CN116186404A

CN116186404A - Knowledge recommendation method and system

Info

Publication number: CN116186404A
Application number: CN202310155645.1A
Authority: CN
Inventors: 葛新
Original assignee: Shanghai Zhidao Knowledge Digital Technology Co ltd
Current assignee: Shanghai Zhidao Knowledge Digital Technology Co ltd
Priority date: 2023-02-23
Filing date: 2023-02-23
Publication date: 2023-05-30

Abstract

The invention provides a knowledge recommendation method and a knowledge recommendation system, wherein the method is executed based on a processor of a symbiotic learning platform, and comprises the following steps: acquiring obstacle information, wherein the obstacle information at least comprises obstacle classification; generating recommended learning content through a symbiotic learning platform based on the obstacle information; and pushing and displaying the obstacle information and the corresponding recommended learning content to the user through the symbiotic learning platform.

Description

Knowledge recommendation method and system

Technical Field

The specification relates to the field of symbiotic learning, in particular to a knowledge recommendation method and system.

Background

The symbiotic learning method is a novel talent culture mode method summarized by combining new teaching mode practice on the basis of the practical background of technology enabling education in the past years. Currently, symbiotic learning methods are widely used in many enterprises and high schools. However, how to better develop the advantages of the internet technology and the platform technology, the technology of recommending the learning content with pertinence for different users is still not perfect.

Aiming at how to recommend learning content to different users, CN111859140B provides a knowledge recommendation method, which comprises the steps of obtaining current learning knowledge points and target learning knowledge points of users, inputting the current learning knowledge points and the target learning knowledge points into a preset knowledge graph to obtain a learning knowledge chain of the users, recommending the knowledge points in the learning knowledge chain to the users according to the sequence, and correcting the learning knowledge chain in real time according to the mastering condition of the users on the current learning knowledge points. However, this knowledge recommendation method is not applicable in all application scenarios. For example, when the scenario using the knowledge recommendation method is a production processing scenario, the learning content matched with the user and the scenario cannot be accurately determined only according to the current learning knowledge point and the target learning knowledge point of the user, and the actual situation (such as a fault situation) of each process in the processing space needs to be further combined.

Therefore, for different production contents of different enterprises, how to recommend learning contents matched with the production contents to staff of the enterprises, a knowledge recommendation method and a knowledge recommendation system are needed to help the staff of the enterprises to timely and accurately acquire the learning contents, and the orderly production flow is ensured.

Disclosure of Invention

One or more embodiments of the present specification provide a knowledge recommendation method, which is executed based on a processor of a symbiotic learning platform, the method comprising: acquiring obstacle information, wherein the obstacle information at least comprises obstacle classification; generating recommended learning content through a symbiotic learning platform based on the obstacle information; and pushing and displaying the obstacle information and the corresponding recommended learning content to the user through the symbiotic learning platform.

One of the embodiments of the present specification provides a knowledge recommendation system, the system including a symbiotic learning platform, the system further including: the first acquisition module is used for acquiring obstacle information, wherein the obstacle information at least comprises obstacle classification; the generation module is used for generating recommended learning content through the symbiotic learning platform based on the obstacle information; the display module is used for pushing and displaying the obstacle information and the corresponding recommended learning content to the user through the symbiotic learning platform.

One or more embodiments of the present specification provide a knowledge recommendation apparatus including a processor for performing a knowledge recommendation method.

One or more embodiments of the present specification provide a computer-readable storage medium storing computer instructions that, when read by a computer in the storage medium, perform a knowledge recommendation method.

Drawings

The present specification will be further elucidated by way of example embodiments, which will be described in detail by means of the accompanying drawings. The embodiments are not limiting, in which like numerals represent like structures, wherein:

FIG. 1 is a schematic illustration of an application scenario of a knowledge recommendation system, shown in accordance with some embodiments of the present description;

FIG. 2 is an exemplary block diagram of a knowledge recommendation system, shown in accordance with some embodiments of the specification;

FIG. 3 is an exemplary flow chart of a knowledge recommendation method shown in accordance with some embodiments of the present description;

FIG. 4 is an exemplary flow chart for generating recommended learning content according to some embodiments of the present description;

FIG. 5 is an exemplary schematic diagram of a knowledge-graph, according to some embodiments of the present description;

FIG. 6 is an exemplary schematic diagram of an assessment model shown in accordance with some embodiments of the present description;

FIG. 7 is an exemplary flow chart for determining a further training program for a user, shown in accordance with some embodiments of the present description;

FIG. 8 is an exemplary schematic diagram of an obstacle prediction model, shown in accordance with some embodiments of the present disclosure.

Detailed Description

In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings that are required to be used in the description of the embodiments will be briefly described below. It is apparent that the drawings in the following description are only some examples or embodiments of the present specification, and it is possible for those of ordinary skill in the art to apply the present specification to other similar situations according to the drawings without inventive effort. Unless otherwise apparent from the context of the language or otherwise specified, like reference numerals in the figures refer to like structures or operations.

It will be appreciated that "system," "apparatus," "unit" and/or "module" as used herein is one method for distinguishing between different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

As used in this specification and the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.

A flowchart is used in this specification to describe the operations performed by the system according to embodiments of the present specification. It should be appreciated that the preceding or following operations are not necessarily performed in order precisely. Rather, the steps may be processed in reverse order or simultaneously. Also, other operations may be added to or removed from these processes.

Fig. 1 is a schematic view of an application scenario of a knowledge recommendation system according to some embodiments of the present disclosure.

In some embodiments, the application scenario 100 of the knowledge recommendation system may include a production space 110, a network 120, a symbiotic learning platform 130, a terminal device 140, and a user 170.

The production space 110 refers to a basic unit of organization and management of production within an enterprise. The production space 110 may be divided into different types of production spaces according to different production requirements of different enterprises. The production space 110 may include a management space 111, an experiment space 112, a processing space 113, and the like. Wherein, different types of production spaces can be further refined according to actual production requirements. For example, the processing space 113 may be a production plant, and different kinds of works and different production contents may correspond to different production plants, e.g., the production plant may include a raw material processing plant, a processing plant, an assembly plant, and the like.

In some embodiments, the production space 110 may send relevant information (e.g., obstacle information, etc.) of the production space 110 to the processor 140 in the symbiotic learning platform 130 over the network 120 for processing to generate relevant learning content (e.g., recommended learning content, etc.). In some embodiments, the production space 110 may send information about the production space 110 to a storage device 150 in the symbiotic learning platform 130 for storage over the network 120. In some embodiments, production space 110 may send relevant information of production space 110 to storage device 150 for storage in real-time. In some embodiments, the production space 110 may send information about the production space 110 to the storage device 150 for storage on a periodic basis (e.g., every hour). For more on obstacle information, recommended learning content see fig. 3 and its associated description.

