Training decision method, apparatus, device, medium, and product

Abstract

This application provides a training decision-making method, apparatus, equipment, medium, and product, relating to the fintech field or other related fields. It includes: acquiring multimodal data; the multimodal data includes development performance data, job knowledge graph data, and employee behavior data; performing natural language processing and graph neural network processing on the multimodal data to obtain a competency gap profile and an organizational competency report; generating a training path based on the competency gap profile and organizational competency report; and conducting training for trainees based on the training path. The solution of this application improves the reliability of training results.

Description

Technical Field

[0001] This application relates to the field of financial technology or other related fields, and in particular to a training decision-making method, apparatus, equipment, medium and product. Background Technology

[0002] With the rapid development of fintech, the banking industry is undergoing a critical period of digital transformation. Traditional banking business models are facing the impact of emerging technologies such as internet finance, mobile payments, and artificial intelligence. Bank personnel need to master new technical knowledge and skills to adapt to the development needs of the digital age. At the same time, with constantly updated regulatory policies and increasingly diversified customer needs, bank personnel need to improve their comprehensive capabilities to better serve customers, prevent risks, and promote business innovation.

[0003] Currently, training for bank staff often requires analyzing multiple heterogeneous data sources to generate corresponding training plans. Traditional training plans struggle to effectively integrate this data, making it difficult to accurately identify training needs and resulting in poor reliability of training outcomes. Summary of the Invention

[0004] This application provides a training decision-making method, apparatus, equipment, medium, and product to improve the reliability of training results.

[0005] Firstly, this application provides a training decision-making method, including:

[0006] Acquire multimodal data; the multimodal data includes development performance data, job knowledge graph data, and employee behavior data;

[0007] The multimodal data is processed using natural language processing and graph neural network processing to obtain a capability gap profile and an organizational capability report;

[0008] Based on the capability gap profile and the organizational capability report, a training path is generated; and based on the training path, the trainees are trained.

[0009] Secondly, this application provides a training decision-making device, comprising:

[0010] The acquisition module is used to acquire multimodal data, including development performance data, job knowledge graph data, and employee behavior data.

[0011] The processing module is used to perform natural language processing and graph neural network processing on the multimodal data to obtain a capability gap profile and an organizational capability report;

[0012] The generation module is used to generate training paths based on the capability gap profile and the organizational capability report; and to train the trainees based on the training paths.

[0013] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0014] The memory stores computer-executed instructions;

[0015] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0017] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0018] The training decision-making method, apparatus, equipment, media, and products provided in this application acquire multimodal data, including development efficiency data, job knowledge graph data, and employee behavior data. The multimodal data undergoes natural language processing and graph neural network processing to obtain a competency gap profile and an organizational competency report. Based on the competency gap profile and organizational competency report, a training path is generated. Training is then conducted on the trainees based on the training path. This application's solution, through joint analysis of multimodal data, obtains an organizational competency report containing the competency gaps of the trainees and corresponding training suggestions, thus improving the accuracy of competency analysis and training needs analysis. The generation of training paths based on accurate competency gap profiles and organizational competency reports makes the generated training paths more aligned with current training needs, improving the reliability of the generated training paths. Furthermore, training is conducted on the trainees based on the training paths, further enhancing the reliability of the training results. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0020] Figure 1 An exemplary flowchart of a training decision-making method is shown. Figure 1 ;

[0021] Figure 2 An exemplary flowchart of a training decision-making method is shown. Figure 2 ;

[0022] Figure 3 An exemplary schematic diagram of a training decision-making device is shown;

[0023] Figure 4 The diagram above illustrates the structure of an electronic device.

[0024] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0025] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0026] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning. The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar or related objects or entities, and do not necessarily mean to limit a specific order or sequence, unless otherwise indicated. It should be understood that such terms can be used interchangeably where appropriate, for example, to be implemented in an order other than those given in the illustrations or descriptions of the embodiments of this application. The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application, are intended to be omnipresent but not exclusive. For example, a product or device that comprises a series of components is not necessarily limited to those components that are explicitly listed, but may include other components that are not explicitly listed or are inherent to such products or devices. The term "module" as used in this application refers to any known or subsequently developed hardware, software, firmware, artificial intelligence, fuzzy logic, or combination of hardware and / or software code capable of performing the functions associated with that element.

