each major outline master, the second initial recommendation weights of the corresponding outline master can be dynamically adjusted and updated using the determined update recommendation weights.

[0162] S350, calculating the current topic distribution of the process query information, and calculating the weighted similarity of the topic distribution between the current topic distribution of the process query information and the topic distribution of each outline master template in the outline master template library.

[0163] The weighted similarity of topic distribution may be a result of weighted calculation of the similarity of topic distribution of each text block.

[0164] Optionally, a topic model building method may be used to calculate the current topic distribution of the process query information.

[0165] In a specific example, taking the process as the manufacturing process object, for a newly added FO master O, after the second target operator inputs the basic information of the four-link processing modules of the FO master O to the process information recommendation platform, the process information recommendation platform can use the information input by the second target operator as the process query information, and automatically calculate the process query information through LDA to obtain the current topic distribution of the process query information At the same time, the process information recommendation platform scans all FO master templates F in the FO master template library, and calculates the weighted similarity of the topic distribution between the current topic distribution and the topic distribution of each major outline master template in the outline master template library based on the following formula:

[0166]

[0167] Among them, Sim(O,F) represents the weighted similarity of the topic distribution between the current topic distribution of FO parent O and the topic distribution of outline parent F. represents the update recommendation weight of the outline master F, W k is a hyperparameter adjusted based on operator experience, Represents the text block T in the FO parent text O k The topic distribution of Indicates the text block T in the parent text F of FO k The topic distribution.

[0168] S360, sorting each of the outline masters according to the weighted similarity of the topic distribution.

[0169] S370. Determine the target process association recommendation result according to the sorting results of each outline master.

[0170] Correspondingly, after calculating the weighted similarity of the topic distribution between the topic distribution of each outline master and the current topic distribution of the process query information according to the updated recommendation weight of the outline master, the calculated weighted similarity of the topic distribution can be used to sort the various outline masters in reverse order from large to small, and determine the target process-related recommendation results based on the sorting results of the various outline masters.

[0171] Through reverse sorting, the FO master template with the highest weighted similarity in topic distribution is the outline master template most similar to the newly added master template O. The outline master template corresponding to the newly added master template O is modified based on the FO master template with the highest weighted similarity in topic distribution, and the change amount is also the least. The FO master template that is frequently selected during use will have a larger score due to the increase in weight, and the recommended ranking will be higher.

[0172] It can be seen that the above technical solution is based on the modification records and instructions of various operators in the process manufacturing outline, forming the log data records of operators as an important source of process knowledge reasoning and knowledge discovery, introducing the adaptive learning technology concept in "AI (Artificial Intelligence, Artificial Intelligence) + Education", forming an "operator in the loop" working mode, using the data fed back by the operator as input to train the knowledge model, thereby attracting more operators to use it, thereby generating more feedback data to further train a larger knowledge model. This is repeated, thus forming a "data flywheel" effect. Ultimately, the process knowledge model will assist operators to accurately understand the actual needs of the newly edited process manufacturing outline and the corresponding relationship between the manufacturing outline master, find the manufacturing outline master with the least modification and the closest content for editing, further reduce the workload and training burden of operators, and at the same time improve process manufacturing efficiency and reduce rework.

[0173] The technical solution of the embodiment of the present invention is studied from the perspective of the FO master theme from the process manufacturing outline, integrating the operator's click behavior feature information, and proposing a recommendation algorithm for the FO master template as a whole. For the FO master recommendation problem with a more complex structure, the implicit FO master theme distribution is used instead of the explicit text word vector analysis, which can avoid excessive deep learning computing resource consumption and improve the query rate of complex structure FO master recommendation. Through the FO master push based on operator feedback, it has strong scalability, can access more knowledge base content, form a more personalized push function, simplify the process of technician training, and solidify the technician's production process experience through weight distribution, which can be used for subsequent technician training.

[0174] In the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information (such as operational behavior characteristics for querying associated reserve databases, etc.) are in compliance with the relevant laws and regulations and do not violate public order and good morals.

[0175] It should be noted that any arrangement and combination of the technical features in the above embodiments also falls within the protection scope of the present invention.