The network 120 may facilitate the exchange of information and/or data. In some embodiments, one or more components in the application scenario 100 (e.g., the production space 110, the processor 140, the storage device 150, etc.) may send information and/or data to another component in the application scenario 100 via the network 120. Network 120 may include a Local Area Network (LAN), wide Area Network (WAN), wired network, wireless network, etc., or any combination thereof. In some embodiments, network 120 may be any one or more of a wired network or a wireless network. In some embodiments, network 120 may include one or more network access points. For example, the network 120 may include wired or wireless network access points, such as base stations and/or network switching points, through which one or more components of the application scenario 100 may connect to the network 120 to exchange data and/or information.

The symbiotic learning platform 130 refers to an automated learning platform for processing relevant information of a production space, generating relevant learning content to assist the user 170 in acquiring knowledge or performing skill training. In some embodiments, symbiotic learning platform 130 can include processor 140 and storage device 150.

The processor 140 may be used to process data and/or information from at least one component of the application scenario 100 or an external data source (e.g., a cloud data center). Processor 140 may be coupled to production space 110, storage device 150, and/or terminal device 160 via network 120 to access and/or receive data and information. For example, processor 140 may receive information regarding production space 110 via network 120.

In some embodiments, processor 140 may be a single processor or a group of processors. The processor complex may be centralized or distributed (e.g., processor 140 may be a distributed system), may be dedicated, or may be concurrently serviced by other devices or systems. In some embodiments, the processor 140 may be connected locally to the network 120 or remotely from the network 120. In some embodiments, the processor 140 may be implemented on a cloud platform. For example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, or the like, or any combination thereof.

Storage device 150 may be used to store data and/or instructions. The data may include data related to the production space 110, etc. In some embodiments, the storage device 150 may store data and/or instructions that the processor 140 uses to execute or use to perform the exemplary methods described in this specification. For example, the storage device 150 may store information related to the production space 110. For another example, the storage device 150 may store one or more machine learning models. In some embodiments, the storage device 150 may be part of the processor 140.

In some embodiments, the storage device 150 may include mass storage, removable storage, volatile read-write memory, read-only memory (ROM), and the like, or any combination thereof. In some embodiments, storage device 150 may be implemented on a cloud platform. In some embodiments, the storage device 150 may be connected to the network 120 to communicate with one or more components of the application scenario 100 (e.g., the production space 110, the processor 140, etc.).

Terminal device 160 may refer to one or more terminal devices or software used by a user. In some embodiments, terminal device 160 can include a mobile device 161, a tablet computer 162, a laptop computer 163, and the like, or any combination thereof. In some embodiments, terminal device 160 may include a signal transmitter and a signal receiver configured to communicate with processor 140 to obtain the relevant information.

In some embodiments, terminal device 160 may be stationary and/or mobile. For example, the terminal device 160 may be directly mounted on the processor 140 as part of the processor 140. As another example, terminal device 160 may be a removable device and user 170 may carry terminal device 160 at a remote location relative to processor 140, terminal device 160 may be coupled to and/or in communication with processor 140 via network 120.

In some embodiments, terminal device 160 may receive user 170 requests and/or feedback information and send information related to the request to processor 140 via network 120. For example, the terminal device 160 may receive a request from the user 170 to generate recommended learning content and send information related to the request to the processor 140 via the network 120. Terminal device 160 can also receive information from processor 140 via network 120. For example, the terminal device 160 may receive, via the network 120, from the processor 140, relevant information of the production space 110, which may be displayed on the terminal device 160. For another example, the processor 140 may transmit the related learning content generated by processing the related information of the production space 110 to the terminal device 160.

The user 170 refers to a staff member who performs learning or training through the symbiotic learning platform 130. Such as a producer, technician, manager, etc. in the production space 110. In some embodiments, different types of production spaces 110 may correspond to different users 170. For example, the management space 111 may correspond to a manager, the experiment space 112 may correspond to a technician, and the processing space 113 may correspond to a producer.

In some embodiments, the user 170 may actively acquire relevant learning content through the terminal device 160. For example, the user 170 may send a request to generate recommended learning content to the symbiotic learning platform 130 through the terminal device 160 and learn based on the content displayed by the terminal device 160. In some embodiments, the enterprise may organize the user 170 periodically (e.g., every month) for learning and/or training of relevant learning content. For example, an enterprise may conduct skill training through the symbiotic learning platform 130 by staff organizing the various production spaces at the end of each month.

It should be noted that the application scenario is provided for illustrative purposes only and is not intended to limit the scope of the present description. Many modifications and variations will be apparent to those of ordinary skill in the art in light of the present description. For example, the application scenario may also include a database. As another example, application scenarios may be implemented on other devices to implement similar or different functionality. However, variations and modifications do not depart from the scope of the present description.

In some embodiments of the present description, the principles of the method of generating recommended learning content, pushing and presenting to users, and determining further training plans for users are open to different types of production spaces through a symbiotic learning platform. A knowledge recommendation method based on the realization of the symbiotic learning platform will be described in detail below by taking the processing space as an example.

FIG. 2 is an exemplary block diagram of a knowledge recommendation system, shown in accordance with some embodiments of the specification. As shown in fig. 2, in some embodiments, the knowledge recommendation system 200 may include a first acquisition module 210, a generation module 220, and a presentation module 230.

The first acquisition module 210 may be configured to acquire obstacle information including at least an obstacle classification. For more information on obstacles, obstacle classification see fig. 3 and its associated description.

The generation module 220 may be configured to generate recommended learning content through a symbiotic learning platform based on the obstacle information. See fig. 1 for more content on the symbiotic learning platform, and fig. 3-6 and their associated description for more content on the recommended learning content and its manner of generation.

In some embodiments, the generating module 220 may be further configured to determine at least one training content satisfying a first preset condition with the production obstacle information based on the production obstacle information in combination with a knowledge graph, the knowledge graph being related to one or more of a user, a production process, a product, an obstacle event, an obstacle classification, and the training content; taking at least one training content as at least one candidate recommended learning content; based on the at least one candidate recommended learning content, generating recommended learning content through the symbiotic learning platform. For more on knowledge maps, graph distances, distance thresholds, training content, candidate recommended learning content, production procedures, products, obstacle events see fig. 4, 5 and their associated descriptions.

In some embodiments, the generation module 220 may also be configured to obtain user information for each user; determining an evaluation score of each user for each candidate recommended learning content through an evaluation model based on the knowledge graph, at least one candidate recommended learning content and user information of each user; the evaluation model is a machine learning model, and the input of the evaluation model comprises a first subgraph; and using the candidate recommended learning content with the evaluation score higher than the score threshold value as the recommended learning content through the symbiotic learning platform. For more on user information, assessment scores, score thresholds, first sub-graph see fig. 4, fig. 5 and their associated description, and for more on assessment model see fig. 6 and their associated description.

The display module 230 may be configured to push and display the obstacle information and the corresponding recommended learning content to the user through the symbiotic learning platform. See fig. 1 and its associated description for more about the user, and fig. 3 and its associated description for more about the manner of pushing and presentation.

In some embodiments, knowledge recommendation system 200 may further include a second acquisition module 240, a prediction module 250, and a determination module 260.

