Talent big data platform system

Abstract

The invention relates to the technical field of big data matching, and particularly discloses a talent big data platform system, which adopts a natural language processing technology based on deep learning to perform semantic analysis on job hunting information and job hunting intention plans issued by job hunters and recruitment demands issued by recruiters, so as to obtain a big data analysis result. The method comprises the following steps: respectively extracting semantic feature representations of job hunting information, job hunting intention planning and recruitment demand information, and then respectively carrying out semantic conjoint analysis on the job hunting information and the job hunting intention planning of job hunters and recruitment demands; according to the method, the matching degree between the job hunters and the requirements of the recruiters in the two aspects of working experience and skills and occupational planning is revealed, so that more comprehensive talent matching recommendation is realized, the accuracy and efficiency of talent matching can be improved, the requirements of enterprises for talents are met, and meanwhile, the job hunting success rate of the job hunters is favorably improved.

Description

Technical Field

[0001] This application relates to the field of big data matching technology, and more specifically, to a talent big data platform system. Background Art

[0002] With the rapid development of information technology and the wide popularization of Internet applications, the talent market has witnessed unprecedented changes. Traditional recruitment methods, such as job fairs, newspaper advertisements, headhunting services, etc., although meeting the recruitment needs of enterprises to a certain extent, have increasingly prominent problems such as low efficiency, information asymmetry, and low matching accuracy. Especially in the current highly competitive market environment, the demand of enterprises for talents is becoming more urgent, and the requirements for recruitment efficiency and matching accuracy are also getting higher and higher.

[0003] Under this background, the talent big data platform system came into being. By collecting, integrating, and analyzing a large amount of job hunting information and recruitment requirements, and using advanced information technology means, it realizes the efficient and accurate matching between job seekers and recruiters, thereby providing better services for both the supply and demand sides of the talent market. However, most of the existing talent big data platforms often rely on relatively basic technical means such as keyword matching and resume screening when processing job hunting information and recruitment requirements. These methods have great limitations when dealing with complex job hunting information and recruitment requirements, and it is difficult to deeply explore the deep matching relationship between job seekers and recruiters, resulting in inaccurate matching results.

[0004] Therefore, an optimized talent big data platform system is expected. Summary of the Invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiments of this application provide a talent big data platform system, which uses natural language processing technology based on deep learning to perform semantic parsing on the job hunting information and job hunting intention planning released by job seekers, and the recruitment requirements released by recruiters, so as to respectively extract the semantic feature representations of job hunting information, job hunting intention planning, and recruitment requirement information. Furthermore, by performing semantic joint analysis on the job hunting information and job hunting intention planning of job seekers with the recruitment requirements respectively, the matching degree of job seekers in terms of work experience skills and career planning with the recruiter's requirements is revealed, so as to achieve a more comprehensive talent matching recommendation. In this way, the accuracy and efficiency of talent matching can be improved, the needs of enterprises for talents can be met, and at the same time, it helps to improve the job hunting success rate of job seekers.

[0006] Correspondingly, according to one aspect of this application, a talent big data platform system is provided, which includes: A job seeker information extraction module, configured to extract the job hunting information and job hunting intention planning released by a first job seeker from the job hunting module; A recruitment requirement information extraction module, which is used to extract the recruitment requirements released by the first recruitment party from the recruitment module; A semantic encoding module, which is used to perform semantic encoding on the job application information, job application intention plan released by the first job applicant, and the recruitment requirements released by the first recruitment party respectively to obtain a job application information semantic encoding vector, a job application intention plan semantic encoding vector, and a recruitment requirement semantic encoding vector; A job application information - recruitment requirement semantic joint analysis module, which is used to perform a job application information - recruitment requirement semantic joint analysis driven by bidirectional attention on the job application information semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job application information - recruitment requirement semantic joint interaction encoding vector; A job application plan - recruitment requirement semantic joint analysis module, which is used to perform a job application plan - recruitment requirement semantic joint analysis driven by bidirectional attention on the job application intention plan semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job application plan - recruitment requirement semantic joint interaction encoding vector; An information push module, which is used to determine whether to push the job application information and job application intention plan released by the first job applicant to the first recruitment party based on the job application information - recruitment requirement semantic joint interaction encoding vector and the job application plan - recruitment requirement semantic joint interaction encoding vector.