[0176] Embodiment 4

[0177] Figure 8 is a schematic diagram of a process information recommendation device provided by a fourth embodiment of the present invention, such as Figure 8 As shown, the device includes: an associated reserve database construction module 410, an operation behavior feature acquisition module 420, a recommendation weight adjustment and update module 430, and a target process associated recommendation result determination module 440, wherein:

[0178] The associated reserve database construction module 410 is used to construct an associated reserve database of manufacturing process objects;

[0179] An operation behavior feature acquisition module 420 is used to acquire the operation behavior feature of the first target operator querying the associated reserve database;

[0180] A recommended weight adjustment and update module 430 is used to adjust and update the recommended weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database;

[0181] The target process association recommendation result determination module 440 is used to determine the matching target process association recommendation result for the process query information input by the second target operator according to the adjusted and updated association reserve database.

[0182] The embodiment of the present invention constructs an associated reserve database for manufacturing process objects and obtains the operational behavior characteristics of the first target operator when querying the associated reserve database, thereby adjusting and updating the recommendation weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database, and then determines the matching target process associated recommendation results for the process query information input by the second target operator according to the adjusted and updated associated reserve database, thereby solving the problem of low recommendation accuracy in the existing process information recommendation methods and being able to continuously improve the accuracy of information recommendation based on the interaction data between users and the process information recommendation platform.

[0183] Optionally, the associated reserve database includes an integrated knowledge base, and the associated reserve database construction module 410 is also used to: obtain multi-source heterogeneous data provided by each associated data platform of the manufacturing process object; format the multi-source heterogeneous data to obtain multi-tuple structure data; configure attribute information for the multi-tuple structure data; wherein the attribute information of the multi-tuple structure data includes quantitative attributes and qualitative attributes.

[0184] Optionally, the recommendation weight adjustment and update module 430 is also used to: determine a set of qualitative attributes of the multi-tuple structure data in the integrated knowledge base; assign a first initial recommendation weight to the set of qualitative attributes of the multi-tuple structure data; determine the number of times each characteristic attribute in the set of qualitative attributes is used according to the operational behavior characteristics of the integrated knowledge base; determine an updated recommendation weight for each characteristic attribute according to the number of times each characteristic attribute is used; and adjust and update the first initial recommendation weight of each characteristic attribute according to the updated recommendation weight of each characteristic attribute.

[0185] Optionally, the target process association recommendation result determination module 440 is also used to: scan the target keywords of all stored process objects in the integrated knowledge base; calculate the similarity between the target keywords and the process query information based on the updated recommendation weights of the characteristic attributes of the target keywords; sort the stored process objects corresponding to each target keyword according to the similarity between each target keyword and the process query information; and determine the target process association recommendation result based on the sorting results of the stored process objects corresponding to each target keyword.

[0186] Optionally, the associated reserve database includes an outline master template library, and the associated reserve database construction module 410 is also used to: construct an outline master for the manufacturing process object, and construct an outline master template library for the manufacturing process object based on the outline master of the manufacturing process object; wherein the outline master includes multiple link processing modules; the process information recommendation device may also include a topic distribution calculation module, which is used to: construct text blocks for each link processing module of the outline master; and use a topic model construction method to calculate the topic distribution of the text block.

[0187] Optionally, the recommendation weight adjustment and update module 430 is also used to: assign a second initial recommendation weight to each outline master in the outline master template library; determine the number of clicks on each outline master according to the operational behavior characteristics of the outline master template library; determine the update recommendation weight of each outline master according to the number of clicks on each outline master; and adjust and update the second initial recommendation weight of each outline master according to the update recommendation weight of each outline master.

[0188] Optionally, the target process associated recommendation result determination module 440 is also used to: calculate the current topic distribution of the process query information; calculate the weighted similarity of the topic distribution between the current topic distribution of the process query information and the topic distribution of each outline master in the outline master template library; sort each outline master according to the weighted similarity of the topic distribution; and determine the target process associated recommendation result based on the sorting results of each outline master.

[0189] The above-mentioned process information recommendation device can execute the process information recommendation method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not fully described in this embodiment, please refer to the process information recommendation method provided by any embodiment of the present invention.

[0190] Since the process information recommendation device described above is a device that can execute the process information recommendation method in the embodiment of the present invention, based on the process information recommendation method described in the embodiment of the present invention, a person skilled in the art can understand the specific implementation of the process information recommendation device in the present embodiment and its various variations, so how the process information recommendation device implements the process information recommendation method in the embodiment of the present invention is not described in detail here. As long as a person skilled in the art implements the device used in the process information recommendation method in the embodiment of the present invention, it belongs to the scope of protection of this application.