The second obtaining module 240 may be configured to obtain a learning situation of the user on the symbiotic learning platform. For more on learning and how it is acquired see fig. 7 and its associated description.

The prediction module 250 may be configured to predict, based on the learning situation, occurrence probabilities of different obstacles corresponding to different obstacle classifications within the target time period through the obstacle prediction model; the obstacle prediction model is a machine learning model. For more on the occurrence probability of different obstacles, see fig. 7 and its related description, and for more on the obstacle prediction model, see fig. 8 and its related description.

The determination module 260 may be used to determine further training plans for the user based on the probability of occurrence of different obstacles. For more on the training program and its manner of determination see FIG. 7 and its associated description.

It should be noted that the above description of the system and its modules is for convenience of description only and is not intended to limit the present description to the scope of the illustrated embodiments. It will be appreciated by those skilled in the art that, given the principles of the system, various modules may be combined arbitrarily or a subsystem may be constructed in connection with other modules without departing from such principles. For example, the generation module 220 and the presentation module 230 may be integrated in one module. For another example, each module may share one storage device, or each module may have a respective storage device. Such variations are within the scope of the present description.

FIG. 3 is an exemplary flow chart of a knowledge recommendation method, shown in accordance with some embodiments of the present description. In some embodiments, the process 300 may be performed by the processor 140. As shown in fig. 3, the process 300 includes the steps of:

in step 310, obstacle information is obtained, the obstacle information including at least an obstacle classification. In some embodiments, step 310 may be performed by the first acquisition module 210.

The obstacle information refers to information related to a fault condition of the production space. Such as the type of fault, the location of the fault, the frequency of the same fault, the cause of the fault, etc. See fig. 1 and its associated description for more details regarding production space.

In some embodiments, the obstacle information includes at least an obstacle classification.

The obstacle classification refers to a fault class included after classifying faults in different production spaces. For example, fault classifications for process spaces may include poor raw materials, utility faults, electrical faults, machine faults, improper operation, and the like. For another example, fault classification of the experimental space may include poor experimental raw materials, equipment faults, improper operation of experimental steps, and the like.

In some embodiments, the first acquisition module 210 may acquire obstacle information of the production space from the storage device 150. In some embodiments, production space 110 may automatically send the obstacle information to storage device 150 for storage. For example, when a production space fails, production devices or monitoring devices (e.g., cameras, etc.) within the production space automatically send obstacle information to the storage device 150. In some embodiments, the staff member may upload the obstacle information by way of manual input. For example, when a production space fails, a worker checks and registers the failure, and inputs registered obstacle information to the terminal device 160, thereby uploading to the storage device 150.

Step 320, generating recommended learning content through the symbiotic learning platform based on the obstacle information. In some embodiments, step 320 may be performed by generation module 220.

The symbiotic learning platform is an automatic learning platform for processing related information of a production space and generating related learning content to help a user acquire knowledge or conduct skill training. For more on the symbiotic learning platform see fig. 1 and its related description.

The recommended learning content refers to learning content which is determined based on the characteristics of different users and is matched with the self condition of the users, and the different users correspond to different recommended learning content. For example, if a user who works in a processing space frequently fails to operate processing equipment, the recommended learning content corresponding to the user mainly includes training learning content about correct operation of the processing equipment. In some embodiments, the recommended learning content may be presented in various ways. For example, the recommended learning content may be text information, picture information, audio information, video information, or any combination.

In some embodiments, the generation module 220 may generate the recommended learning content by modeling or employing various possible data processing approaches, processing the obstacle information through a symbiotic learning platform.

In some embodiments, the generation module 220 may construct the feature vector based on the obstacle information. There are various ways of constructing the feature vector based on the obstacle information. For example, the feature vector p constructed based on the obstacle information (a, b, c, d, e, f) which may indicate that the obstacle information is from the a production space, the obstacle information contains b kinds of obstacle classifications in total, the first kind of obstacle classification is c, the occurrence probability of the failure corresponding to the obstacle classification is d, the second kind of obstacle classification is e, and the occurrence probability of the failure corresponding to the obstacle classification is e.

The storage device 150 includes a plurality of reference vectors, and recommended learning content corresponding to each of the plurality of reference vectors.

The reference vector is constructed based on the history obstacle information, and the recommended learning content corresponding to the reference vector is the recommended learning content corresponding to the history obstacle information. The vector to be matched is constructed based on the current obstacle information. The construction modes of the reference vector and the vector to be matched are referred to the characteristic vector.

In some embodiments, the generating module 220 may calculate vector distances (such as cosine distances) between the reference vector and the vector to be matched, respectively, and determine recommended learning content corresponding to the vector to be matched. For example, a reference vector whose vector distance from the vector to be matched satisfies a preset condition is taken as a target vector, and recommended learning content corresponding to the target vector is taken as recommended learning content corresponding to the vector to be matched. The preset conditions may be set according to circumstances. For example, the preset condition may be that the vector distance is minimum or that the vector distance is less than a distance threshold, or the like.

In some embodiments, the generating module 220 may determine at least one candidate recommended learning content based on the obstacle information in combination with the knowledge graph, and further generate the recommended learning content through the symbiotic learning platform based on the at least one candidate recommended learning content. For further content regarding knowledge maps, candidate recommended learning content, generating recommended learning content based on candidate recommended learning content, see fig. 4-6 and their associated descriptions.

And 330, pushing and displaying the obstacle information and the corresponding recommended learning content to the user through the symbiotic learning platform. In some embodiments, step 330 may be performed by presentation module 230.

The user refers to staff who learn or train through the symbiotic learning platform. Such as production personnel, technicians, management personnel, etc. in the production space. See fig. 1 and its associated description for more details regarding the user.

In some embodiments, presentation module 230 may push recommended learning content to a user after the user sends a request instruction to obtain recommended learning content. In some embodiments, presentation module 230 may push recommended learning content to a user in time when a production space fails. In some embodiments, presentation module 230 may push recommended learning content to the user on a regular basis (e.g., every month). In some embodiments, the presentation module 230 may present the recommended learning content to the user via the terminal device 160. In some embodiments, the presentation module 230 may simultaneously present the obstacle information of the production space corresponding to the recommended learning content to the user while pushing and presenting the recommended learning content to the user.

In some embodiments of the present disclosure, based on obstacle information, through a symbiotic learning platform, recommended learning content is generated and sent to a user, so that matched recommended learning content can be automatically provided for staff corresponding to different production spaces according to fault conditions of the different production spaces, the user is helped to acquire the learning content required by the user in time, and the working skills of the user are improved.

FIG. 4 is an exemplary flow chart for generating recommended learning content according to some embodiments of the present description. In some embodiments, the flow 400 may be performed by the processor 140 or the generation module 220. As shown in fig. 4, the process 400 includes the steps of:

step 410, determining at least one training content meeting a first preset condition with the production obstacle information based on the production obstacle information in combination with a knowledge graph, wherein the knowledge graph is related to one or more of a user, a production process, a product, an obstacle event, an obstacle classification and the training content.