[0027] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize or refuse.

[0028] Furthermore, the technical solution involved in this application, which involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.) and the use of artificial intelligence technology for automated decision-making, and makes decisions that have a significant impact on personal rights based on the results of automated decision-making, provides users with corresponding operation entry points for users to choose to agree to or reject the results of automated decision-making; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0029] It should be noted that the training decision-making methods, devices, equipment, media and products provided in this application can be used in the field of fintech, or in any field other than fintech. The application fields of the training decision-making methods, devices, equipment, media and products in this application are not limited.

[0030] With the rapid development of fintech, the banking industry faces the dual challenges of digital transformation and accelerated technological iteration. Banks have a large number of heterogeneous data sources, including development efficiency data (such as code commit quality and production event logs), structured business data (such as job descriptions and policy libraries), and unstructured behavioral data (such as employee learning records and interview transcripts). Traditional training systems struggle to effectively integrate this data, making it difficult to accurately identify training needs. For example, code quality defects among development center programmers are not correlated with courses; branch employees lack targeted training resources when dealing with new regulatory technology rules; and the skills gap between head office and frontline employees cannot be dynamically tracked. Furthermore, bank training needs to adapt to the differentiated scenarios across multiple levels—head office, branches, and development centers. For instance, cloud platform development engineers need to master fault diagnosis skills, while tellers need to enhance their intelligent customer service interaction capabilities. Current technical solutions cannot build a closed-loop system covering the entire "learning-application-evaluation" chain, making it difficult to quantify and verify training effectiveness.

[0031] Currently, the training management of banking institutions generally adopts the following model:

[0032] 1. Demand gathering methods: relying on manual questionnaires and interviews

[0033] 2. Course Development Process: Instructors manually create PPTs / lecture notes.

[0034] 3. Teacher matching mechanism: resume screening + trial lecture evaluation

[0035] 4. Training Implementation Format: Standardized face-to-face instruction or online pre-recorded lectures

[0036] 5. Effectiveness Evaluation System: Post-lesson Satisfaction Score

[0037] The disadvantages of the above model include:

[0038] 1. Delayed demand insight

[0039] Manual surveys struggle to quantify dynamic skill gaps across different roles and levels. They also fail to link performance data with training needs (e.g., the lack of correlation between code quality defects among programmers in development centers and course content).

[0040] 2. Inefficient content production

[0041] Lecturers spend 60% of their time creating courseware, and the update speed lags behind technological evolution.

[0042] 3. Resource mismatch

[0043] The selection of high-quality instructors relies on subjective experience, and internal trainers lack empowerment tools. Standardized courses cannot adapt to the differentiated scenarios of "head office-branch-development center".

[0044] 4. The closed-loop effect is broken.

[0045] Training is disconnected from job competency development, and a "learning-application-assessment" data chain has not been established.

[0046] The above solutions fail to accurately identify training needs, resulting in poor reliability of training outcomes.

[0047] The training decision-making methods, devices, equipment, media, and products provided in this application, through joint analysis of multimodal data, generate an organizational capability report that provides trainees' competency gaps and corresponding training recommendations; this improves the accuracy of competency analysis and training needs analysis. Training paths are generated based on accurate competency gap profiles and organizational capability reports, making the generated training paths more aligned with current training needs; this improves the reliability of the generated training paths. Furthermore, training is conducted on trainees based on these training paths, further enhancing the reliability of training outcomes.

[0048] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0049] Example 1

[0050] Figure 1 An exemplary flowchart of a training decision-making method is shown. Figure 1 ;like Figure 1 As shown, the method includes:

[0051] Step 101: Obtain multimodal data; multimodal data includes development performance data, job knowledge graph data, and employee behavior data.

