Job description optimization methods, equipment, storage media, and computer program products

Abstract

This application discloses a method, device, storage medium, and computer program product for optimizing job descriptions, relating to the field of enterprise resource management technology. The method includes: acquiring a user-input job description optimization instruction and a job description to be optimized; performing natural language semantic analysis on the job description optimization instruction and entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list; retrieving benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights, wherein the construction process of the dynamic knowledge base employs heterogeneous data alignment; and optimizing the job description to be optimized item by item according to the benchmark job data to obtain an optimized job description and optimization instructions. By comprehensively utilizing semantic analysis, entity extraction, dynamic knowledge base retrieval, and item-by-item optimization techniques, the accuracy and timeliness of job description optimization are improved.

Description

Technical Field

[0001] This application relates to the field of enterprise resource management technology, and in particular to methods, equipment, storage media and computer program products for optimizing job descriptions. Background Technology

[0002] In the field of human resource management, optimizing job descriptions is of great significance for enterprises' talent selection, training, and management. With the continuous adjustment of corporate strategies and rapid changes in industry trends, higher demands are placed on the accuracy and timeliness of job descriptions. Currently, job description optimization mainly relies on manual operation, and its process includes manual data collection, ambiguity resolution of responsibilities, static update mechanisms, and reliance on subjective experience. Specifically, manual data collection is usually based on limited internal analysis or reference to industry-standard templates, lacking comprehensiveness and specificity; in terms of responsibility description, most job descriptions only focus on "the tasks actually performed" rather than "the tasks that should be performed," resulting in unclear and inaccurate responsibility definitions, easily leading to overlaps or omissions; the optimized documents lack dynamic adjustment capabilities and cannot adapt to changes in corporate strategy or industry trends in real time; furthermore, when adjusting content, human resource departments often rely on personal experience, lacking data support, which can easily lead to biases.

[0003] Therefore, how to accurately and efficiently optimize job descriptions has become a technical problem that this application urgently needs to solve.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a method, device, storage medium, and computer program product for optimizing job descriptions, aiming to solve the technical problem of accurately and efficiently optimizing job descriptions.

[0006] To achieve the above objectives, this application proposes a method for optimizing job descriptions, the method comprising:

[0007] Obtain the user-inputted job description optimization instructions and the job description to be optimized;

[0008] The job description optimization instructions are parsed using natural language semantics, and entity extraction and semantic analysis are performed on the job descriptions to be optimized to obtain a candidate job list.

[0009] Based on the candidate job list and matching weights, benchmark job data is retrieved from a pre-built dynamic knowledge base, wherein the construction process of the dynamic knowledge base adopts heterogeneous data alignment.

[0010] Based on the benchmark job data, the job description to be optimized is optimized item by item to obtain the optimized job description and optimization instructions.

[0011] In one embodiment, the steps of performing natural language semantic parsing on the job description optimization instructions and performing entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list include:

[0012] The job title is extracted from the job description entity to be optimized, and the optimization instructions for the job description are parsed using natural language to extract the user's optimization intent.

[0013] The pre-built large model is invoked to perform semantic analysis on the user's optimization intention and the job description to be optimized, and explicit and implicit conditions are extracted respectively.

[0014] A candidate job list is generated based on the explicit conditions, the implicit conditions, and the job title.

[0015] In one embodiment, the step of retrieving benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights further includes:

[0016] Collect diverse original job description data, and perform anonymization and data cleaning on the original job description data to obtain job description data;

[0017] A pre-trained model is selected to perform entity recognition on the job description data and key information is labeled to obtain a labeled job description dataset.

[0018] Based on predefined encoding rules and label dictionaries, the labeled job description dataset is mapped to a unified encoding system at multiple levels to form a multi-level encoded mapping dataset;

[0019] A dynamic knowledge base is constructed based on the multi-level encoding mapping dataset.

[0020] In one embodiment, the step of mapping the labeled job description dataset to a unified encoding system at multiple levels based on predefined encoding rules and a label dictionary to form a multi-level encoded mapping dataset further includes:

[0021] Define encoding rules, and based on the encoding rules, perform data source integration and tag standardization processing on the job description data to obtain a tag dictionary.

[0022] In one embodiment, the step of mapping the labeled job description dataset to a unified encoding system at multiple levels based on predefined encoding rules and a label dictionary to form a multi-level encoded mapping dataset includes:

[0023] A hierarchical tag system is constructed based on the predefined encoding rules and tag dictionary;

[0024] Based on the hierarchical labeling system, the labeled job description dataset is mapped to a unified coding system through a precise matching module and a rule generalization module, forming a multi-level coding mapping dataset.

[0025] In one embodiment, the step of retrieving benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights includes:

[0026] A reinforcement learning model is used to adjust the matching weights, and benchmark job data that match the candidate job list are retrieved from a pre-built dynamic knowledge base based on the adjusted matching weights.

[0027] The feedback data of the benchmark job data is recorded, and the online incremental training of the reinforcement learning model is triggered based on the feedback data to update the matching weights.

[0028] In one embodiment, the job description to be optimized and the benchmark job data are compared and analyzed item by item to generate optimization suggestions;

[0029] The job description to be optimized is optimized item by item according to the optimization suggestions to obtain the optimized job description, and an optimization description is generated based on the optimized job description.