[0007] Compared with the prior art, the talent big data platform system provided by this application uses natural language processing technology based on deep learning to perform semantic parsing on the job application information and job application intention plan released by job applicants, and the recruitment requirements released by recruitment parties, so as to respectively extract the semantic feature representations of job application information, job application intention plans, and recruitment requirement information. Furthermore, by performing semantic joint analysis on the job application information and job application intention plan of job applicants with the recruitment requirements respectively, the matching degree of job applicants with the recruitment party's requirements in terms of work experience skills and career planning is revealed, thereby realizing a more comprehensive talent matching recommendation. In this way, the accuracy and efficiency of talent matching can be improved, meeting the enterprise's demand for talents, and at the same time helping to improve the job application success rate of job applicants. Brief Description of the Drawings

[0008] By describing the embodiments of the present application in more detail in combination with the drawings, the above - mentioned and other purposes, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application, and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 It is a block diagram of the talent big data platform system according to the embodiment of the present application.

[0010] Figure 2 It is a schematic diagram of data flow of the talent big data platform system according to an embodiment of the present application.

[0011] Figure 3 It is a block diagram of the semantic joint analysis module for job hunting information and recruitment requirements in the talent big data platform system according to an embodiment of the present application.

[0012] Figure 4 It is a block diagram of the information push module in the talent big data platform system according to an embodiment of the present application. Detailed implementation manners

[0013] As shown in the present application and the claims, unless the context clearly indicates an exception, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the clearly identified steps and elements, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0014] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or the server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0015] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the operations before or below are not necessarily executed precisely in order. On the contrary, as needed, various steps can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0016] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0017] It should be noted that all actions of obtaining data in the present application are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where the location is located and obtaining the authorization given by the owner of the corresponding device.

[0018] In view of the technical problems described in the above background art, the present application proposes a talent big data platform system, which uses natural language processing technology based on deep learning to semantically analyze the job hunting information and job hunting intention plans published by job seekers, as well as the recruitment requirements published by recruiters, so as to extract the semantic feature representations of job hunting information, job hunting intention plans, and recruitment requirement information respectively. Furthermore, by semantically jointly analyzing the job hunting information and job hunting intention plans of job seekers with the recruitment requirements respectively, the matching degree of job seekers with the requirements of recruiters in terms of work experience skills and career planning is revealed, thereby realizing a more comprehensive talent matching recommendation. In this way, the accuracy and efficiency of talent matching can be improved, the needs of enterprises for talents can be met, and at the same time, it helps to improve the job hunting success rate of job seekers.

[0019] Figure 1 FIG. is a block diagram of a talent big data platform system according to an embodiment of the present application. Figure 2 FIG. is a schematic diagram of data flow of a talent big data platform system according to an embodiment of the present application. As Figure 1 and Figure 2 shown, the talent big data platform system 100 includes: a job seeker information extraction module 110, configured to extract the job hunting information and job hunting intention plans published by a first job seeker from a job hunting module; a recruitment requirement information extraction module 120, configured to extract the recruitment requirements published by a first recruiter from a recruitment module; a semantic encoding module 130, configured to perform semantic encoding on the job hunting information and job hunting intention plans published by the first job seeker and the recruitment requirements published by the first recruiter respectively to obtain a job hunting information semantic encoding vector, a job hunting intention plan semantic encoding vector, and a recruitment requirement semantic encoding vector; a job hunting information - recruitment requirement semantic joint analysis module 140, configured to perform a job hunting information - recruitment requirement semantic joint analysis based on bidirectional attention on the job hunting information semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job hunting information - recruitment requirement semantic joint interaction encoding vector; a job hunting plan - recruitment requirement semantic joint analysis module 150, configured to perform a job hunting plan - recruitment requirement semantic joint analysis based on bidirectional attention on the job hunting intention plan semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job hunting plan - recruitment requirement semantic joint interaction encoding vector; an information push module 160, configured to determine whether to push the job hunting information and job hunting intention plans published by the first job seeker to the first recruiter based on the job hunting information - recruitment requirement semantic joint interaction encoding vector and the job hunting plan - recruitment requirement semantic joint interaction encoding vector.

[0020] In the above-mentioned talent big data platform system, the job seeker information extraction module 110 is used to extract the job seeking information and job intention plan released by the first job seeker from the job seeking module. It should be understood that the job seeking information includes information such as the work experience, skill level, and educational background of the first job seeker, which is a comprehensive description of the personal ability and professional background of the first job seeker; while the job intention plan includes information such as the position, salary level, work location, and personal career plan expected by the first job seeker, reflecting the expectations of the first job seeker for future career development. By obtaining the job seeking information and job intention plan released by the job seeker as the basis for job recommendation, this application can comprehensively understand the professional background, skill level, and career plan of the job seeker, thus providing strong support for subsequent job matching.

