Human resource management method based on generative adversarial network and large language model

Abstract

A human resource management method based on a generative adversarial network and a large language model comprises the following steps: learning obtained human resource management data through the generative adversarial network to generate multiple groups of qualified virtual human resource management data; training the large language model by adopting the obtained human resource management data and the qualified virtual human resource management data to obtain a human resource management model; and analyzing human resource data to be analyzed by adopting the human resource management model to obtain an analysis result, and optimizing a human resource decision based on the analysis result. Diversified virtual human resource management data are generated by adopting the generative adversarial network, and the data are learned and comprehensively analyzed by adopting the large language model, so that diversification of generated virtual human resource data is realized, and human resource management decisions are optimized by fully utilizing the virtual human resource data and real human resource data.

Description

Technical Field

[0001] The present invention relates to the field of human resource management, and in particular to a human resource management method based on a generative adversarial network and a large language model. Background Art

[0002] At present, the application of artificial intelligence technology in the field of human resource management has made some progress, but there are still some problems that need to be solved. For example, at present, the generative adversarial network (GAN) is only used to generate virtual resume data, although it expands the resume data set of the traditional human resource management system; however, the quantity and diversity of the generated data are the key factors that determine the application effect. At present, the human resource management data generated by the generative adversarial network (GAN) is of a single type, which may affect the accuracy of human resource management decisions.

[0003] In addition, the existing large language model (LLM) technology can already assist in human resource management tasks. However, the current decision support capabilities of the large language model (LLM) are mainly focused on the processing of unstructured data. When faced with the complexity of HRM and the needs of multimodal data fusion, it is still insufficient and cannot make full use of existing data to optimize human resource management decisions. Summary of the invention

[0004] In order to solve the problem that the human resource management data generated in the prior art is insufficient and not authentic enough, and the problem that the existing data cannot be fully utilized to optimize human resource management decisions, the present invention proposes a human resource management method based on a generative adversarial network and a large language model, including:

[0005] Generate multiple sets of qualified virtual human resource management data by learning the acquired human resource management data through generative adversarial networks;

[0006] Using the obtained human resource management data and the qualified virtual human resource management data to train the large language model to obtain a human resource management model;

[0007] The human resource management model is used to analyze the human resource data to be analyzed, to obtain analysis results, and to optimize human resource decisions based on the analysis results.

[0008] Preferably, the generative adversarial network is a Wasserstein generative adversarial network.

[0009] Preferably, the large language model is a Qwen-1.5-72B large language model.

[0010] Preferably, the obtained human resource management data has been processed by the large language model in the following ways: text completion and sentiment analysis.

[0011] Preferably, the human resource management data includes: applicant resumes, interview records, background check reports, recruitment results, performance appraisal records, employee feedback records, and employee behavior data.

[0012] Preferably, the employee behavior data includes: employee behavior changes when the work environment changes, and employee behavior changes when the human resource management system changes.

[0013] Preferably, the method of learning the acquired human resource management data by generating an adversarial network to generate multiple sets of qualified virtual human resource management data includes:

[0014] Generating original virtual human resource management data through the generator of the generative adversarial network;

[0015] The discriminator of the generative adversarial network judges the authenticity of the original virtual human resource management data according to the obtained human resource management data to obtain a judgment result, and improves the generator and the discriminator according to the judgment result;

[0016] The generator and the discriminator are subjected to multiple confrontations and improvements. When the original virtual human resource management data generated by the generator meets the system requirements, it is regarded as qualified virtual human resource management data, and multiple groups of the qualified virtual human resource management data are generated by the generator.

[0017] Preferably, the step of using the obtained human resource management data and the qualified virtual human resource management data to train the large language model to obtain a human resource management model comprises:

[0018] Cleaning the acquired human resource management data and the qualified virtual human resource management data, and retaining data related to the human resource management model;

[0019] Convert the data related to the human resource management model into a data set that conforms to the model input format;

[0020] Substituting the data set that conforms to the model input format into the large language model for training to obtain a human resource management model;

[0021] The human resource management model-related data include: recruitment-related data, resignation risk-related data, performance appraisal-related data, and employee incentive-related data.