[0191] Embodiment 5

[0192] Fig. 9 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0193] like Fig. 9 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0194] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0195] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the process information recommendation method.

[0196] Optionally, the process information recommendation method may include: constructing an associated reserve database for manufacturing process objects; obtaining operational behavior characteristics of a first target operator when querying the associated reserve database; adjusting and updating recommendation weights of process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database; and determining matching target process associated recommendation results for process query information input by a second target operator according to the adjusted and updated associated reserve database.

[0197] In some embodiments, the process information recommendation method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the process information recommendation method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the process information recommendation method in any other appropriate manner (e.g., by means of firmware).

[0198] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0199] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0200] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0201] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0202] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0203] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

Claims

1. A process information recommendation method, characterized in that: include: Build an associated reserve database of manufacturing process objects; Acquire the operation behavior characteristics of the first target operator querying the associated reserve database; Adjusting and updating the recommended weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database; The process query information input by the second target operator is determined according to the adjusted and updated associated reserve database to match the target process associated recommendation result.

2. The method according to claim 1, characterized in that The associated reserve database includes an integrated knowledge base, and the associated reserve database for building manufacturing process objects includes: Acquire multi-source heterogeneous data provided by various associated data platforms of the manufacturing process object; Formatting the multi-source heterogeneous data to obtain multi-tuple structure data; Attribute information is configured for the multi-tuple structure data; wherein the attribute information of the multi-tuple structure data includes quantitative attributes and qualitative attributes.

3. The method according to claim 2, characterized in that The adjusting and updating the recommended weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database includes: Determining a set of qualitative attributes of multi-tuple structure data in the integrated knowledge base; assigning a first initial recommendation weight to the qualitative attribute set of the multi-tuple structure data; Determining the usage count of each characteristic attribute in the qualitative attribute set according to the operational behavior characteristics of the integrated knowledge base; Determining an update recommendation weight for each of the characteristic attributes according to the number of times each of the characteristic attributes is used; The first initial recommendation weight of each of the feature attributes is adjusted and updated according to the updated recommendation weight of each of the feature attributes.

4. The method according to claim 3, characterized in that The step of determining a matching target process association recommendation result for the process query information input by the second target operator according to the adjusted and updated association reserve database includes: Scanning target keywords of all stored process objects in the integrated knowledge base; Calculating the similarity between the target keyword and the process query information according to the updated recommendation weight of the characteristic attribute of the target keyword; Sorting the stored process objects corresponding to each of the target keywords according to the similarity between each of the target keywords and the process query information; The target process association recommendation result is determined according to the sorting result of the stored process objects corresponding to each of the target keywords.

5. The method according to claim 1, characterized in that The associated reserve database includes an outline master template library, and the associated reserve database for constructing the manufacturing process object includes: Constructing an outline master of the manufacturing process object, and constructing an outline master template library of the manufacturing process object according to the outline master of the manufacturing process object; wherein the outline master includes a plurality of link processing modules; After the associated reserve database of manufacturing process objects is constructed, the method further includes: Constructing text blocks for each link processing module of the outline master; A topic model building method is adopted to calculate the topic distribution of the text block.

6. The method according to claim 5, characterized in that The adjusting and updating the recommended weights of the process objects stored in the associated reserve database according to the operational behavior characteristics of the associated reserve database includes: Allocating a second initial recommendation weight to each outline master template in the outline master template library; Determine the number of clicks on each of the outline masters according to the operational behavior characteristics of the outline master template library; Determine the update recommendation weight of each outline master according to the number of clicks on each outline master; The second initial recommendation weight of each outline master is adjusted and updated according to the update recommendation weight of each outline master.

7. The method according to claim 6, characterized in that The step of determining a matching target process association recommendation result for the process query information input by the second target operator according to the adjusted and updated association reserve database includes: Calculating the current topic distribution of the process query information; Calculating the weighted similarity of the current topic distribution of the process query information and the topic distribution of each outline master template in the outline master template library; Sorting each of the outline masters according to the weighted similarity of the topic distribution; The target process association recommendation result is determined according to the sorting results of each outline master.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the process information recommendation method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the process information recommendation method described in any one of claims 1-7 when executed.

10. A computer program product comprising a computer program / instructions, wherein: When the computer program / instructions are executed by a processor, the process information recommendation method according to any one of claims 1 to 7 is implemented.

Images