[0021] Specifically, the work experience part in the job seeking information details the past career experiences of the first job seeker. This part of the information usually covers the names of the enterprises where the job seeker has worked, the departments, the specific positions held, the length of service, the main responsibilities, and the achievements obtained. These details help the system evaluate the professional field, industry experience, and project management ability of the first job seeker, etc., so as to judge whether they meet the requirements of a specific position. For example, if a job seeker has worked as a software engineer in a technology company and successfully led the team to complete several important projects, this indicates strong technical strength and leadership skills, and may be very suitable for positions such as technical director or senior development manager. The description of the skill level focuses on the specific skills mastered by the first job seeker, whether they are hard skills or soft skills. These skills are the core elements for the first job seeker to be competent for the target position and are also the key points that enterprises focus on during recruitment. The information on educational background is also indispensable, including the graduating institution, the major studied, the degree, and the relevant certificates and awards obtained during school. These information provide important clues for understanding the knowledge structure of the first job seeker. Degrees from top universities are often regarded as symbols of high intelligence and strong learning ability, and the background of certain specific majors also means in-depth understanding and professional knowledge in related fields. For example, graduates majoring in computer science or information technology will naturally be more competitive when applying for positions in the IT industry.

[0022] In addition to the information reflecting the current status, the job intention plan demonstrates the expectations of the first job seeker for future career development. This part of the content not only includes the expected position, salary level, and work location, but also involves more long-term career planning. For example, the first job seeker may hope to be promoted to the management level within the next few years, or seek career development opportunities across departments or even across national boundaries. The job intention plan reflects career aspirations and personal development goals, and is of guiding significance for predicting future career paths. In addition, the geographical location preference in the job intention plan is also very important. Different cities or regions may offer different types of employment opportunities and living environments. Some job seekers may be inclined to the broad development space and generous salary offered by big cities, while others may choose to work in cities near their hometown due to family reasons or other personal factors. Understanding geographical location preferences can help the system more accurately screen out eligible positions, improving job search efficiency and satisfaction.

[0023] In the above-mentioned talent big data platform system, the recruitment requirement information extraction module 120 is used to extract the recruitment requirements published by the first recruiter from the recruitment module. It should be understood that the recruitment requirements include the specific requirements of the first recruiter for the required position, such as information on job responsibilities, skill requirements, educational background, work experience, etc., which is a detailed description of the employment needs of the first recruiter. By obtaining the recruitment requirements published by the recruiter, the employment standards and expectations of the recruiter can be accurately understood, providing an important basis for subsequent matching and recommendation.

[0024] Specifically, the part of the recruitment requirements regarding job responsibilities elaborates in detail the work content and scope of responsibilities of the specific position. This part of the information usually includes daily tasks, project management responsibilities, and requirements for collaboration with other departments or teams, etc. For example, in the job description of a software engineer, it may be clearly stated that one needs to be responsible for code writing, software testing, maintaining the existing system, and participating in the development of new products, etc. The clear definition of job responsibilities helps to understand the nature and intensity of the work, ensuring that job seekers possess the skills and experience required to complete the tasks. In addition, job responsibilities also reveal a part of the internal operation mode of the enterprise, allowing job seekers to evaluate whether they are suitable for this working environment.

[0025] Skill requirements are another important part of the recruitment needs, which are directly related to whether job seekers can be competent for the target position. Skill requirements cover two major categories: hard skills and soft skills. Hard skills refer to professional techniques and the ability to use tools, such as programming languages (Python, Java), design software (Adobe Illustrator, AutoCAD), etc.; while soft skills involve communication skills, teamwork spirit, problem-solving ability, and leadership, etc. Different types of positions have different emphases on skill requirements. For example, a marketing manager may need to have excellent communication skills and data analysis ability to develop effective marketing strategies; while a data scientist should be proficient in machine learning algorithms and big data processing technologies. By carefully analyzing the skill requirements, the system can screen out the candidates who best meet the position needs.