[0052] In this embodiment, development performance data may include code commit quality and production event logs; whereby code commit quality refers to the evaluation of code quality derived through quantitative analysis. This evaluation mainly includes the following dimensions:

[0053] Code style compliance: Does the code conform to the team's agreed-upon style guidelines (such as naming conventions, comments, and formatting)? This is typically checked by automated tools in merge requests, which provide scores or reports. Numerous style violations in a submission will lower the quality score. Code complexity: Is the code structure clear, easy to understand, and maintain? Overly complex code can easily hide errors. Static code analysis tools can be used to calculate metrics such as cyclomatic complexity and cognitive complexity. High-complexity submissions imply higher potential risk. Test coverage: Does the submitted new code have corresponding unit tests or integration tests, and what percentage of the code are covered by these tests? Test coverage tools can generate reports; submissions with high test coverage are generally more reliable. Merge request review feedback: Comments from other engineers on the code submission. If a submission receives numerous modification suggestions from multiple colleagues, requires multiple revisions before merging, or is rejected outright, the quality score will be low. Submissions with many positive reviews and quick approval are of high quality. Code submission quality can be objectively and quantitatively evaluated through a series of automated tools and peer review processes. High-quality submissions mean robust, maintainable code with few defects.

[0054] Production events refer to unexpected events that occur in the environment where a software system is officially launched and running (i.e., the production environment), resulting in service interruptions, performance degradation, functional abnormalities, or data errors. Examples include: payment functions suddenly becoming unusable, application (APP) pages failing to load, and extremely slow system responses.

[0055] For example, a production event ticket is a standardized record or work order used to track and manage the event. It typically contains the following information:

[0056] Event title and description: What happened?

[0057] Time of occurrence: When did it happen?

[0058] Priority and scope of impact: How many users were affected? What level of incident was it?

[0059] Root cause analysis: What was the root cause found after investigation?

[0060] Solution and repair process: How it was resolved.

[0061] Follow-up improvement measures: How to prevent similar incidents from happening again.

[0062] Production incident tickets are complete and structured records of online system failures; they are an important basis for problem review, accountability, and process improvement; the number and severity of production incident tickets directly reflect the software quality and operational capabilities.

[0063] The job knowledge graph is generated by constructing skill nodes based on job descriptions and an internal policy database. For example, the job knowledge graph is a structured data model that graphically represents all the knowledge, skills, and their inherent relationships required for a job; it is a topological network built around the fintech skills system. Its nodes include entities such as job roles (e.g., cloud platform development engineers), skill tags, and course resources, while edge relationships represent skill dependency paths. The graph continuously updates node weights and association strengths through a reinforcement learning mechanism, driving intelligent training decisions.

[0064] For example, a job knowledge graph includes skill nodes. In the graph, each specific skill or knowledge point is an independent "node," a basic unit constituting the graph. These skill nodes can be automatically or semi-automatically extracted, identified, and created from unstructured text data. Specifically, the first input is the job description and an internal policy database. The job description contains documents such as job responsibilities and qualifications; the internal policy database includes various business specifications, technical standards, compliance requirements, and operation manuals. Deep analysis is performed on these documents, breaking sentences down into words and labeling the attributes of each word (e.g., nouns, verbs). Entities representing "skills" or "knowledge" in the text are identified and extracted. Standardization and normalization are then implemented because skill names extracted from different documents and description methods may be inconsistent. The system needs to perform standardization to ensure the uniqueness and standardization of skill nodes in the graph. When creating nodes, rich attributes are added to them, making them not isolated labels but entities containing information.

[0065] Employee behavior data can include: employee behavior profiles: Office Automation (OA) system learning records, Application Access Management (AAM) system login frequency; by collecting and analyzing employee behavior data in the digital work environment, quantitative insights can be formed regarding their learning preferences, activity levels, areas of interest, and potential status. For example, OA learning records come from the learning and training module of the enterprise's office automation system; the record content can include a list of online courses and training projects completed by the employee; course learning duration and completion progress (e.g., whether 100% was watched); quiz scores or assessment results; which course content was actively searched or viewed (even if no learning has started). Attendance records for live training sessions and online seminars participated in.

[0066] The AAM system login frequency refers to how often an employee logs into a specific business system (such as the number of times they log in daily or weekly), the active time periods, and the duration of their online time. It reflects the employee's work activity and proficiency. High-frequency logins and long online times usually mean that the employee is an active user of the relevant system and may have more experience.

[0067] Step 102: Perform natural language processing and graph neural network processing on the multimodal data to obtain a capability gap profile and an organizational capability report.