[0030] In addition, to achieve the above objectives, this application also proposes a job description optimization device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the job description optimization method as described above.

[0031] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the job description optimization method described above.

[0032] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the job description optimization method described above.

[0033] One or more technical solutions proposed in this application have at least the following technical effects:

[0034] The process involves: acquiring user-inputted job description optimization instructions and the job description to be optimized; performing natural language semantic analysis on the job description optimization instructions and entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list; retrieving benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights, wherein the construction process of the dynamic knowledge base employs heterogeneous data alignment; and optimizing the job description to be optimized item by item according to the benchmark job data to obtain the optimized job description and optimization instructions. First, the user's intent is accurately understood through natural language semantic analysis, and the deeper semantics of the job are mined by entity extraction and semantic analysis. Then, the benchmark job data retrieved based on matching weights is optimized to generate precise and personalized optimization solutions tailored to the specific needs and job characteristics of different enterprises. This avoids the subjective bias of manual optimization and the limitations of a single template, making the job description more aligned with the actual business needs of the enterprise. Furthermore, the pre-built dynamic knowledge base employs heterogeneous data alignment technology to ensure the breadth and accuracy of the retrieved benchmark job data. Secondly, the job descriptions are optimized item by item based on benchmark job data, resulting in optimized job descriptions and optimization explanations. This item-by-item optimization process achieves precise optimization of the job descriptions. By comprehensively utilizing semantic parsing, entity extraction, dynamic knowledge base retrieval, and item-by-item optimization techniques, the accuracy and timeliness of job description optimization are improved, meeting the technical needs of enterprises in job description optimization. Attached Figure Description

[0035] 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.

[0036] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a flowchart illustrating the first embodiment of the job description optimization method in this application.

[0038] Figure 2 This is a flowchart illustrating the second embodiment of the job description optimization method in this application.

[0039] Figure 3 This is a flowchart illustrating the third embodiment of the job description optimization method in this application.

[0040] Figure 4 This is a flowchart illustrating the fourth embodiment of the job description optimization method in this application.

[0041] Figure 5 This is a schematic diagram of the module structure of the job description optimization device according to an embodiment of this application;

[0042] Figure 6 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the job description optimization method in this application embodiment.

[0043] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0044] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0045] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0046] The main solution of this application embodiment is as follows: Obtain the user-inputted job description optimization instruction and the job description to be optimized; perform natural language semantic parsing on the job description optimization instruction, and perform entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list; retrieve benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights, wherein the construction process of the dynamic knowledge base adopts heterogeneous data alignment; optimize the job description to be optimized item by item according to the benchmark job data to obtain the optimized job description and optimization instructions.

[0047] This application's embodiments take into account that: In the field of human resource management, the optimization of job descriptions is of great significance for enterprises' talent selection, training, and management. Furthermore, with the continuous adjustment of corporate strategies and rapid changes in industry trends, higher demands are placed on the accuracy and timeliness of job descriptions. Currently, job description optimization mainly relies on manual operation, and its process includes manual data collection, ambiguity processing of responsibilities, static update mechanisms, and reliance on subjective experience. Specifically, manual data collection is usually based on limited internal analysis or reference to industry-standard templates, lacking comprehensiveness and specificity; in terms of responsibility description, most job descriptions only focus on "the tasks actually performed" rather than "the tasks that should be performed," resulting in unclear and inaccurate responsibility definitions, easily leading to overlaps or omissions; the optimized documents lack dynamic adjustment capabilities and cannot adapt to changes in corporate strategy or industry trends in real time; in addition, when adjusting content, human resource departments often rely on personal experience, lacking data support, which can easily lead to biases.

[0048] Therefore, this application provides a solution to obtain user-inputted job description optimization instructions and job descriptions to be optimized; perform natural language semantic parsing on the job description optimization instructions, and perform entity extraction and semantic analysis on the job descriptions to be optimized to obtain a candidate job list; retrieve benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights, wherein the construction process of the dynamic knowledge base adopts heterogeneous data alignment; optimize the job descriptions to be optimized item by item according to the benchmark job data to obtain optimized job descriptions and optimization instructions. First, the user's intent is accurately understood through natural language semantic parsing, and the deep semantics of the job are mined by entity extraction and semantic analysis. Then, the benchmark job data retrieved according to the matching weights is optimized to generate accurate and personalized optimization schemes for the specific needs and job characteristics of different enterprises, avoiding the subjective bias of manual optimization and the limitations of a single template, making the job descriptions more in line with the actual business needs of enterprises. Furthermore, the pre-built dynamic knowledge base adopts heterogeneous data alignment technology to ensure that the retrieved benchmark job data is extensive and accurate. Secondly, the job descriptions are optimized item by item based on benchmark job data, resulting in optimized job descriptions and optimization explanations. This item-by-item optimization process achieves precise optimization of the job descriptions. By comprehensively utilizing semantic parsing, entity extraction, dynamic knowledge base retrieval, and item-by-item optimization techniques, the accuracy and timeliness of job description optimization are improved, meeting the technical needs of enterprises in job description optimization.