[0026] Educational background is also an important consideration in recruitment needs, which records the requirements for the candidate's academic level, major, and relevant certificates. Degrees from top universities are often regarded as symbols of high intelligence and strong learning ability, while a background in a specific major means in-depth understanding and professional knowledge in the relevant field. For example, graduates majoring in computer science or information technology will naturally be more competitive when applying for positions in the IT industry. In addition, some industries also require specific professional qualification certificates, such as the Certified Public Accountant (CPA), which is of great importance for financial positions. Information on educational background provides important clues for understanding the knowledge structure of job seekers, helping the system better evaluate whether candidates meet the expectations of the first employer.

[0027] Work experience reflects the degree of importance that enterprises attach to the candidate's past career experiences in recruitment needs. The work experience section usually covers the name of the company where the candidate has worked, the department, the specific position held, the length of service, the main responsibilities, and the achievements obtained, etc. These details help to evaluate the job seeker's professional field, industry experience, and project management ability, etc., so as to judge whether they meet the requirements of a specific position. For example, a job seeker who has worked as a software engineer in a technology company and successfully led the team to complete several important projects will undoubtedly be a very attractive candidate for employers looking for senior development managers or technical directors. Work experience is not only an important indicator for measuring professional skills but also can reflect the career growth trajectory and development potential of job seekers.

[0028] In addition to the information reflecting specific job requirements mentioned above, the recruitment needs also include descriptions of aspects such as salary range, welfare benefits, work location, and working hours. The setting of the salary range directly affects the ability to attract high-quality talents, and a reasonable compensation system can effectively enhance the interest of job seekers. Welfare benefits such as medical insurance, paid vacations, and employee training are also one of the important factors considered by job seekers when choosing a position. The choice of work location cannot be ignored either. Different cities or regions may offer different types of employment opportunities and living environments. Some job seekers may prefer the broad development space and generous salary offered by big cities, while others may choose to work in a city near their hometown due to family reasons or other personal factors. Understanding geographical location preferences can help the system more accurately screen out eligible positions, improving job search efficiency and satisfaction.

[0029] The recruitment needs also involve content such as corporate culture, values, and long-term development goals. Corporate culture is the code of conduct and values commonly followed within an enterprise, which profoundly affects employees' work attitudes and the effectiveness of teamwork. For example, an Internet company that emphasizes innovation and technology-driven may prefer job seekers who are creative and dare to try new technologies. Long-term development goals reflect the strategic planning and vision of the enterprise in the coming period. Job seekers can evaluate whether their career development direction aligns with the enterprise by understanding this information. When personal career planning is consistent with the long-term development of the enterprise, it is easier to establish a stable and productive cooperative relationship.

[0030] In the above talent big data platform system, the semantic encoding module 130 is used to perform semantic encoding on the job seeking information, job seeking intention plan released by the first job seeker, and the recruitment requirements released by the first recruitment party respectively to obtain a job seeking information semantic encoding vector, a job seeking intention plan semantic encoding vector, and a recruitment requirement semantic encoding vector. In a specific example of the present application, the semantic encoding module 130 is used to: use a semantic encoder based on the BERT model to perform semantic encoding on the job seeking information, the job seeking intention plan, and the recruitment requirements respectively to obtain the job seeking information semantic encoding vector, the job seeking intention plan semantic encoding vector, and the recruitment requirement semantic encoding vector. In particular, considering that traditional talent matching and recommendation methods usually adopt simple keyword matching and often cannot accurately understand the deep semantic association between job seeking information and recruitment requirements. Therefore, in order to achieve an in-depth understanding of personnel job seeking information and recruitment requirements, the present application adopts the BERT model with excellent natural language understanding and representation capabilities as the semantic encoder to perform semantic encoding on the job seeking information, the job seeking intention plan, and the recruitment requirements respectively to obtain a job seeking information semantic encoding vector, a job seeking intention plan semantic encoding vector, and a recruitment requirement semantic encoding vector. Those of ordinary skill in the art should be aware that the BERT model adopts a multi-layer bidirectional Transformer encoder structure and can utilize the context information on both sides of each word in the sentence at the same time. When performing semantic encoding on the job seeking information, the job seeking intention plan, and the recruitment requirements respectively, the BERT model will convert the input text into a series of fixed-length vectors, which not only contain the meaning of the words themselves but also incorporate the context information of their positions. In this way, it can effectively capture the deep semantic association between job seeking information and recruitment requirements, rather than just based on the surface similarity of keywords.