[0022] Preferably, the adopting of the human resource management model to analyze the human resource data to be analyzed to obtain analysis results, and optimizing human resource decisions based on the analysis results includes:

[0023] Using the human resource management model to analyze the recruitment-related data to obtain a recruitment analysis result;

[0024] Using the human resource management model to analyze the resignation risk related data to obtain a resignation risk analysis result;

[0025] Using the human resource management model to analyze the performance appraisal related data, and obtaining the performance appraisal method analysis results;

[0026] Using the human resource management model to analyze the employee incentive related data, and obtaining employee incentive method analysis results;

[0027] Optimize human resource decisions by comprehensively analyzing the recruitment analysis results, resignation risk analysis results, performance appraisal method analysis results, and employee incentive method analysis results.

[0028] On the other hand, based on the same concept, the present application also provides a human resource management system based on a generative adversarial network and a large language model, characterized in that it includes: a human resource management data generation module, a human resource management model training module, and a human resource decision-making module;

[0029] The human resource management data generation module is used to learn the acquired human resource management data through a generative adversarial network to generate multiple sets of qualified virtual human resource management data;

[0030] The human resource management model training module is used to train the large language model using the obtained human resource management data and the qualified virtual human resource management data to obtain a human resource management model;

[0031] The human resource decision-making module is used to analyze the human resource data to be analyzed using the human resource management model, obtain analysis results, and optimize human resource decisions based on the analysis results.

[0032] Preferably, the human resource management data generation module comprises: an original data generation submodule, a generator and discriminator improvement submodule, and a qualified data generation submodule;

[0033] The original data generation submodule is used to generate original virtual human resource management data through the generator of the generative adversarial network;

[0034] The generator and discriminator improvement submodule is used to judge the authenticity of the original virtual human resource management data according to the obtained human resource management data through the discriminator of the generative adversarial network, obtain a judgment result, and improve the generator and the discriminator according to the judgment result;

[0035] The qualified data generation submodule is used to perform multiple confrontations and improvements between the generator and the discriminator. When the original virtual human resource management data generated by the generator meets the system requirements, it is used as qualified virtual human resource management data, and multiple groups of the qualified virtual human resource management data are generated by the generator.

[0036] Preferably, the human resource management model training module includes: a data cleaning submodule, a format conversion submodule, and a substitution and training submodule;

[0037] The data cleaning submodule is used to clean the obtained human resource management data and the qualified virtual human resource management data, and retain the human resource management model related data;

[0038] The format conversion submodule is used to convert the format of the data related to the human resource management model to obtain a data set that conforms to the model input format;

[0039] The substitution and training submodule is used to substitute the data set that conforms to the model input format into the large language model for training to obtain a human resource management model.

[0040] Preferably, the human resource decision-making module includes: a recruitment analysis submodule, a resignation risk analysis submodule, a performance appraisal method analysis submodule, an employee incentive method analysis submodule, and a comprehensive analysis submodule;

[0041] The recruitment analysis submodule is used to analyze the recruitment related data using the human resource management model to obtain a recruitment analysis result;

[0042] The resignation risk analysis submodule is used to analyze the resignation risk related data using the human resource management model to obtain a resignation risk analysis result;

[0043] The performance appraisal method analysis submodule is used to analyze the performance appraisal related data using the human resource management model to obtain the performance appraisal method analysis results;

[0044] The employee incentive method analysis submodule is used to analyze the employee incentive related data using the human resource management model to obtain the employee incentive method analysis results;

[0045] The comprehensive analysis submodule is used to optimize human resource decisions by comprehensively analyzing the recruitment analysis results, resignation risk analysis results, performance appraisal method analysis results, and employee incentive method analysis results.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] The present invention provides a human resource management method based on a generative adversarial network and a large language model, comprising: learning the acquired human resource management data through a generative adversarial network to generate multiple sets of qualified virtual human resource management data; training the large language model with the acquired human resource management data and the qualified virtual human resource management data to obtain a human resource management model; analyzing the human resource data to be analyzed with the human resource management model to obtain an analysis result, and optimizing human resource decisions based on the analysis result;