[0049] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as a job description optimization system. The following description uses a job description optimization system, hereinafter referred to as the system, as an example to illustrate this embodiment and the subsequent embodiments.

[0050] Based on this, this application provides a method for optimizing job descriptions, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the job description optimization method in this application.

[0051] In this embodiment, the job description optimization method includes steps S10 to S40:

[0052] Step S10: Obtain the user-inputted job description optimization instruction and the job description to be optimized;

[0053] A job description optimization instruction refers to a user's instruction to the system to initiate the job description optimization process; a job description to be optimized refers to the original job description text or file that the user wants to optimize, which usually describes detailed information such as the responsibilities, requirements, and working conditions of a job in text form, and its format includes PDF, Word, or plain text.

[0054] In one possible implementation, user input can be received through a web-based user interface or a local application. The user interface provides file upload functionality, supporting various file formats such as PDF and Word, and provides text boxes for users to input optimization instructions, such as "Please optimize this job description" or "Please optimize this JD (JobDescription)". Upon receiving this input, the system performs preliminary format checks and content extraction on the files to ensure correct parsing of the text information.

[0055] Step S20: Perform natural language semantic parsing on the job description optimization instructions, and perform entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list;

[0056] Natural language semantic parsing of job description optimization instructions refers to using natural language processing technology to understand the semantic information in user-input instructions and clarify the user's specific optimization needs. For example, if the user inputs "Please optimize this job description," the system needs to identify that the user wants to optimize the job description. Natural language semantic parsing enables the system to accurately grasp the user's intent and avoid deviations in the optimization direction due to ambiguity or vagueness of the instructions.

[0057] Entity extraction and semantic analysis for optimizing job descriptions refers to extracting meaningful entity information from the text content of job descriptions, such as key elements like job title, responsibilities, and qualifications, and analyzing the semantic relationships between these elements.

[0058] The candidate job list refers to a list of job names that may be related to or similar to the user's optimization needs, generated during the job description optimization process based on the user's input optimization instructions and semantic parsing and entity extraction operations on the original job description.

[0059] In one possible implementation, after receiving user instructions and job descriptions, the system first uses a deep learning-based natural language processing model, such as a language model based on the Transformer architecture, to semantically parse the instructions. This model analyzes keywords and context in the instructions to determine the user's optimization intent, such as whether it's a comprehensive optimization or optimization targeting a specific aspect. For the job description to be optimized, the system uses Named Entity Recognition (NER) technology combined with semantic analysis algorithms to identify and extract various entities from the text, such as job title, job responsibilities, and job requirements. It then analyzes the relationships between these entities to construct a semantic graph of the job. Based on this graph and the user instructions, the system generates a candidate job list, which may include information on other jobs similar to or related to the target job, providing diverse references for subsequent optimization.

[0060] It should be noted that the system may employ various techniques to improve accuracy and comprehensiveness during entity extraction and semantic analysis. For example, in the entity extraction stage, in addition to using the NER model, rule-based matching methods can be combined to formulate specific recognition rules for some common job entities, thus compensating for the model's shortcomings in certain specific situations. In the semantic analysis stage, word vector technology is used to convert text information into vector form, thereby making it easier to calculate the semantic similarity between different entities, uncover implicit semantic relationships, and support the generation of a more reasonable candidate job list.

[0061] Step S30: Based on the candidate job list and matching weights, retrieve benchmark job data from a pre-built dynamic knowledge base. The construction process of the dynamic knowledge base adopts heterogeneous data alignment.

[0062] The dynamic knowledge base is a repository that integrates recruitment data from globally renowned companies, job descriptions (JDs) from recruitment platforms, and corporate strategy documents. It reflects in real-time changes in market job demands, industry trends, and internal job adjustments within companies, providing the latest and most accurate data support for optimizing job descriptions. The construction process of the dynamic knowledge base employs heterogeneous data alignment. Heterogeneous data alignment refers to mapping job-related data from different data sources, with different formats, structures, and semantics, into a unified coding system and data framework through a series of technical means and rules to achieve data consistency and comparability.

[0063] In one possible implementation, the system's retrieval module compares each job in the candidate job list with a massive amount of job data in a dynamic knowledge base. During the comparison process, a similarity score is calculated between each candidate job and the jobs in the knowledge base based on a preset matching weight.

[0064] Step S40: Optimize the job description of the job to be optimized item by item based on the benchmark job data to obtain the optimized job description and optimization instructions.

[0065] The job description to be optimized is then improved item by item based on benchmark job data. This mainly refers to using the retrieved benchmark job data as a reference standard to conduct a detailed review and improvement of each component of the job description (such as basic job information, job purpose, main responsibilities, qualifications, etc.) to make it more in line with industry standards, market trends, and the company's internal strategic needs. The optimization description is a detailed explanation and record of the optimization process and results, including which parts were optimized, the reasons for the optimization, and which benchmark job data were used as references. This aims to provide users with more background information and basis, enabling them to understand the rationality and necessity of the optimization, and to trace and adjust the optimization results when necessary.