[0031] In the above talent big data platform system, the job seeking information - recruitment requirement semantic joint analysis module 140 is used to perform a job seeking information - recruitment requirement semantic joint analysis driven by bidirectional attention on the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job seeking information - recruitment requirement semantic joint interaction encoding vector. It should be understood that considering that there may be different matching relationships between job seekers and the requirements of the recruitment party in terms of their own work experience, skills, and career planning, therefore, in order to perform a more detailed matching analysis between job seekers and the recruitment party, in this application, the job seeking information semantic encoding vector and the job seeking intention planning semantic encoding vector are respectively subjected to semantic interaction analysis with the recruitment requirement semantic encoding vector, so as to more finely reveal the matching degree between job seekers and the recruitment party in terms of work experience, skills, and career planning. In particular, in order to improve the accuracy and efficiency of semantic interaction analysis, this application proposes a semantic interaction analysis method driven by bidirectional attention, which automatically focuses on the most critical and relevant semantic information parts in both parties' data through bidirectional attention calculation, so as to accurately mine the deep - level matching relationship between job seekers and the recruitment party and improve the accuracy of matching analysis. Among them, Figure 3 is a block diagram of the job seeking information - recruitment requirement semantic joint analysis module in the talent big data platform system according to an embodiment of the present application. As Figure 3 shown, the job seeking information - recruitment requirement semantic joint analysis module 140 includes: a bidirectional attention calculation unit 141, which is used to calculate the bidirectional attention scores between the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job seeking information - recruitment requirement semantic association positive attention score field and a job seeking information - recruitment requirement semantic association negative attention score field; a bidirectional attention balance field construction unit 142, which is used to construct a job seeking information - recruitment requirement semantic association positive - negative bidirectional attention balance field based on the job seeking information - recruitment requirement semantic association positive attention score field and the job seeking information - recruitment requirement semantic association negative attention score field; and a semantic mapping interaction analysis unit 143, which is used to perform semantic mapping interaction analysis on the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector based on the job seeking information - recruitment requirement semantic association positive - negative bidirectional attention balance field to obtain the job seeking information - recruitment requirement semantic joint interaction encoding vector.

[0032] Specifically, in a specific example of this application, the bidirectional attention calculation unit 141 is used to: respectively perform a homography projection transformation on the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job seeking information semantic homography projection encoding vector and a recruitment requirement semantic homography projection encoding vector, which is expressed by the formula: Among them, represents the semantic encoding vector of the job application information, represents the semantic encoding vector of the recruitment requirement, and are different homography projection matrices, and respectively represent the homography projection encoding vector of the job application information semantics and the homography projection encoding vector of the recruitment requirement semantics.

[0033] That is, in the semantic interaction process between the semantic encoding vector of the job application information and the semantic encoding vector of the recruitment requirement, through the implementation of the homography projection transformation, the semantic encoding vector of the job application information and the semantic encoding vector of the recruitment requirement are mapped to a shared semantic feature space, so as to ensure their comparability at the semantic level.

[0034] In a specific example of the present application, the bidirectional attention calculation unit 141 is further configured to: calculate the forward attention score field of the homography projection encoding vector of the job application information semantics relative to the homography projection encoding vector of the recruitment requirement semantics to obtain the job application - recruitment requirement semantic association forward attention score field, which is expressed by the formula: Among them, represents the transpose of the vector, represents the matrix multiplication operation, is the attention score scaling factor, represents the job application - recruitment requirement semantic association forward attention score field.

[0035] That is, after the homography projection transformation is completed, the forward attention score field of the semantic encoding vector of the job application information and the semantic encoding vector of the recruitment requirement is calculated. From the perspective of the job application information, it reveals the importance of each part of the features of the job application information when interacting semantically with the recruitment requirement information. By analyzing which job application information features are highly relevant to the recruitment requirement, the system can generate an attention weight distribution to form a forward attention score field. In this way, in the subsequent semantic interaction, it is possible to focus on those job application information features that are most relevant to the recruitment requirement, ensuring that the matching process is more accurate and effective.

[0036] In a specific example of the present application, the bidirectional attention calculation unit 141 is further configured to: calculate the reverse attention score field of the homography projection encoding vector of the recruitment requirement semantics relative to the homography projection encoding vector of the job application information semantics to obtain the job application - recruitment requirement semantic association reverse attention score field, which is expressed by the formula: Among them, represents the reverse attention score field of the semantic association between job application information and recruitment requirements.