The production obstacle information refers to obstacle information related to the processing space. For more about the process space, see fig. 1 and its associated description, and for more about the obstacle information, see fig. 3 and its associated description.

The knowledge graph refers to a graph constructed based on the related information of the production space and the user, reflecting the relationship between various factors in a symbolic form, and includes nodes and edges. As shown in fig. 5, the knowledge graph 510 includes nodes (e.g., node 1, etc.) and edges (e.g., edge a, etc.).

In some embodiments, the generation module 220 may generate the first sub-graph, the second sub-graph, and the third sub-graph based on the knowledge-graph. For more on the first sub-graph and the second sub-graph see fig. 6 and its associated description, and for more on the third sub-graph see fig. 8 and its associated description.

The nodes of the knowledge graph may include user nodes, production process nodes, product nodes, obstacle event nodes, obstacle classification nodes, and training content nodes.

The user nodes refer to nodes corresponding to different users, and one user corresponds to one user node. For example, node 1 shown in fig. 5 is a user node, which may be a producer of a production space. The attributes of the user node may include user information of the user. Such as the gender, age, production space to which the user belongs, the time of job entry of the user, etc. For more about the user see fig. 1 and its associated description, for more about the user information see the associated description of sub-step 421, later.

The production process nodes refer to nodes corresponding to different processes in the production process, and one production process corresponds to one production process node. The production process refers to the combination of continuous production activities of one worker (or a group of workers) on one work place (or a plurality of workers) and is a basic unit for the production process. For example, the node 2 shown in fig. 5 is a production process node, and the production process may be an assembly process of a certain component. The attributes of the production process node may include information related to the production process. For example, the type of the production process, the number of steps (i.e., steps) included in the production process, parameters of the machine equipment corresponding to the production process, and the like.

The product nodes refer to nodes corresponding to different products finally produced, and one product corresponds to one product node. The product refers to a final product obtained after being processed by each production procedure of the production space. For example, node 3 shown in fig. 5 is a product node, which may be some part of an automobile. The attributes of the product nodes may include the type of product, the yield of the product, etc.

The obstacle event node refers to a node corresponding to a specific fault at a time, and one obstacle event node corresponds to a specific fault at a time. The obstacle is a fault event affecting the normal running of the production process in the production space. For example, the node 4 shown in fig. 5 is an obstacle event node, and the obstacle event may be that a certain machine stops working. The attribute of the obstacle event node may include a parameter of the corresponding mechanical device at the time of occurrence of the fault, and the like.

The obstacle classification nodes refer to nodes corresponding to different obstacle types, and one obstacle type corresponds to one obstacle classification node. For example, node 5 shown in fig. 5 is an obstacle classification node, which may be a machine failure. The attribute of the obstacle classification node may include an obstacle type corresponding to the obstacle classification, an occurrence frequency of the obstacle classification, a hazard level of the obstacle classification, and the like. For more details regarding the classification of disorders see fig. 3 and its associated description, and for more details regarding the extent of hazard of the classification of disorders see the associated description below.

The training content nodes are nodes corresponding to different training contents, and one training content corresponds to one training content node. The training content may include detailed instructions for basic knowledge and canonical operations involved in the production process. For example, the node 6 shown in FIG. 5 is a training content node, which may be a utility usage manual. The attributes of the training content nodes may include a presentation form (e.g., text, pictures, audio, video, etc.) of the training content, a file size of the training content, a time when the training content was last updated, etc.

The edges of the knowledge graph may reflect the connection relationship between different nodes. The edges of the knowledge graph may include edges (e.g., edge a) connecting the production process node and the user node, edges (e.g., edge b) connecting the production process node and the product node, edges (e.g., edge c) connecting the obstacle event node and the production process node, edges (e.g., edge d) connecting the obstacle event node and the user node, edges (e.g., edge e) connecting the obstacle event node and the obstacle classification node, edges (e.g., edge f) connecting the obstacle classification node and the production process node, edges (e.g., edge g) connecting the training content node and the product node, edges (e.g., edge h) connecting the training content node and the obstacle classification node, and edges (e.g., edge j) connecting the training content node and the user node.

In some embodiments, the attribute of the edge connecting the training content node and the user node may include whether the user has learned the training content. In some embodiments, the attributes of the edge connecting the training content node and the user node may further include a learning time used by the user to learn the training content. In some embodiments, the attributes of the edges connecting the training content node and the user node may further include a score of the user after learning the training content.

The first preset condition refers to a condition for screening training contents satisfying the requirement. In some embodiments, the first preset condition may be that a graph distance from the production barrier information is less than a distance threshold.

The graph distance may refer to an attribute of an edge in the knowledge graph that connects the obstacle classification node and the training content node. In some embodiments, the attributes of the edges connecting the obstacle classification node and the training content node include a degree of correlation.

The correlation degree refers to the correlation degree between the obstacle classification node and the training content node, which are connected by the edges. The degree of correlation can be expressed by a percentage, and the larger the percentage is, the higher the degree of correlation is. For example, if the degree of correlation between the obstacle classification node a, which is a utility obstacle, and the training content node B, which is a utility instruction manual, is 80%, the degree of correlation between the side connecting the obstacle classification node a and the training content node B is 80%; for another example, if the degree of correlation between the obstacle classification node a, which is an obstacle type of the utility obstacle, and the training content node C, which is a training content of the utility maintenance skill, is 90%, the degree of correlation between the sides connecting the obstacle classification node a and the training content node C is 90%; for another example, if the degree of correlation between the obstacle classification node a, whose type is a utility fault, and the training content node D, whose raw material component and property are the training content, is 10%, the degree of correlation between the side connecting the obstacle classification node a and the training content node D is 10%.

In some embodiments, the relevance may be annotated by manual presets. For example, the user may manually annotate the relevance based on expert advice or prior experience. In some embodiments, the relevance may be determined based on a frequent item set.

The frequent item set refers to a set with a support degree equal to or greater than a minimum support degree (min_sup), wherein the support degree refers to the frequency of occurrence of a certain set in all events. Determining the relevance based on the frequent item set may be achieved by the following sub-steps S1-S3:

s1: a historical training record of the symbiotic learning platform 130 is obtained, including historical obstacle classifications and their corresponding historical training content, where each obstacle classification may correspond to a plurality of training content. For example, the historical obstacle classification 1 may correspond to the historical training content a, the historical training content B, the historical training content C; the historical obstacle classification 2 may correspond to historical training content B; the historical obstacle classification 3 may correspond to the historical training content a, the historical training content D, the historical training content E.

S2: based on the historical training records, the support degree of each obstacle classification and each training content is calculated. In some embodiments, the support of each obstacle classification with each training content may be calculated by the following equation (1):

Wherein X represents obstacle classification, and Y represents training content.

For example, there are 1000 pieces of history training records in which the number of times that history obstacle class 1 and history training content a co-occur is 60, and the degree of support of history obstacle class 1 and history training content a is 6%.