[0066] In one possible implementation, when optimizing each item, the system compares and analyzes each section of the job description with benchmark job data. For example, when optimizing the "Main Responsibilities" section, the system references detailed descriptions of similar job responsibilities in the benchmark data, and expands, refines, or adjusts the responsibilities based on the company's own characteristics and needs to better align with actual work scenarios and corporate strategic goals. For the "Qualifications" section, the system may update and optimize the corresponding content in the original job description based on the requirements for educational background, work experience, and professional skills in the benchmark data, ensuring it accurately reflects the reasonable expectations the position places on candidates.

[0067] It should be noted that during the optimization process, the system comprehensively utilizes various technologies and methods, such as text mining, data analysis, and machine learning, to ensure the accuracy and effectiveness of the optimization results. At the same time, it also needs to consider the specific circumstances and individual needs of each enterprise, making appropriate customized adjustments based on benchmark data to avoid the optimized job descriptions becoming too generic and lacking specificity.

[0068] This embodiment provides a method for optimizing job descriptions. It involves obtaining a user-inputted job description optimization instruction and a job description to be optimized; performing natural language semantic analysis on the optimization instruction and entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list; retrieving benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights, wherein the construction process of the dynamic knowledge base employs heterogeneous data alignment; and optimizing the job description item by item according to the benchmark job data to obtain an optimized job description and optimization instructions. First, it accurately understands the user's intent through natural language semantic analysis, combines entity extraction and semantic analysis to mine the deep semantics of the job, and then optimizes the benchmark job data retrieved based on matching weights. This generates precise and personalized optimization solutions tailored to the specific needs and job characteristics of different enterprises, avoiding the subjective bias of manual optimization and the limitations of a single template, making the job description more aligned with the actual business needs of the enterprise. Furthermore, the pre-built dynamic knowledge base employs heterogeneous data alignment technology to ensure the breadth and accuracy of the retrieved benchmark job data. Secondly, the job descriptions were optimized item by item based on benchmark job data, resulting in optimized job descriptions and optimization explanations. This item-by-item optimization process achieved precise optimization of the job descriptions. By comprehensively utilizing semantic parsing, entity extraction, dynamic knowledge base retrieval, and item-by-item optimization techniques, the accuracy and timeliness of job description optimization were improved, meeting the technical needs of enterprises in job description optimization.

[0069] In one feasible implementation, step S30 may include steps S31 to S32:

[0070] Step S31: Adjust the matching weights using a reinforcement learning model, and retrieve benchmark job data that matches the candidate job list from the pre-built dynamic knowledge base based on the adjusted matching weights.

[0071] Adjusting matching weights using a reinforcement learning model refers to using reinforcement learning algorithms to dynamically adjust matching weights based on historical feedback data accumulated by the system during the optimization of job descriptions, so that the system can more accurately retrieve benchmark job data that matches the candidate job list.

[0072] The reinforcement learning model is trained using states, actions, and rewards. States include information such as the job description content to be optimized, a list of candidate jobs, and optimization instructions input by the user; actions refer to adjusting the matching weights; and rewards are based on user feedback (such as HR adopting the optimization result or marking the optimization as unreasonable) to measure the effectiveness of the adjustments. In this way, the system gradually finds the optimal matching weights through continuous trial and error, making the retrieved benchmark job data more aligned with the user's actual needs.

[0073] Specifically, the system takes the current job description content to be optimized, the candidate job list after parsing user instructions, and standard coded data in the dynamic knowledge base as state inputs. The reinforcement learning model generates weight adjustment actions for each coding dimension based on preset initial weights (such as term importance based on TF-IDF) and online feedback history. For example, if the skill "Python data analysis" is frequently adopted in a job description, the model will increase the association weight between "technical skills - programming language - Python" and "technical skills - application scenarios - data analysis". The adjusted weights are applied to the retrieval algorithm of the dynamic knowledge base, prioritizing the matching of high-frequency demand for similar positions in globally renowned companies or emerging industry trend data, such as filtering benchmark job descriptions from the knowledge base that contain both "Python" and "data analysis" and match the job level and years of work experience. The retrieval process uses multi-dimensional weighted similarity calculation to ensure that the returned results not only meet explicit conditions (such as job name and job level) but also cover implicit associations (such as the combined requirement of cross-team collaboration ability and project management experience).

[0074] Step S32: Record the feedback data of the benchmark job data, and trigger online incremental training of the reinforcement learning model based on the feedback data to update the matching weights.

[0075] Recording feedback data from benchmark job data primarily refers to collecting user (e.g., HR personnel) evaluations and feedback on the optimization results after the system completes the job description optimization. This feedback data may include satisfaction with the optimized job descriptions, the adoption or rejection of specific optimization items, and evaluations of the relevance and usefulness of the retrieved benchmark job data. Feedback data is a crucial basis for measuring the effectiveness of system optimization.

[0076] Online incremental training of the reinforcement learning model, triggered by feedback data to update matching weights, refers to the system using accumulated feedback data to update and train the reinforcement learning model in real-time or periodically. Online incremental training enables the model to adapt promptly to changes in user needs and dynamic adjustments in the market environment, continuously improving the accuracy and optimization effect of matching weights.