[0068] In this embodiment, the aforementioned data are processed through natural language processing and combined with graph neural network processing to obtain a competency gap profile of the current employees to be trained and an organizational competency report. For example, the competency gap profile is a data-driven, refined, and visualized description of the difference between the required competencies for an individual or specific position and their current actual competencies. Its core is "comparison," that is, comparing "standard requirements" with "current performance." The standard requirements come from job competency models, project requirements, strategic development goals, etc.; the current performance is derived through analysis of multimodal data (such as work documents, code, communication records, and project deliverables). The competency gap profile is not merely a vague description like "insufficient technical skills," but rather a specific statement; for example, a lack of knowledge about "the application of graph neural networks in risk control scenarios"; "programming proficiency" at an intermediate level, not reaching advanced requirements; "unfamiliarity with project management tools"; and "cross-team collaboration skills" needing improvement.

[0069] An organizational capability report is a systematic analysis and summary of the organization's core capabilities status, distribution, bottlenecks, and risks, based on the integration of "capability gap profiles" of all individuals or teams, from the perspective of the organization as a whole.

[0070] Step 103: Generate training paths based on the capability gap profile and organizational capability report; and conduct training for trainees based on the training paths.

[0071] In this embodiment, based on the aforementioned capability gap profile and organizational capability report, a personalized learning path can be generated for the trainees. For example, a personalized learning path can be generated based on the different skill levels of the trainees and the courses they will learn. This may include a daily learning plan and specifying which chapter of the course to start learning.

[0072] In this embodiment, by jointly analyzing multimodal data, an organizational capability report is obtained, which includes the competency gaps of trainees and corresponding training recommendations. This improves the accuracy of competency and training needs analysis for trainees. Training paths are derived based on the competency gap profile and organizational capability report, making the generated paths more aligned with current training needs and improving their reliability. Furthermore, training is conducted on trainees based on these training paths, further enhancing the reliability of the training results.

[0073] Figure 2 An exemplary flowchart of a training decision-making method is shown. Figure 2 ;like Figure 2 As shown, based on the capability gap profile and organizational capability report, a training path is generated, including:

[0074] Step 201: Generate scenario-based courseware based on the capability gap profile, organizational capability report, technical documents, and historical courseware.

[0075] In this embodiment, based on the aforementioned capability gap profile and organizational capability report, as well as industry technical documents and historical courseware, an automatically generated scenario-based course component can be output; for example, it can be "Distributed Transaction Fault Handling - Based on Real Work Order Cases"; it supports outputting presentation slides (PPT) scripts, virtual lecturer broadcasts, and interactive sandbox experimental environments.

[0076] Step 202: Based on the capability gap profile, organizational capability report, and scenario-based courseware, generate a training path using reinforcement learning methods.

[0077] In this embodiment, a personalized training path is generated for each trainee based on an individual's skill gap profile, organizational capability report, and generated scenario-based courseware. For example, reinforcement learning methods can be used to make the generated training path more suitable for the trainee's individual skill level. For instance, the trainee is tested using the current training path. If the training effect is poor, a lower reward value is assigned to the training path; if the training effect is good, a higher reward value is assigned. Ultimately, the training path with the higher reward value is selected as the trainee's training path.

[0078] In this embodiment, scenario-based courseware is generated based on capability gap profiles, organizational capability reports, technical documents, and historical courseware. Then, training paths are generated based on these scenario-based courseware in conjunction with capability gap profiles and organizational capability reports, thereby improving the reliability of the generated training paths.

[0079] Optionally, after generating the training path, the following may also be included:

[0080] Based on the instructor's competency tags, the trainees' skill level, and the training path, a multi-objective optimization algorithm is used to generate an instructor matching scheme.

[0081] In this embodiment, the instructor's competency tag can be a quantifiable indicator of the instructor's skill level, such as "containerized deployment practical score = 4.8 / 5". The multi-objective optimization algorithm simultaneously optimizes multiple objectives (such as competency fit and teaching style preference). The generated instructor matching scheme is a mentor recommendation scheme generated based on the instructor's competency tag and the student's level; for example, it recommends instructors with high practical scores for the "fault diagnosis" skill requirement.

[0082] In this embodiment, a multi-objective optimization algorithm is used to achieve a globally optimal match between instructors' abilities and students' needs, reducing the waste of high-quality instructors and improving the overall resource utilization rate.