S3: and taking the support degree of each obstacle classification and each training content as the correlation degree of the edge connecting the obstacle classification node and the training content node. For example, if the support degree between the history obstacle classification 1 and the history training content a is 6%, the correlation degree between the side connecting the obstacle classification node corresponding to the obstacle classification 1 and the training content node corresponding to the history training content a is 6%.

In some embodiments of the present disclosure, the degree of correlation of the edges is introduced when determining the distance of the graph, and the degree of correlation between the obstacle classification nodes and the training content nodes connected by the edges is further considered, so that the determined distance of the graph is more in line with the actual situation, and a large amount of historical data is introduced when determining the degree of correlation of the edges based on the frequent item set, thereby effectively improving the accuracy and scientificity of the calculated degree of correlation.

In some embodiments, the graph distance may also include other attributes of the edges of the knowledge graph that connect the obstacle classification node and the training content node. In some embodiments, other attributes of the edges connecting the obstacle classification node and the training content node include a priority.

In some embodiments, one obstacle classification node may be coupled to a plurality of training content nodes. The priority refers to the priority learning degree of the training content corresponding to the obstacle classification. The priority may be expressed by a percentage, the higher the priority, the more preferably the training content is learned. For example, the obstacle classification node 1 is connected to the training content node A, B, C, the side connecting the obstacle classification node 1 to the training content node a is x, the side connecting the obstacle classification node 1 to the training content node B is y, and the side connecting the obstacle classification node 1 to the training content node C is z. When a user performs training learning, firstly learning training content a corresponding to a training content node A, wherein the priority learning degree of the training content a is 90%; training content B corresponding to the training content node B is learned again, and the preferential learning degree of the training content B is 70%; and finally, learning the training content C corresponding to the training content node C, wherein the priority learning degree of the training content C is 50%. Then, the priority of side a is 90%, the priority of side b is 70%, and the priority of side c is 50%.

In some embodiments, the priorities may be annotated by manual presets. For example, the user can judge the preferential learning degree of different training contents based on expert advice or priori experience, and then manually marks the preferential degree.

In some embodiments of the present disclosure, priority is introduced when determining the graph distance, and the priority learning degree of each training content can be determined in combination with the specific situation of a certain obstacle classification under the condition that a certain obstacle classification corresponds to a plurality of training contents, so as to help a user to formulate a reasonable learning sequence, and to preferentially train the content with higher priority, so as to improve the accuracy of the graph distance.

In some embodiments, graph distance may be determined based on relevance and priority. In some embodiments, the generation module 220 may determine the graph distance by weighted summation based on the relevance and the priority.

The weights may include a weight of the degree of correlation and a weight of the degree of priority. In some embodiments, the weights may be annotated by manual presets. For example, the user may manually label weights based on expert advice or prior experience, e.g., for a certain edge connecting the obstacle classification node and the training content node, the weight of the relevance is manually labeled as 0.4, and the weight of the priority is labeled as 0.6.

In some embodiments, the weight of the priority is related to the hazard level of the obstacle classification corresponding to the obstacle classification node in the nodes to which the graph is connected. The hazard level of the obstacle classification refers to the magnitude of the hazard of the type of fault to the effect of production of the production space. For example, if a user operates a machine improperly during a production process, the user may be severely injured, and the poor raw material does not pose a threat to the personal safety of the user, for example, the obstacle classification of the type of obstacle being improperly operated and the obstacle classification of the type of obstacle being poor raw material are more dangerous than the obstacle classification of the type of obstacle being poor raw material. The greater the degree of hazard of the obstacle classification, the greater the weight of the priority of the edge connecting the obstacle classification nodes.

In some embodiments, the generation module 220 may determine the graph distance, which may be the inverse of the calculation, by weighted summation based on the relevance and the priority. For example, the obstacle classification nodes 1 are connected to the training content nodes A, B, respectively. The side connecting the obstacle classification node 1 and the training content node A is x, and the graph distance is p; the side connecting the obstacle classification node 1 and the training content node B is y, and the graph distance is q. Wherein, the correlation of the side x is 20%, the weight of the correlation is 0.1, the priority is 80%, and the weight of the priority is 0.7, and the graph distance p is 1/(20%. 0.1+80%. 0.7) =50/29; the correlation of the side y is 30%, the weight of the correlation is 0.4, the priority is 60%, and the weight of the priority is 0.6, and the graph distance q is 1/(30% 0.4+60% 0.6) =25/12.

In some embodiments of the present disclosure, the graph distance is determined by weighted summation based on the correlation degree and the priority degree, and the weight is determined in combination with the hazard degree of the obstacle classification, so that the weight can be set in a targeted manner according to the specific situations of different obstacle types, and the accuracy and the rationality of the graph distance are further ensured.

The distance threshold refers to a specific graph distance value used to screen training content. For example, the distance threshold may be 2. In some embodiments, the distance threshold may be annotated by a manual preset. For example, the manager may manually annotate the distance threshold based on expert advice or prior experience. In some embodiments, the distance threshold may be automatically determined by the symbiotic learning platform 130. For example, the symbiotic learning platform 130 can automatically determine the distance threshold based on the number of training content.

The training content refers to relevant data for training and learning by the user. Different obstacle classifications may correspond to different training content, for example, the obstacle type is an electrical fault, and the corresponding training content is an electrical safety training manual; the type of obstacle is a utility fault, and the corresponding training content is a utility instruction manual. The same obstacle classification may correspond to a plurality of training contents, for example, an obstacle classification in which the obstacle type is an electrical fault, and the corresponding training contents may include an electrical safety training manual, electrical basic knowledge, electrical equipment operation training contents, and the like.

And step 420, taking at least one training content as at least one candidate recommended learning content.

The candidate recommended learning content refers to training content in which the distance of the graph is smaller than the distance threshold. For example, the training content whose graph distance is smaller than the distance threshold is an electric safety training manual, and the electric safety training manual is taken as a candidate recommended learning content.

At step 430, recommended learning content is generated by the symbiotic learning platform based on the at least one candidate recommended learning content.

In some embodiments, generating recommended learning content by the symbiotic learning platform based on at least one candidate recommended learning content may be accomplished by sub-steps 431-433.

In sub-step 431, user information of each user is acquired.

The user information refers to information related to the user's behavior. In some embodiments, the user information may be an attribute of the user node. For example, the user information may include a job number of the user, a production space to which the user belongs, a number of times the user has caused a malfunction, a number of times the user has handled the malfunction, a time period for the user to enter, a history of the user's learning through the symbiotic learning platform, and the like.

In some embodiments, the generation module 220 may obtain user information for the user from an employee information database of the enterprise and/or the storage device 150 of the symbiotic learning platform 130. In some embodiments, the user information of the user may be updated in real-time or periodically (e.g., every month).

In a substep 432, based on the knowledge graph, the at least one candidate recommended learning content, and the user information of each user, an evaluation score of each user for each candidate recommended learning content in the at least one candidate recommended learning content is determined by an evaluation model, where the evaluation model is a machine learning model.