[0077] Specifically, the system records the adoption rate of recommended benchmark job data by the human resources department (such as the percentage of clicks on "Adopt Optimization Suggestions") and negative feedback from manual annotations (such as marking "Skill Requirements Mismatch" or "Experience Years Deviation Too Large"), transforming this into reward signals for reinforcement learning. For example, if HR adopts a recommended encoding mapping of "5+ years of Java development experience," the corresponding weight receives a positive reward; if HR reports "PyTorch framework requirement not recognized," a negative penalty is triggered on the "Technical Skills-Framework-PyTorch" encoding. This feedback data, along with the current state and actions, is stored in an experience pool, triggering incremental training every 100 new data points. During training, the reinforcement learning model balances exploration and utilization through an epsilon-greedy strategy (randomly adjusting weights with a 10% probability to discover potential correlations), while simultaneously updating the weight allocation strategy using a deep neural network. For example, when the system detects a continuous increase in the implicit demand for "cross-departmental communication skills" in financial industry positions, the reinforcement learning model automatically prioritizes "soft skills - communication skills - cross-departmental coordination" in the search, ensuring that subsequent recommendations are more aligned with actual industry needs. Additionally, the system updates the dynamic knowledge base synchronously through scheduled tasks, such as weekly scraping and recoding the latest job descriptions from recruitment platforms, forming a closed-loop optimization mechanism.

[0078] In one feasible implementation, step S40 may include steps S41 to S42:

[0079] Step S41: Compare and analyze the job description to be optimized and the benchmark job data item by item to generate optimization suggestions;

[0080] The system parses the document to be optimized into structured fields (such as job responsibilities, qualifications, and skills) and converts them into a standard coding format based on a unified coding system. For example, "responsible for Python data analysis" is mapped to "technical skills - programming language - Python / technical skills - application scenarios - data analysis". Subsequently, the system retrieves matching benchmark job data from a dynamic knowledge base and compares explicit indicators (such as skill requirements and years of work experience) and implicit relationships (such as the coupling degree between "project management experience" and "cross-team collaboration ability") item by item through coding similarity calculation. During the comparison process, the large model combines semantic association analysis to identify differences. For example, if benchmark data generally requires "familiarity with the TensorFlow framework" but the document to be optimized does not mention it, an optimization suggestion of "add skill requirement: TensorFlow" is generated. If the "job description" in the document to be optimized contains vague expressions (such as "assist in completing projects"), it is recommended to break it down into specific tasks. Additionally, forward-looking optimization suggestions can be proposed by referring to industry trend data.

[0081] Step S42: Optimize the job description to be optimized item by item according to the optimization suggestions to obtain the optimized job description, and generate an optimization description based on the optimized job description.

[0082] Specifically, the system employs natural language generation technology (such as GPT-4's text reconstruction capabilities) to embed optimization suggestions into the original document. For example, it inserts "Proficient in the TensorFlow framework and has model deployment experience" into the "Job Qualifications" section, while removing redundant entries (such as the repeated requirement of "basic Excel operations"). The optimized document retains the company's original format to ensure compatibility with internal templates. The optimization explanation document visually displays the changes through a comparison view. For example, it lists the differences between the original description "Assisted in completing projects" and the new description "Independently responsible for data cleaning and feature engineering" in a table format, and includes benchmark company cases (such as job descriptions for similar positions in well-known companies) and large-scale model analysis conclusions (such as "Specific job responsibilities can improve candidate matching"). The final output includes a directly usable and editable Word document and a PDF version for archiving and publication, while also providing version history tracking functionality to facilitate the HR team's review of the optimization logic.

[0083] Based on the first embodiment of this application, a second embodiment of this application is proposed. In the second embodiment of this application, content that is the same as or similar to that in the first embodiment described above can be referred to the above description and will not be repeated hereafter.

[0084] Based on this, please refer to Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the present application. In this embodiment, step S10 includes steps S11 to S13:

[0085] Step S11: Extract the job title from the job description entity to be optimized, and perform natural language parsing on the job description optimization instructions to extract the user's optimization intent;

[0086] Extracting the job title from the job description to be optimized refers to using Named Entity Recognition (NER) in Natural Language Processing to accurately identify and extract the key entity information—the job title—from the text content of the job description. The job title is one of the core elements of the job description, directly reflecting the job's position and basic functions within the company's organizational structure.

[0087] First, the system uses multimodal input parsing and natural language processing (NLP) technologies to extract job titles and user intent. Specifically, the system receives job descriptions uploaded by users (supporting PDF, Word, or plain text). For unstructured documents, OCR technology is used to extract the text content, and an entity recognition model (such as a NER model fine-tuned based on Qwen2.5-3B-Instruct) is used to locate the job title field in the document. For example, "Senior Data Analyst" is extracted as the standard job title from "Senior Data Analyst Job Description".

[0088] Furthermore, the system parses the optimization instructions input by the user, such as "Please optimize the job description for this position and add AI skill requirements." Through the intent recognition module of the large model, it breaks down the core requests in the instruction—"optimize the job description" and "supplement AI skills"—and, combined with the context, eliminates ambiguity, distinguishing whether "optimization" refers to streamlining content or supplementing missing items, ultimately generating a structured optimization instruction.