[0048] By adopting a generative adversarial network to learn all kinds of human resource management data that have been obtained, multiple groups of qualified virtual human resource management data are generated; by adopting the human resource management model to analyze the human resource data to be analyzed, the analysis results are obtained, and the human resource decisions are optimized based on the analysis results, the diversification of the generation of virtual human resource data is achieved, and the virtual human resource data and real human resource data are fully utilized to optimize the human resource management decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a flow chart of the human resource management method provided by the present invention. DETAILED DESCRIPTION

[0050] In order to better understand the present invention, the content of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0051] Embodiment 1:

[0052] The present invention provides a human resource management method based on generative adversarial networks and large language models, aiming to optimize various aspects of human resource management, including recruitment process, employee turnover prediction and performance management, through the deep combination of data generation, text analysis and decision support. Figure 1 As shown, the specific steps include:

[0053] Step 1: Generate multiple sets of qualified virtual human resource management data by learning the acquired human resource management data through generative adversarial networks;

[0054] Step 2: Use the acquired human resource management data and qualified virtual human resource management data to train the large language model to obtain a human resource management model;

[0055] Step 3: Use the human resource management model to analyze the human resource data to be analyzed, obtain analysis results, and optimize human resource decisions based on the analysis results;

[0056] Wherein, step 1 specifically includes:

[0057] Step 1.1: Generate original virtual human resource management data through the generator of the generative adversarial network;

[0058] Step 1.2: The discriminator of the generative adversarial network is used to judge the authenticity of the original virtual human resource management data based on the obtained human resource management data, and the judgment result is obtained, and the generator and the discriminator are improved according to the judgment result;

[0059] Step 1.3: The generator and the discriminator are confronted and improved with each other multiple times. When the original virtual human resource management data generated by the generator meets the system requirements, it is regarded as qualified virtual human resource management data, and multiple groups of qualified virtual human resource management data are generated by the generator.

[0060] Wherein, step 2 specifically includes:

[0061] Step 2.1: Clean the acquired human resource management data and qualified virtual human resource management data, and retain the data related to the human resource management model;

[0062] Step 2.2: Convert the format of the data related to the human resource management model to obtain a data set that conforms to the model input format;

[0063] Step 2.3: Substitute the data set that conforms to the model input format into the large language model for training to obtain the human resource management model;

[0064] Among them, the data related to the human resource management model include: recruitment related data, resignation risk related data, performance appraisal related data, and employee incentive related data.

[0065] Wherein, step 3 specifically includes:

[0066] Step 3.1: Use the human resource management model to analyze recruitment-related data and obtain recruitment analysis results;

[0067] Step 3.2: Use the human resource management model to analyze the data related to turnover risk and obtain the turnover risk analysis results;

[0068] Step 3.3: Use the human resource management model to analyze the performance appraisal related data and obtain the performance appraisal method analysis results;

[0069] Step 3.4: Use the human resource management model to analyze employee incentive-related data and obtain employee incentive method analysis results;

[0070] Step 3.5: Optimize human resource decisions by comprehensively analyzing the recruitment analysis results, resignation risk analysis results, performance appraisal method analysis results, and employee incentive method analysis results.

[0071] Among them, the generative adversarial network can adopt the Wasserstein generative adversarial network, and the large language model can adopt the Qwen-1.5-72B large language model.

[0072] Among them, the obtained human resource management data has been subjected to text completion and sentiment analysis using a large language model before being used to train the human resource management model.

[0073] Among them, human resource management data include: applicant resumes, interview records, background investigation reports, recruitment results, performance appraisal records, employee feedback records, and employee behavior data; employee behavior data include: changes in employee behavior when the work environment changes, and changes in employee behavior when the human resource management system changes.

[0074] Embodiment 2:

[0075] In the embodiment of the present invention, an intelligent human resource management system based on a generative adversarial network (GAN) and a large language model (LLM) is implemented through the method of the present invention, aiming to improve the efficiency and accuracy of recruitment processes, employee turnover prediction and performance management. Based on the generation capability of the generative adversarial network (GAN) and the text analysis capability of the large language model (LLM), the system establishes a data-driven human resource management (HRM) optimization and decision-making platform, realizing the integration of data generation, analysis and decision support.