The evaluation score refers to a score obtained by predicting the grasp condition of the user after learning the candidate recommended learning content and evaluating and scoring the grasp condition. In some embodiments, the score after the user learns the candidate recommended learning content may be used as the evaluation score. For example, after the user learns the candidate recommended learning content, the related problem is completed, and the score is 80, and the evaluation score of the user for the candidate recommended learning content is 80.

In some embodiments, the generation module 220 may determine the evaluation score of each user for each of the at least one candidate recommended learning content by processing the knowledge-graph, the at least one candidate recommended learning content, and the user information for each user through an evaluation model. For more on the assessment model, determining an assessment score based on the assessment model see fig. 6 and its associated description.

In step 433, candidate recommended learning content whose evaluation score satisfies the second preset condition is used as recommended learning content by the symbiotic learning platform.

The second preset condition refers to a condition for screening candidate recommended learning content meeting the requirement. In some embodiments, the second preset condition may be that the evaluation score is above a score threshold.

The score threshold refers to a value of a specific evaluation score used to filter candidate recommended learning content. For example, the score threshold may be 80 points. In some embodiments, the score threshold may be annotated by a manual preset. For example, the user may manually annotate the score threshold based on expert advice or prior experience. In some embodiments, the score threshold may be automatically determined by the symbiotic learning platform 130. For example, the symbiotic learning platform 130 can automatically determine the distance threshold based on the number of candidate recommended learning content.

In some embodiments of the present disclosure, the training content is initially screened based on the graph distance, and then the recommended learning content is further determined through the evaluation model based on the candidate recommended learning content obtained after screening, so that excessive data for prediction can be avoided, the subsequent calculation is too complex, the operation load of the symbiotic learning platform is reduced, and the data processing efficiency of the symbiotic learning platform is improved.

FIG. 6 is an exemplary schematic diagram of an assessment model shown in accordance with some embodiments of the present description.

In some embodiments, the generation module 220 may determine, via the evaluation model, an evaluation score for each user for each of the at least one candidate recommended learning content based on the knowledge-graph, the at least one candidate recommended learning content, and the user information for each user.

The evaluation model refers to a model for predicting an evaluation score of each user for each of at least one candidate recommended learning content.

In some embodiments, the assessment model 650 may process the first sub-graph 610, the second sub-graph 620, the user information 630, and the at least one candidate recommendation learning content 640, obtaining an assessment score 660 for each candidate recommendation learning content for the user.

In some embodiments, the evaluation model 650 may be a graph neural network (GraphNeural Network, GNN).

As shown in fig. 6, the inputs to the assessment model 650 may include a first sub-graph 610, a second sub-graph 620, user information 630, and at least one candidate recommendation learning content 640.

The first subgraph is generated based on the knowledge graph and corresponds to the graph of each user. In some embodiments, the first sub-graph may be generated based on a correlation graph structure of a certain user node included in the knowledge graph, and obstacle classification nodes and training content nodes related to the user node. For example, as shown in fig. 5, the first sub-graph 520 is a graph corresponding to a certain user, the node corresponding to the user is a user node 1, and the first sub-graph 520 is generated based on the user node 1 in the knowledge graph 510 and the related graph structures of the obstacle classification node and the training content node related to the user node. The first sub-graph 610 of the input assessment model 650 can be the first sub-graph 520 shown in FIG. 5.

The second sub-graph is a graph formed by adding related obstacle event nodes and three-degree inner adjacent edges of the user in a period of time on the basis of the first sub-graph. The three-degree inner adjacent edges refer to nodes connected by three continuous edges connected with different nodes. For example, as shown in fig. 5, the "user node 1-obstacle event node 8-obstacle classification node 9-training content node 10" in the second sub-graph 530, where each node is different from the other, and includes three consecutive edges p, q, r, then the three edges and the node connected thereto are three-degree adjacent edges. In some embodiments, the second sub-graph may be generated based on the first sub-graph and obstacle events that occur within a period of time (e.g., within a month) of the user to which the first sub-graph corresponds. For example, as shown in fig. 5, the second sub-graph 530 is obtained by adding the maps generated by the obstacle event node 7 and the obstacle event node 8 on the basis of the first sub-graph 520, where the obstacle event corresponding to the obstacle event node 7 and the obstacle event corresponding to the obstacle event node 8 are the obstacle events that occur within one month for the user, and the obstacle event corresponding to the obstacle event node 7 occurs twice. The second sub-graph 620 of the input assessment model 650 can be the second sub-graph 530 shown in FIG. 5.

For more on knowledge graph, nodes and edges of the first sub-graph, nodes and edges of the second sub-graph, user nodes, obstacle classification nodes, training content nodes, obstacle event nodes, see fig. 4 and its associated description.

In some embodiments of the present disclosure, a first sub-graph corresponding to a user in a knowledge graph is generated based on the user, so that content related to the user can be screened out, thereby simplifying the knowledge graph, avoiding overlarge data input into an evaluation model, and improving data processing capability and speed of the evaluation model.

The input user information 630 is user information of the user corresponding to the input first sub-graph 610, for example, an job-in duration of the user corresponding to the first sub-graph 610. The at least one candidate recommended learning content 640 entered into the assessment model 650 may be an electrical safety training manual. For more on user information, recommended learning content see fig. 4 and its associated description.

The output of the assessment model 650 may include an assessment score 660 for each candidate recommended learning content by the user. For example, the output evaluation score of the user for the electric safety training manual is 70 points. See fig. 4 and its associated description for more details regarding evaluation scores.

In some embodiments, the assessment model 650 may be trained solely based on historical data. In some embodiments, the evaluation model 650 may be trained based on multiple sets of first training samples and first tags.

In some embodiments, the first training sample of the training evaluation model includes a sample first sub-graph, a sample second sub-graph, sample user information, and at least one sample candidate recommended learning content, and the first label corresponding to the first training sample is an evaluation score of the user for each sample candidate recommended learning content. The first training sample may be obtained based on historical data, the first label may be determined based on a historical score when the user learns the recommended learning content in the past, and the first label may be determined by means of manual labeling or automatic labeling. The above description is by way of example only and not limitation, and the tags of the training data may be obtained in various ways.

During training, a first training sample is input into an initial evaluation model, a loss function is constructed based on the output of the initial evaluation model and a first label, and parameters of the initial evaluation model are updated through the loss function until the trained initial evaluation model meets a preset condition, and a trained evaluation model 650 is obtained, wherein the preset condition can be that the loss function is smaller than a threshold value, convergence is achieved, or a training period reaches the threshold value.

In some embodiments of the present disclosure, the evaluation score of each user for each candidate recommended learning content is obtained through an evaluation model, so that the mastering situation of each user for each candidate recommended learning content can be predicted in advance, and further, the recommended learning content that is more matched with the user can be determined according to the actual situation of different users.

FIG. 7 is an exemplary flow chart for determining a further training program for a user, according to some embodiments of the present description. In some embodiments, the flow 700 may be performed by the processor 140. As shown in fig. 7, the process 700 includes the steps of:

step 710, obtaining the learning condition of the user on the symbiotic learning platform. In some embodiments, step 710 may be performed by the second acquisition module 240.