[0089] Step S12: Call the pre-built large model to perform semantic analysis on the user's optimization intention and the job description to be optimized, and extract explicit conditions and implicit conditions respectively;

[0090] The process of using a pre-built large model to perform semantic analysis on the user's optimization intent and the job description to be optimized mainly refers to leveraging the semantic analysis capabilities of large language models (such as GPT-4 and BERT) to deeply understand and parse the semantic information in the job description and the user's optimization intent. Specifically, the system inputs the job description text to be optimized and the user's optimization intent into the pre-trained large model GPT-4. The model is guided by chained prompts to analyze in stages: first, it parses explicit conditions (such as "5+ years of Python experience" explicitly listed in the job description); second, it uncovers implicit conditions (such as inferring "cross-team communication skills" from "coordinating cross-departmental needs"). The large model GPT-4 also correlates with benchmark data in a dynamic knowledge base to identify the gap between explicit conditions and industry standards (such as the original document only requiring "basic SQL," while the benchmark data requires "complex query optimization"). Finally, it outputs a structured list of explicit conditions and a matrix of implicit conditions.

[0091] Step S13: Generate a candidate job list based on the explicit conditions, the implicit conditions, and the job name.

[0092] The system first precisely matches the core keywords of the job titles. Then, it filters jobs in the knowledge base that meet the minimum requirements based on explicit criteria. Finally, it calculates similarity by combining implicit criteria with weighted similarity, and outputs candidate jobs sorted by matching degree. The candidate list includes explanations of the matching criteria (e.g., "Due to the detection of user intent to supplement AI skills, jobs with MLOps requirements are given priority"), allowing HR personnel to easily view benchmark case comparisons.

[0093] In this embodiment, by accurately extracting job titles from the job descriptions to be optimized, the specific targets of the optimization work are quickly identified, avoiding subsequent optimization deviations caused by incorrect job title identification. Natural language parsing of user optimization instructions enables the system to accurately grasp user intent. Using a large model for semantic analysis, the system deeply mines the explicit and implicit conditions in the job descriptions and user instructions, covering clearly listed requirements as well as potential ability and competency requirements captured through semantic reasoning. This helps explore optimization directions from different angles, improving the innovation and adaptability of the optimization results.

[0094] Based on the first and / or second embodiments of this application, a third embodiment of this application is proposed. In this third embodiment, content that is the same as or similar to the above embodiments can be referred to the above description, and will not be repeated hereafter.

[0095] Based on this, please refer to Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment of the present application. In this embodiment, steps S231 to S233 are included before step S30:

[0096] Step S231: Collect diverse original job description data, perform anonymization and data cleaning on the original job description data to obtain job description data;

[0097] The original job description data refers to a collection of relevant recruitment data collected from external vendors and historical recruitment data from within the company. The internal data within the original job description data undergoes automated anonymization, such as removing the company name and anonymizing sensitive information like employee numbers and workstation addresses in internal documents. During the data cleaning phase, regular expressions are used to remove HTML tags and invalid symbols, and text similarity algorithms are used to eliminate duplicate job descriptions. Finally, structured fields and unstructured descriptive text are retained to form standardized job description data.

[0098] Step S232: Select a pre-trained model to perform entity recognition on the job description data and annotate key information to obtain an annotated job description dataset;

[0099] Selecting a pre-trained model for entity recognition of the job description data refers to using a deep learning-based pre-trained language model to process the cleaned job description data and identify key information, mainly including job title, job responsibilities, qualifications, educational background, work experience, and skill requirements.

[0100] Labeling key information refers to annotating key entities in job description data based on entity recognition results, forming a structured dataset. The annotation process can employ manual or semi-automatic methods to ensure accuracy and consistency. The annotated dataset not only retains the original text information but also adds entity labels, facilitating subsequent data analysis and processing.

[0101] Specifically, accurate entity recognition and annotation are achieved by combining a pre-trained model with manual verification. Qwen2.5-3B-Instruct is selected as the base model, with domain-specific fine-tuning tailored to the characteristics of job description texts. For example, the pre-trained model extracts the entities "Technical Skills - Distributed Computing - Spark" and "Work Experience - Big Data - 3-5 Years" from the phrase "Requires proficiency in Spark and over 3 years of big data processing experience." The annotation process employs a semi-automated workflow: after model pre-annotation, key fields are reviewed by HR experts, and erroneous labels are corrected using annotation tools, ultimately generating an annotated job description dataset containing entity types, locations, and encoding mappings.

[0102] Step S233: Based on predefined encoding rules and tag dictionary, the labeled job description dataset is mapped to a unified encoding system at multiple levels to form a multi-level encoding mapping dataset, and a dynamic knowledge base is constructed based on the multi-level encoding mapping dataset.

[0103] According to pre-designed coding rules and tag dictionary, the information of each entity in the labeled job description data is mapped to a unified and standardized coding system. This process is achieved through multi-level mapping, such as mapping "work experience" to a coding structure of "industry-years-job type-experience depth", and mapping "skill requirements" to "skill category-skill subcategory-specific skill-application scenario".

[0104] A multi-level encoding mapping dataset refers to a dataset formed after mapping processing, in which each entity's information is converted into a specific code or identifier in a unified encoding system. Building a dynamic knowledge base based on this multi-level encoding mapping dataset means integrating the mapped dataset into a dynamic and updatable knowledge base.