[0083] Optionally, the multimodal data can be processed using natural language processing and graph neural networks to obtain a competency gap profile of the trainees and an organizational competency report, including:

[0084] Natural language processing is used to extract skill tags from development performance data.

[0085] Graph neural network modeling is performed on job knowledge graph data and employee behavior data to generate relationships between capability nodes.

[0086] Based on the correlation between skill tags and capability nodes, a capability gap profile of the trainees and an organizational capability report are obtained.

[0087] In this embodiment, skill tags can be extracted using natural language processing. For example, natural language processing refers to the technology of parsing unstructured text data (such as job descriptions and interview texts) using algorithms to extract skill tags, such as identifying the skill "containerized operations and maintenance" from a description of "fault diagnosis". Graph neural network modeling is a technology that models job knowledge graph data (such as skill nodes) and employee behavior data (such as learning records) using graph structures, such as constructing a skill dependency relationship of "mastering cloud microservice solutions → learnable service circuit breaker technology". The ability node association relationship is described through the relationship between skill nodes generated by graph neural network modeling, such as the association strength between the skills "microservice architecture design" and "distributed transaction fault handling".

[0088] In this embodiment, through the collaborative modeling of natural language processing and graph neural networks, deep semantic association analysis of multi-source heterogeneous data is realized, thereby more comprehensively identifying employee skill gaps; improving the accuracy of skill gap profiling and the relevance of organizational-level demand reports, and providing a more reliable data foundation for subsequent course generation and resource matching.

[0089] Optionally, the method also includes:

[0090] Simulate real business scenarios in a virtual sandbox to generate student operation records.

[0091] A training effectiveness evaluation report is generated based on trainees' operation records and code quality scanning results.

[0092] In this embodiment, the virtual sandbox can be an interactive training platform that simulates a real production environment, such as simulating a "microservice interface timeout" failure scenario for trainees to fix. Code quality scanning can detect code defects using static analysis tools, such as identifying potential deadlock issues in the code. The training effectiveness evaluation report refers to a skills improvement analysis report generated based on trainees' operation records and code quality scanning results, such as the trainees' improvements in "distributed transaction failure handling" skills.

[0093] For example, a virtual sandbox simulates real-world business scenarios (such as database deadlock logs). Trainees are required to complete troubleshooting operations in an interactive environment, and the system records their operation steps and code modifications. Secondly, code quality scanning tools are used to analyze the defect rate of the trainees' repaired code, and this analysis, combined with the operation records, generates a training effectiveness evaluation report (such as the trainees' success rate in repairing "microservice interface timeout" scenarios).

[0094] In this embodiment, the combination of virtual sandbox and code quality scanning enables quantitative verification of training effectiveness. Through a closed-loop verification process of "learn-practice-combat-evaluation," the quantifiable nature of training outcomes is ensured, providing data support for continuous optimization of subsequent course content and instructor matching.

[0095] Optionally, the method also includes:

[0096] Based on the training effectiveness evaluation report, update the skill node weights and association strengths in the dynamic knowledge graph; the dynamic knowledge graph is used to represent employee capabilities, organizational assets, and the relationships between them.

[0097] In this embodiment, the dynamic knowledge graph is a topological network that continuously updates node weights and association strengths through a reinforcement learning mechanism. For example, based on training effectiveness evaluation reports (such as the success rate of trainees in fixing "microservice interface timeout" scenarios), the skill node weights and association strengths in the dynamic knowledge graph are updated through a reinforcement learning mechanism.

[0098] In this embodiment, the knowledge system is continuously optimized through the closed-loop update of the training effectiveness evaluation report and the dynamic knowledge graph; the reinforcement learning mechanism ensures that the knowledge system and the training effect evolve in sync, forming a complete closed loop of "demand identification-intervention-verification-iteration".

[0099] The training decision-making method provided in this embodiment obtains an organizational capability report, including a capability gap analysis of trainees and corresponding training recommendations, through joint analysis of multimodal data. This improves the accuracy of capability analysis and training needs analysis of trainees. Training paths are generated based on the capability gap profile and organizational capability report, making the generated paths more aligned with current training needs and improving their reliability. Furthermore, training is conducted on trainees based on these training paths, further enhancing the reliability of training outcomes.