[0076] The human resources intelligent management method based on GAN and LLM in the embodiment of the present invention uses a framework such as Figure 1 As shown, it includes a data generation module, a text analysis module, a decision support module and a system interaction module, and specifically includes the following steps:

[0077] Step 1. Virtual data generation: First, use GAN to generate virtual resume data and employee behavior models. GAN consists of a generator and a discriminator, which optimizes the quality of generated data through mutual adversarial learning. The generator G takes a random noise vector z as input to generate samples G(z) similar to the real data distribution, while the discriminator D accepts real data and generated data and tries to distinguish between the two. The goal of GAN is to make the data generated by the generator as close to the real data distribution as possible, making it difficult for the discriminator to distinguish. The optimized GAN objective function is as follows:

[0078]

[0079] Among them, P data (x) is the real data distribution, D(x) represents the probability that the discriminator judges the real data X as real; p z (z) is the data distribution generated by the generator; D(G(z)) is the probability that the discriminator judges the generated data to be real.

[0080] During the training process, the generator and the discriminator are optimized alternately to continuously improve the fidelity of the generated data. To further improve the performance of GAN, the present invention adopts Wasserstein GAN (WGAN), whose objective function is:

[0081]

[0082] Among them, WGAN solves the problem of mode collapse and instability that may occur in the training process of traditional GAN ​​by introducing Wasserstein distance. The large amount of virtual resume data and employee behavior models generated by GAN can enrich the data set, especially providing diversified reference data in recruitment and employee turnover prediction.

[0083] Step 2, recruitment process optimization: In the recruitment process, the present invention uses a large language model (LLM) to conduct in-depth analysis of job seekers' resumes, cover letters and other related text data. LLM learns language features in a large text corpus through pre-training to achieve natural language understanding and generation. For recruitment tasks, LLM mainly relies on the following conditional probability formula modeling:

[0084]

[0085] Among them, y represents the target text, yt represents the tth word in the target text, x is the input text, and T represents the number of words in the target text. Through the autoregressive model, LLM can perform semantic analysis on the input resume, extract key information such as work experience, skills, educational background, etc., and match it with job requirements. In addition, the virtual resume data generated by GAN is also used to train LLM to improve its performance in recruitment tasks. Through the fusion analysis of a large amount of virtual and real data, LLM can build a more comprehensive recruitment screening model to achieve intelligent screening and recommendation of candidates. The goal of recruitment process optimization is to maximize the match between candidates and positions. The matching calculation formula can be defined as:

[0086]

[0087] Among them, w i is the weight of the i-th feature, si is the candidate's score on the i-th feature, and n is the number of features. Through this matching function, the system can sort and screen candidates to improve recruitment efficiency.

[0088] Step 3, employee turnover prediction: In employee turnover prediction, the method of the present invention analyzes employee feedback and performance data through the employee behavior model generated by GAN and LLM. First, GAN is used to generate employee behavior trajectories in different situations to simulate the reactions of employees under different environments and pressures. The discriminator is used to distinguish between employee turnover behavior and normal behavior, and the optimized generator can generate a realistic employee behavior model. LLM identifies potential resignation risks by analyzing employee performance reports, communication records, and feedback. Resignation risk prediction is modeled based on the following Bayesian formula:

[0089]

[0090] Among them, Features represents the behavior and performance characteristics of employees, and Attrition represents employee resignation. By modeling the probability distribution of employee characteristics, the system can predict the employee's tendency to resign. The virtual employee behavior data generated by GAN enhances the generalization ability of the prediction model and improves the accuracy of employee turnover prediction.

[0091] Step 4: Performance management and decision support: In terms of performance management, LLM is used to analyze unstructured text data such as employee performance reports, team feedback, and project summaries. Through sentiment analysis and semantic understanding, LLM can identify employees' work attitudes, teamwork capabilities, and project contributions. The core of sentiment analysis is to calculate the sentiment score of the text. The formula for calculating the sentiment score of the text is:

[0092]

[0093] Among them, wj is the weight of the i-th sentiment word, s i is the score of sentiment words. Through the sentiment score, the system can identify the emotional state of employees and take timely measures. The virtual employee behavior data generated by GAN is used to simulate different performance scenarios and provide multi-dimensional decision support for management. Finally, the system evaluates the overall performance of employees and forms a decision support report to assist the HR department in performance management and reward decisions.