[0105] Specifically, in one feasible implementation, step S233 may include steps S2331 to S2332:

[0106] Step S2331: Construct a hierarchical tag system based on the predefined encoding rules and tag dictionary;

[0107] A hierarchical tagging system is constructed based on the predefined encoding rules and tag dictionary. Following the pre-designed encoding rules and tag dictionary, a hierarchical tag structure is established to classify and encode various entity information in job description data. By decomposing complex job information into multiple levels of tags, the hierarchical tagging system enables more refined and structured data management and retrieval.

[0108] Step S2332: Based on the hierarchical labeling system, the labeled job description dataset is mapped to a unified coding system through the precise matching module and the rule generalization module to form a multi-level coding mapping dataset.

[0109] Based on the hierarchical labeling system, and through the precise matching module and the rule generalization module, the labeled job description dataset is mapped to a unified coding system at multiple levels. By using the hierarchical labeling system and combining precise matching and rule generalization techniques, the entity information in the labeled job description data is mapped to a unified coding system, forming a structured multi-level coding mapping dataset.

[0110] The exact matching module refers to a fast retrieval based on a hash table of a tag dictionary, which can achieve an efficient matching module with a time complexity of O(1). For entity information in the labeled data, if a tag that matches exactly can be found in the tag dictionary, it can be directly mapped to the corresponding code.

[0111] The rule generalization module refers to predicting the entity category of out-of-vocabulary (OV) words (i.e., words for which no exact matching entity information can be found in the tag dictionary) using pre-trained models (such as Qwen2.5-3B-Instruct, Bert) and performing generalization processing according to the rules of the hierarchical tag system. The system also monitors the frequency of OV words. For high-frequency new words, manual review is triggered. After approval, they are added to the tag dictionary, and the model is incrementally trained periodically to maintain its ability to recognize new terms.

[0112] In this embodiment, diverse raw job description data is collected. Entity recognition and key information annotation transform unstructured text data into a structured dataset, preserving not only the original information but also adding entity labels, thus improving data readability and usability. A hierarchical labeling system and multi-level coding mapping integrate the annotated data into a unified coding system, achieving data standardization and normalization. This standardization enables efficient data retrieval and analysis, providing a solid foundation for the construction of a dynamic knowledge base.

[0113] Based on the above embodiments of this application, a fourth embodiment of this application is proposed. In this fourth embodiment, content that is the same as or similar to that in the above embodiments can be referred to the above description, and will not be repeated hereafter.

[0114] Based on this, please refer to Figure 4 , Figure 4 The flowchart provided for the third embodiment of this application is as follows: Figure 4 As shown, in this embodiment, before step S233, which maps the labeled job description dataset to a unified encoding system at multiple levels based on predefined encoding rules and a label dictionary to form a multi-level encoded mapping dataset, step S1 is also included:

[0115] Step S1: Define encoding rules, and based on the encoding rules, perform data source integration and tag standardization processing on the job description data to obtain a tag dictionary.

[0116] Defining coding rules refers to developing a systematic coding system based on the common content and structure of job descriptions. This system is used to classify and encode various types of information in job descriptions. Coding rules typically include multiple levels and categories, such as basic job information, job purpose, main responsibilities, and qualifications. Each category is further subdivided into more specific subcategories and tags.

[0117] Data source integration based on coding rules refers to the unified processing and integration of job description data from different data sources (such as internal enterprise systems, recruitment platforms, industry reports, etc.) according to coding rules. The purpose of data source integration is to eliminate differences between different data sources and ensure data consistency and integrity.

[0118] Tag standardization refers to the unified standardization of various tags in job description data to ensure that the same or similar tags have the same expression and meaning in different data sources. This mainly includes operations such as merging synonyms and building multi-level tree structures.

[0119] A tag dictionary is a collection containing all standard tags and their corresponding codes, used to guide subsequent entity recognition, data mapping, and knowledge base construction processes.

[0120] In this embodiment, the tag dictionary not only improves data consistency and comparability but also provides a reliable reference for subsequent entity recognition, data mapping, and knowledge base construction, ensuring the efficiency and accuracy of the entire optimization process. Through this standardized process, enterprises can more easily manage and optimize job descriptions, improving the efficiency and quality of human resource management.

[0121] This application also provides a job description optimization device; please refer to... Figure 5 The job description optimization device includes:

[0122] Module 10 is used to obtain the job description optimization instructions and the job description to be optimized input by the user;

[0123] The generation module 20 is used to perform natural language semantic parsing on the job description optimization instructions and to perform entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list.

[0124] The retrieval module 30 is used to retrieve benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights. The construction process of the dynamic knowledge base adopts heterogeneous data alignment.

[0125] The optimization module 40 is used to optimize the job description of the job to be optimized item by item based on the benchmark job data, so as to obtain the optimized job description and optimization instructions.

[0126] The job description optimization device provided in this application, employing the job description optimization method in the above embodiments, can solve the technical problem of job description optimization. Compared with the prior art, the beneficial effects of the job description optimization device provided in this application are the same as those of the job description optimization method provided in the above embodiments, and other technical features in the job description optimization device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0127] This application provides a job description optimization device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the job description optimization method in Embodiment 1 above.