The learning situation refers to the relevant situation that the user learns through the symbiotic learning platform. For example, the learning situation may include learning content of the user, learning duration, learning effect, interval time between learning the same content twice, and the like.

In some embodiments, the second acquisition module 240 may acquire the learning situation of the user through the storage device 150 of the symbiotic learning platform 130. In some embodiments, the learning situation of the user may be updated in real-time or periodically (e.g., every month).

Step 720, based on the learning condition, predicting occurrence probabilities of different obstacles corresponding to different obstacle classifications in a target time period through an obstacle prediction model; the obstacle prediction model is a machine learning model. In some embodiments, step 720 may be performed by prediction module 250.

The target period refers to a future certain period set in advance. For example, a month in the future after the last learning of the user is completed may be taken as the target period.

The occurrence probability of the different obstacle refers to the probability of occurrence of the different obstacle within the target period. The occurrence probability may be expressed by a percentage, for example, the obstacle classification may be an electrical fault, and the occurrence probability of the electrical fault within the target period may be 90%.

In some embodiments, the prediction module 250 may process the learning situation of the user through the obstacle prediction model, and obtain the occurrence probabilities of different obstacles corresponding to different obstacle classifications in the target time period. For more details regarding the obstacle prediction model, see fig. 8 and its associated description.

Step 730, determining a further training program for the user based on the probability of occurrence of the different obstacles. In some embodiments, step 730 may be performed by determination module 260.

Further training plans refer to subsequent learning plans for the user that are determined based on the probability of occurrence of different obstacles. For example, the further training program may learn training contents corresponding to the obstacle classification to which the obstacle having a high occurrence probability belongs.

In some embodiments, the determination module 260 regenerates the supplemental recommended learning content for the user in response to the probability of occurrence of the obstacle being greater than the probability threshold.

The supplementary recommended learning content refers to training content related to an obstacle whose occurrence probability is greater than a probability threshold. For more content on recommended learning content see fig. 3 and its associated description, for more content on training content see fig. 4 and its associated description.

The probability threshold value refers to a value for determining whether or not it is necessary to regenerate a specific obstacle occurrence probability of the supplementary recommended learning content. For example, the probability threshold may be 80%. In some embodiments, the probability threshold may be annotated by a manual preset. For example, the user may manually annotate the probability threshold based on expert advice or prior experience. In some embodiments, the probability threshold may be automatically determined by the symbiotic learning platform 130. For example, the symbiotic learning platform 130 can automatically determine the probability threshold based on the probability of occurrence of the historical obstacle.

The method for generating the supplementary recommended learning content is the same as the method for generating the recommended learning content described in fig. 4 to 6, and will not be described again here.

In some embodiments of the present disclosure, based on the learning situation of the user in the symbiotic learning platform and the occurrence probability of different obstacles in the target time period, determining a further training plan of the user can purposefully help the user learn or review relevant knowledge in time, so as to correctly cope with the obstacles or prevent the obstacles from occurring, and ensure orderly progress of the production flow of the production space.

It should be noted that the above description of the

flows

300, 400, 700 is for illustration and description only, and is not intended to limit the scope of applicability of the present description. Various modifications and changes to the

processes

300, 400, 700 may be made by those skilled in the art under the guidance of this specification. However, such modifications and variations are still within the scope of the present description.

In some embodiments, the prediction module 250 may process the learning situation of the user through the obstacle prediction model, and obtain the occurrence probabilities of different obstacles corresponding to different obstacle classifications in the target time period.

The obstacle prediction model is a model for predicting occurrence probabilities of different obstacles corresponding to different obstacle classifications in a target period.

In some embodiments, obstacle prediction model 850 may process third sub-graph 810, environment information 820, device information 830, and target time period 840 to obtain probabilities of occurrence 860 of different obstacles.

In some embodiments, the obstacle prediction model 850 may be a graph neural network (Graph Neural Network, GNN).

As shown in fig. 8, inputs to obstacle prediction model 850 may include third sub-graph 810, environment information 820, device information 830, and target time period 840.

The third sub-graph refers to a graph generated based on the knowledge graph, which reflects the learning situation of a certain user. In some embodiments, the third subgraph may be generated based on a related graph structure of user nodes, production process nodes, obstacle classification nodes, and training content nodes contained in the knowledge graph. For example, the third sub-graph 540 shown in fig. 5 is a graph reflecting the learning situation of the user corresponding to the user node 1. The attribute of the edge connecting the user node and the training content node may include a learning condition of the user learning the training content, for example, whether the user learns the training content, a time used by the user to learn the training content, a score of the user learning the training content, and the like. The third sub-graph 810 input the obstacle prediction model 850 may be the third sub-graph 540 shown in fig. 5. For more on knowledge graph, edges and nodes of the third sub-graph, user nodes, production process nodes, obstacle classification nodes, training content nodes see FIG. 4 and its associated description.

The environmental information refers to related information of the production space environment. Such as temperature, air pressure, etc. within the production space. In some embodiments, the environmental information may be obtained by various detection devices. For example, the temperature of the production space may be obtained by a thermometer; for another example, the air pressure of the production space may be obtained by an air pressure sensor. In some embodiments, the detection device may upload the acquired environmental information to the storage device 150 in real-time or periodically (e.g., every month). The environmental information 820 input to the obstacle prediction model 850 may be the temperature of the production space.

The device information refers to related information of different devices. Such as the model number of each type of device, the age of the device, etc. In some embodiments, the device information may be entered into the terminal device 160 manually and uploaded to the storage device 150 in real-time or periodically (e.g., every month). For example, when a worker periodically checks the device, the worker records the device information, inputs the device information into the terminal device 160, and uploads the device information to the storage device 150. The device information 830 input to the obstacle prediction model 850 may be a life span of the device.

The target period 840 of time for which the obstacle prediction model 850 is input may be one month after the user corresponding to the third sub-graph 810 finishes the last learning. For more on the target time period see fig. 7 and its associated description.

In some embodiments of the present disclosure, a third sub-graph corresponding to a user in the knowledge graph is generated, and content related to learning conditions of the user may be screened out, so that the knowledge graph is simplified, excessive data input into the obstacle prediction model is avoided, and data processing capability and speed of the obstacle prediction model are improved.

The output of the obstacle prediction model 850 may include the probability of occurrence 860 of different obstacles. For example, the probability of occurrence of a certain obstacle of the output may be 80%. In some embodiments, the probability of occurrence of the different obstacle may be based on the obstacle classification node output in the third subgraph.

In some embodiments, obstacle prediction model 850 may be trained alone based on historical data. In some embodiments, the obstacle prediction model 850 may be trained based on multiple sets of second training samples and second labels.