[0094] Embodiment 3: The present invention based on the same inventive concept also provides a human resource management system based on a generative adversarial network and a large language model, comprising:

[0095] A human resource management data generation module is used to learn the acquired human resource management data through a generative adversarial network to generate multiple sets of qualified virtual human resource management data;

[0096] A human resource management model training module is used to train a large language model using the acquired human resource management data and qualified virtual human resource management data to obtain a human resource management model;

[0097] The human resource decision-making module is used to analyze the human resource data to be analyzed using a human resource management model, obtain analysis results, and optimize human resource decisions based on the analysis results.

[0098] The human resource management data generation module includes:

[0099] A raw data generation submodule, used to generate raw virtual human resource management data through a generator of a generative adversarial network;

[0100] The generator and discriminator improvement submodule is used to judge the authenticity of the original virtual human resource management data according to the obtained human resource management data through the discriminator of the generative adversarial network, obtain the judgment result, and improve the generator and the discriminator according to the judgment result;

[0101] The qualified data generation submodule is used to conduct multiple confrontations and improvements between the generator and the discriminator. When the original virtual human resource management data generated by the generator meets the system requirements, it is used as qualified virtual human resource management data, and multiple groups of qualified virtual human resource management data are generated through the generator.

[0102] Human resource management model training module, including:

[0103] The data cleaning submodule is used to clean the acquired human resource management data and qualified virtual human resource management data, and retain the data related to the human resource management model;

[0104] The format conversion submodule is used to convert the format of data related to the human resource management model to obtain a data set that conforms to the model input format;

[0105] The substitution and training submodule is used to substitute the data set that conforms to the model input format into the large language model for training to obtain the human resource management model.

[0106] Human resource decision-making module, including:

[0107] The recruitment analysis submodule is used to analyze recruitment-related data using a human resource management model to obtain recruitment analysis results;

[0108] The resignation risk analysis submodule is used to analyze resignation risk related data using a human resource management model to obtain resignation risk analysis results;

[0109] The performance appraisal method analysis submodule is used to analyze the performance appraisal related data using the human resource management model to obtain the performance appraisal method analysis results;

[0110] The employee incentive method analysis submodule is used to analyze employee incentive related data using the human resource management model to obtain employee incentive method analysis results;

[0111] The comprehensive analysis submodule is used to optimize human resource decisions by comprehensively analyzing recruitment analysis results, resignation risk analysis results, performance appraisal method analysis results, and employee incentive method analysis results.

[0112] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0113] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0116] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A human resource management method based on generative adversarial networks and large language models, characterized in that: include: Generate multiple sets of qualified virtual human resource management data by learning the acquired human resource management data through generative adversarial networks; Using the obtained human resource management data and the qualified virtual human resource management data to train the large language model to obtain a human resource management model; The human resource management model is used to analyze the human resource data to be analyzed, to obtain analysis results, and to optimize human resource decisions based on the analysis results.

2. The method according to claim 1, characterized in that The method of learning the acquired human resource management data by generating an adversarial network to generate multiple sets of qualified virtual human resource management data includes: Generating original virtual human resource management data through the generator of the generative adversarial network; The discriminator of the generative adversarial network judges the authenticity of the original virtual human resource management data according to the obtained human resource management data to obtain a judgment result, and improves the generator and the discriminator according to the judgment result; The generator and the discriminator are subjected to multiple confrontations and improvements. When the original virtual human resource management data generated by the generator meets the system requirements, it is regarded as qualified virtual human resource management data, and multiple groups of the qualified virtual human resource management data are generated by the generator.

3. The method according to claim 1, characterized in that The step of training the large language model using the acquired human resource management data and the qualified virtual human resource management data to obtain a human resource management model includes: Cleaning the acquired human resource management data and the qualified virtual human resource management data, and retaining data related to the human resource management model; Convert the data related to the human resource management model into a data set that conforms to the model input format; Substituting the data set that conforms to the model input format into the large language model for training to obtain a human resource management model; The human resource management model-related data include: recruitment-related data, resignation risk-related data, performance appraisal-related data, and employee incentive-related data.