[0128] The following is for reference. Figure 6 The diagram illustrates a structural schematic suitable for implementing the job description optimization device in the embodiments of this application. The job description optimization device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The job description and optimized equipment shown are merely examples and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0129] like Figure 6As shown, the job description optimization device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the job description optimization device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the job description optimization device to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows a job description optimization device with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0130] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0131] The job description optimization device provided in this application, employing the job description optimization method described in the above embodiments, can solve the technical problem of job description optimization. Compared with the prior art, the beneficial effects of the job description optimization device provided in this application are the same as those of the job description optimization method provided in the above embodiments, and other technical features of the job description optimization device are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0132] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0133] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0134] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the job description optimization method in the above embodiments.

[0135] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0136] The aforementioned computer-readable storage medium may be included in the job description optimization device; or it may exist independently and not be assembled into the job description optimization device.

[0137] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the job description optimization device, the job description optimization device performs the following actions: acquires a user-input job description optimization instruction and a job description to be optimized; performs natural language semantic parsing on the job description optimization instruction and entity extraction and semantic analysis on the job description to be optimized to obtain a candidate job list; retrieves benchmark job data from a pre-built dynamic knowledge base based on the candidate job list and matching weights, wherein the construction process of the dynamic knowledge base employs heterogeneous data alignment; and optimizes the job description to be optimized item by item according to the benchmark job data to obtain an optimized job description and optimization instructions.

[0138] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0139] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0140] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0141] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described job description optimization method, and is capable of solving the technical problem of job description optimization. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the job description optimization method provided in the above embodiments, and will not be repeated here.

[0142] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the job description optimization method described above.

[0143] The computer program product provided in this application can solve the technical problem of job description optimization. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the job description optimization method provided in the above embodiments, and will not be repeated here.

[0144] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for optimizing job descriptions, characterized in that, The methods for optimizing the job description include: Obtain the user-inputted job description optimization instructions and the job description to be optimized; The job description optimization instructions are parsed using natural language semantics, and entity extraction and semantic analysis are performed on the job descriptions to be optimized to obtain a candidate job list. Collect diverse original job description data, and perform anonymization and data cleaning on the original job description data to obtain job description data; A pre-trained model is selected to perform entity recognition on the job description data and key information is labeled to obtain a labeled job description dataset. Based on predefined encoding rules and tag dictionaries, the labeled job description dataset is mapped to a unified encoding system at multiple levels to form a multi-level encoding mapping dataset, and a dynamic knowledge base is constructed based on the multi-level encoding mapping dataset; Based on the candidate job list and matching weights, benchmark job data is retrieved from a pre-built dynamic knowledge base. The construction process of the dynamic knowledge base adopts heterogeneous data alignment, including: adjusting the matching weights using a reinforcement learning model; retrieving benchmark job data that matches the candidate job list from the pre-built dynamic knowledge base according to the adjusted matching weights; recording feedback data of the benchmark job data; and triggering online incremental training of the reinforcement learning model to update the matching weights based on the feedback data. The job description to be optimized is optimized item by item based on the benchmark job data to obtain the optimized job description and optimization instructions, including: a comparative analysis of each item of the job description to be optimized with the benchmark job data.

2. The job description optimization method as described in claim 1, characterized in that, The steps of performing natural language semantic parsing on the job description optimization instructions and performing entity extraction and semantic analysis on the job descriptions to be optimized to obtain a candidate job list include: The job title is extracted from the entity of the job description to be optimized, and the optimization instructions of the job description are parsed using natural language to extract the user's optimization intent. The pre-built large model is invoked to perform semantic analysis on the user's optimization intent and the job description to be optimized, and explicit and implicit conditions are extracted respectively; the explicit conditions include the job title, and the implicit conditions include the combination of cross-team collaboration ability and project management experience. A candidate job list is generated based on the explicit conditions, the implicit conditions, and the job title.

3. The job description optimization method as described in claim 1, characterized in that, Before the step of mapping the labeled job description dataset to a unified encoding system at multiple levels based on predefined encoding rules and a label dictionary to form a multi-level encoded mapping dataset, the following steps are also included: Define encoding rules, and based on the encoding rules, perform data source integration and tag standardization processing on the job description data to obtain a tag dictionary.

4. The job description optimization method as described in claim 1, characterized in that, The steps of mapping the labeled job description dataset to a unified encoding system at multiple levels based on predefined encoding rules and a label dictionary to form a multi-level encoded mapping dataset include: A hierarchical tag system is constructed based on the predefined encoding rules and tag dictionary; Based on the hierarchical labeling system, the labeled job description dataset is mapped to a unified coding system through a precise matching module and a rule generalization module, forming a multi-level coding mapping dataset.

5. The job description optimization method as described in claim 1, characterized in that, The step of optimizing the job description of the job to be optimized item by item based on the benchmark job data to obtain the optimized job description and optimization instructions includes: The job descriptions to be optimized and the benchmark job data are compared and analyzed item by item to generate optimization suggestions; The job description to be optimized is optimized item by item according to the optimization suggestions to obtain the optimized job description, and an optimization description is generated based on the optimized job description.

6. A job description optimization device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the job description optimization method as described in any one of claims 1 to 5.

7. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the job description optimization method as described in any one of claims 1 to 5.

8. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the job description optimization method as described in any one of claims 1 to 5.

Images