[0100] Example 2

[0101] Figure 3 An exemplary schematic diagram of a training decision-making device is shown; as follows: Figure 3 As shown, the device includes:

[0102] The acquisition module 21 is used to acquire multimodal data, which includes development performance data, job knowledge graph data, and employee behavior data.

[0103] Processing module 22 is used to perform natural language processing and graph neural network processing on multimodal data to obtain a capability gap profile and an organizational capability report.

[0104] The generation module 23 is used to generate training paths based on the capability gap profile and organizational capability report; and to train the trainees based on the training paths.

[0105] The training decision-making device provided in this embodiment can execute the training decision-making method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0106] Example 3

[0107] Figure 4 The diagram above illustrates the structure of an electronic device, which includes:

[0108] The device includes a processor 291 and a memory 292; it may also include a communication interface 293 and a bus 294. The processor 291, memory 292, and communication interface 293 can communicate with each other via the bus 294. The communication interface 293 can be used for information transmission. The processor 291 can invoke logical instructions stored in the memory 292 to execute the methods described in the example above.

[0109] Furthermore, the logic instructions in the aforementioned memory 292 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0110] The memory 292, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this application. The processor 291 executes functional applications and data processing by running the software programs, instructions, and modules stored in the memory 292, that is, it implements the methods in the above method examples.

[0111] The memory 292 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory 292 may include high-speed random access memory and may also include non-volatile memory.

[0112] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method in any of the embodiments.

[0113] This application also provides a computer program product, including a computer program that, when executed by a processor, is used to implement the method in any of the embodiments.

[0114] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0115] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0116] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0117] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0118] When integrated units / modules are implemented in hardware, the hardware can be digital circuits, analog circuits, etc. The physical implementation of the hardware structure includes, but is not limited to, transistors, memristors, etc. Unless otherwise specified, the processor can be any suitable hardware processor, such as a CPU, GPU, FPGA, DSP, and ASIC, etc. Unless otherwise specified, the storage unit can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc.

[0119] If the integrated unit / module is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0120] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0121] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0122] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A training decision-making method, characterized in that, include: Acquire multimodal data; the multimodal data includes development performance data, job knowledge graph data, and employee behavior data; The multimodal data is processed using natural language processing and graph neural network processing to obtain a capability gap profile and an organizational capability report; Based on the capability gap profile and the organizational capability report, a training path is generated; and based on the training path, the trainees are trained.

2. The method according to claim 1, characterized in that, The step of generating a training path based on the capability gap profile and the organizational capability report includes: Based on the capability gap profile, the organizational capability report, technical documents, and historical courseware, generate scenario-based courseware. Based on the capability gap profile, the organizational capability report, and the scenario-based courseware, the training path is generated using reinforcement learning methods.

3. The method according to claim 2, characterized in that, After generating the training path, the process also includes: Based on the instructor's competency tags, the trainee's skill level, and the training path, a teacher matching scheme is generated using a multi-objective optimization algorithm.

4. The method according to claim 1, characterized in that, The process of performing natural language processing and graph neural network processing on the multimodal data to obtain a competency gap profile and organizational competency report for the trainees includes: Natural language processing is performed on the development performance data to extract skill tags; Graph neural network modeling is performed on the job knowledge graph data and the employee behavior data to generate the relationship between capability nodes; Based on the relationship between the skill tags and the capability nodes, a capability gap profile of the trainees and an organizational capability report are obtained.

5. The method according to any one of claims 1-4, characterized in that, The method further includes: Simulate real-world business scenarios in a virtual sandbox to generate student operation records; Based on the trainees' operation records and code quality scan results, a training effectiveness evaluation report is generated.

6. The method according to claim 5, characterized in that, The method further includes: Based on the training effectiveness evaluation report, the skill node weights and association strengths in the dynamic knowledge graph are updated; the dynamic knowledge graph is used to represent employee capabilities, organizational assets, and the relationships between them.

7. A training decision-making device, characterized in that, include: The acquisition module is used to acquire multimodal data, including development performance data, job knowledge graph data, and employee behavior data. The processing module is used to perform natural language processing and graph neural network processing on the multimodal data to obtain a capability gap profile and an organizational capability report; The generation module is used to generate training paths based on the capability gap profile and the organizational capability report; and to train the trainees based on the training paths.

8. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

Images