In some embodiments, the second training sample of the training obstacle prediction model includes a sample third sub-graph, sample environment information, sample equipment information, and a sample target time period, and the second label corresponding to the second training sample is whether different obstacles corresponding to different obstacle classifications in the third sub-graph occur in the target time period, if so, the label is 1, and if not, the label is 0. The second training sample can be obtained based on historical data, and the second label can be determined by means of manual labeling or automatic labeling. The above description is by way of example only and not limitation, and the tags of the training data may be obtained in various ways.

During training, the second training sample is input into the initial obstacle prediction model, a loss function is constructed based on the output of the initial obstacle prediction model and the second label, and parameters of the initial obstacle prediction model are updated through the loss function until the trained initial obstacle prediction model meets the preset condition, and the trained obstacle prediction model 850 is obtained, wherein the preset condition can be that the loss function is smaller than a threshold value, convergence is achieved, or the training period reaches the threshold value and the like.

In some embodiments of the present disclosure, the occurrence probability of different obstacles is predicted by processing the third subgraph, the environmental information, the device information, and the target time period through the obstacle prediction model, and the occurrence probability of different obstacles can be predicted by combining the historical obstacle situation and the actual situation in the production space, so that the accuracy of the prediction result is improved.

While the basic concepts have been described above, it will be apparent to those skilled in the art that the foregoing detailed disclosure is by way of example only and is not intended to be limiting. Although not explicitly described herein, various modifications, improvements, and adaptations to the present disclosure may occur to one skilled in the art. Such modifications, improvements, and modifications are intended to be suggested within this specification, and therefore, such modifications, improvements, and modifications are intended to be included within the spirit and scope of the exemplary embodiments of the present invention.

Meanwhile, the specification uses specific words to describe the embodiments of the specification. Reference to "one embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the present description. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the present description may be combined as suitable.

Furthermore, the order in which the elements and sequences are processed, the use of numerical letters, or other designations in the description are not intended to limit the order in which the processes and methods of the description are performed unless explicitly recited in the claims. While certain presently useful inventive embodiments have been discussed in the foregoing disclosure, by way of various examples, it is to be understood that such details are merely illustrative and that the appended claims are not limited to the disclosed embodiments, but, on the contrary, are intended to cover all modifications and equivalent arrangements included within the spirit and scope of the embodiments of the present disclosure. For example, while the system components described above may be implemented by hardware devices, they may also be implemented solely by software solutions, such as installing the described system on an existing server or mobile device.

Likewise, it should be noted that in order to simplify the presentation disclosed in this specification and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof. This method of disclosure, however, is not intended to imply that more features than are presented in the claims are required for the present description. Indeed, less than all of the features of a single embodiment disclosed above.

In some embodiments, numbers describing the components, number of attributes are used, it being understood that such numbers being used in the description of embodiments are modified in some examples by the modifier "about," approximately, "or" substantially. Unless otherwise indicated, "about," "approximately," or "substantially" indicate that the number allows for a 20% variation. Accordingly, in some embodiments, numerical parameters set forth in the specification and claims are approximations that may vary depending upon the desired properties sought to be obtained by the individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and employ a method for preserving the general number of digits. Although the numerical ranges and parameters set forth herein are approximations that may be employed in some embodiments to confirm the breadth of the range, in particular embodiments, the setting of such numerical values is as precise as possible.

Each patent, patent application publication, and other material, such as articles, books, specifications, publications, documents, etc., referred to in this specification is incorporated herein by reference in its entirety. Except for application history documents that are inconsistent or conflicting with the content of this specification, documents that are currently or later attached to this specification in which the broadest scope of the claims to this specification is limited are also. It is noted that, if the description, definition, and/or use of a term in an attached material in this specification does not conform to or conflict with what is described in this specification, the description, definition, and/or use of the term in this specification controls.

Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments of this specification. Other variations are possible within the scope of this description. Thus, by way of example, and not limitation, alternative configurations of embodiments of the present specification may be considered as consistent with the teachings of the present specification. Accordingly, the embodiments of the present specification are not limited to only the embodiments explicitly described and depicted in the present specification.

Claims

1. A knowledge recommendation method executed by a processor based symbiotic learning platform, the method comprising:

Acquiring obstacle information, wherein the obstacle information at least comprises obstacle classification;

generating recommended learning content through the symbiotic learning platform based on the obstacle information;

and pushing and displaying the obstacle information and the corresponding recommended learning content to the user through the symbiotic learning platform.

2. The method of claim 1, wherein the type of obstacle information comprises production obstacle information;

the generating, based on the obstacle information, recommended learning content by the symbiotic learning platform includes:

determining at least one training content meeting a first preset condition with the production disorder information based on the production disorder information and combining a knowledge graph; the knowledge graph is related to one or more of the user, a production process, a product, an obstacle event, the obstacle classification, and the training content;

taking the at least one training content as at least one candidate recommended learning content;

and generating the recommended learning content through the symbiotic learning platform based on the at least one candidate recommended learning content.

3. The method of claim 2, wherein the generating, by the symbiotic learning platform, the recommended learning content based on the at least one candidate recommended learning content comprises:

Acquiring user information of each user;

determining an evaluation score of each of the users for each of the at least one candidate recommended learning content by an evaluation model based on the knowledge graph, the at least one candidate recommended learning content, and the user information of each of the users; the evaluation model is a machine learning model;

and taking the candidate recommended learning content with the evaluation score meeting a second preset condition as the recommended learning content through the symbiotic learning platform.

4. The method according to claim 1, wherein the method further comprises:

acquiring the learning condition of the user on the symbiotic learning platform;

based on the learning condition, predicting occurrence probabilities of different barriers corresponding to different barrier classifications in a target time period through a barrier prediction model; the obstacle prediction model is a machine learning model;

based on the probability of occurrence of the different obstacle, a further training plan for the user is determined.

5. A knowledge recommendation system comprising a symbiotic learning platform, the system further comprising:

The first acquisition module is used for acquiring obstacle information, wherein the obstacle information at least comprises obstacle classification;

the generation module is used for generating recommended learning content through the symbiotic learning platform based on the obstacle information;

and the display module is used for pushing and displaying the obstacle information and the recommended learning content corresponding to the obstacle information to a user through the symbiotic learning platform.

6. The system of claim 5, wherein the type of obstacle information comprises production obstacle information, the generation module further configured to:

7. The system of claim 6, wherein the generation module is further configured to:

Acquiring user information of each user;

8. The system of claim 5, wherein the system further comprises:

the second acquisition module is used for acquiring the learning condition of the user on the symbiotic learning platform;

the prediction module is used for predicting the occurrence probability of different barriers corresponding to different barrier classifications in a target time period through a barrier prediction model based on the learning condition; the obstacle prediction model is a machine learning model;

a determination module for determining a further training plan for the user based on the probability of occurrence of the different obstacle.

9. A knowledge recommendation device, the device comprising at least one processor and at least one memory;

The at least one memory is configured to store computer instructions;

the at least one processor is configured to execute at least some of the computer instructions to implement the knowledge recommendation method of any one of claims 1-4.

10. A computer readable storage medium storing computer instructions which, when executed by a processor, implement the knowledge recommendation method of any one of claims 1 to 4.