4. The method according to claim 3, characterized in that The method of using the human resource management model to analyze the human resource data to be analyzed, obtaining analysis results, and optimizing human resource decisions based on the analysis results includes: Using the human resource management model to analyze the recruitment-related data to obtain a recruitment analysis result; Using the human resource management model to analyze the resignation risk related data to obtain a resignation risk analysis result; Using the human resource management model to analyze the performance appraisal related data, and obtaining the performance appraisal method analysis results; Using the human resource management model to analyze the employee incentive related data, and obtaining employee incentive method analysis results; Optimize human resource decisions by comprehensively analyzing the recruitment analysis results, resignation risk analysis results, performance appraisal method analysis results, and employee incentive method analysis results.

5. The method according to claim 1, characterized in that The human resources management data include: applicant resumes, interview records, background check reports, recruitment results, performance appraisal records, employee feedback records, and employee behavior data.

6. The method according to claim 5, characterized in that The employee behavior data includes: employee behavior changes when the work environment changes, and employee behavior changes when the human resource management system changes.

7. A human resource management system based on generative adversarial networks and large language models, characterized in that: include: Human resource management data generation module, human resource management model training module, and human resource decision-making module; The human resource management data generation module is used to learn the acquired human resource management data through a generative adversarial network to generate multiple sets of qualified virtual human resource management data; The human resource management model training module is used to train the large language model using the obtained human resource management data and the qualified virtual human resource management data to obtain a human resource management model; The human resource decision-making module is used to analyze the human resource data to be analyzed using the human resource management model, obtain analysis results, and optimize human resource decisions based on the analysis results.

8. The system according to claim 7, characterized in that The human resource management data generation module includes: an original data generation submodule, a generator and discriminator improvement submodule, and a qualified data generation submodule; The original data generation submodule is used to generate original virtual human resource management data through the generator of the generative adversarial network; The generator and discriminator improvement submodule is used to judge the authenticity of the original virtual human resource management data according to the obtained human resource management data through the discriminator of the generative adversarial network, obtain a judgment result, and improve the generator and the discriminator according to the judgment result; The qualified data generation submodule is used to perform multiple confrontations and improvements between the generator and the discriminator. When the original virtual human resource management data generated by the generator meets the system requirements, it is used as qualified virtual human resource management data, and multiple groups of the qualified virtual human resource management data are generated by the generator.

9. The system according to claim 8, characterized in that The human resource management model training module includes: a data cleaning submodule, a format conversion submodule, and a substitution and training submodule; The data cleaning submodule is used to clean the obtained human resource management data and the qualified virtual human resource management data, and retain the human resource management model related data; The format conversion submodule is used to convert the format of the data related to the human resource management model to obtain a data set that conforms to the model input format; The substitution and training submodule is used to substitute the data set that conforms to the model input format into the large language model for training to obtain a human resource management model.

10. The system according to claim 9, characterized in that The human resource decision-making module includes: a recruitment analysis submodule, a resignation risk analysis submodule, a performance appraisal method analysis submodule, an employee incentive method analysis submodule, and a comprehensive analysis submodule; The recruitment analysis submodule is used to analyze the recruitment related data using the human resource management model to obtain a recruitment analysis result; The resignation risk analysis submodule is used to analyze the resignation risk related data using the human resource management model to obtain a resignation risk analysis result; The performance appraisal method analysis submodule is used to analyze the performance appraisal related data using the human resource management model to obtain the performance appraisal method analysis results; The employee incentive method analysis submodule is used to analyze the employee incentive related data using the human resource management model to obtain the employee incentive method analysis results; The comprehensive analysis submodule is used to optimize human resource decisions by comprehensively analyzing the recruitment analysis results, resignation risk analysis results, performance appraisal method analysis results, and employee incentive method analysis results.

11. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, the human resource management method based on a generative adversarial network and a large language model as described in any one of claim 6 is implemented.

12. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, the human resource management method based on a generative adversarial network and a large language model as described in any one of claim 6 is implemented.

Images