[0037] That is, in order to comprehensively evaluate the matching relationship between job applicants and recruiters, the reverse attention score field is further calculated from the perspective of recruitment requirements. In this way, when revealing the matching between recruitment requirements and job application information, it can be determined which specific requirements highly match the skills and experience of job applicants. Through this analysis, the system can generate the reverse attention weight distribution to form the reverse attention score field. This helps to highlight the requirements in the recruitment needs that are most relevant to the background of job applicants, ensuring that these key factors are fully considered during the matching process.

[0038] Specifically, in a specific example of this application, the two-way attention balance field construction unit 142 is used to: input the forward attention score field of the semantic association between job application information and recruitment requirements and the reverse attention score field of the semantic association between job application information and recruitment requirements into the attention balance modulation module based on a convolutional neural network to obtain the forward and reverse two-way attention balance field of the semantic association between job application information and recruitment requirements, which is expressed by the formula: Among them, represents concatenation, represents the forward and reverse two-way attention balance field of the semantic association between job application information and recruitment requirements, represents 3 3 convolution operations.

[0039] That is, in order to ensure that the most critical and relevant semantic information in both parties' data can be comprehensively captured during the final semantic interaction process, this application further performs attention balance modulation on the forward attention score field and the reverse attention score field to construct the forward and reverse two-way attention balance field of the semantic association between job application information and recruitment requirements. Thus, by comprehensively considering the forward and reverse attention information, it is ensured that in the final semantic interaction coding, neither party will be overly biased, nor will any important information be ignored, better balancing the semantic association interaction relationship between job application information and recruitment requirements.

[0040] Specifically, in a specific example of this application, the semantic mapping interaction analysis unit 143 is used to: map the homography projection coding vector of job application information and the homography projection coding vector of recruitment requirements to the forward and reverse two-way attention balance field of the semantic association between job application information and recruitment requirements to obtain the homography projection attention modulation coding vector of job application information and the homography projection attention modulation coding vector of recruitment requirements, which is expressed by the formula: in, and They represent the job search information semantic homography projection attention modulation coding vector and the recruitment demand semantic homography projection attention modulation coding vector respectively; The job search information-recruitment demand semantics joint interaction coding vector is obtained by calculating the position point-by-position division between the job search information semantics homography projection attention modulation coding vector and the recruitment demand semantics homography projection attention modulation coding vector, which is expressed as: in, Represents the job search information-recruitment demand semantic joint interaction encoding vector.

[0041] That is, after completing the homography projection transformation, the job search information semantic coding vector and the recruitment demand semantic coding vector are projected into the job search information-recruitment demand semantic association positive and negative bidirectional attention balance field respectively, and the feature representations of the two are adjusted according to the attention weight distribution in the balance field. In this way, the feature representations with significant correlation can be enhanced, and these key features can be highlighted in the final semantic interaction coding, while the features with weak correlation are weakened accordingly. In this way, it is possible to more effectively focus on the most critical and matching potential semantic information parts of both data. The adjusted feature representation generates the job search information semantic homography projection attention modulation coding vector and the recruitment demand semantic homography projection attention modulation coding vector. Finally, through the position-by-position point division operation, the job search information semantic homography projection attention modulation coding vector and the recruitment demand semantic homography projection attention modulation coding vector are subjected to element-by-element semantic interaction response coding, and the job search information-recruitment demand semantic joint interaction coding vector is obtained, thereby realizing the accurate capture and representation of the deep semantic association between job search information and recruitment demand.

[0042] In the talent big data platform system, the job search plan recruitment demand semantic joint analysis module 150 is used to perform a job search plan-recruitment demand semantic joint analysis based on a two-way attention-driven semantic coding vector of the job search intention plan semantic coding vector and the recruitment demand semantic coding vector to obtain a job search plan-recruitment demand semantic joint interactive coding vector. Specifically, in the semantic interaction analysis process between the job search intention plan semantic coding vector and the recruitment demand semantic coding vector, the deep-level semantic association relationship between the job search intention plan and the recruitment demand is also excavated through the construction of the two-way attention balance field and the element-by-element semantic interaction response coding method, so as to obtain a job search plan-recruitment demand semantic joint interactive coding vector, thereby providing another important basis for the association matching between job seekers and recruiters.

[0043] In the above talent big data platform system, the information push module 160 is configured to determine whether to push the job seeking information and job seeking intention plan published by the first job seeker to the first recruiter based on the job seeking information - recruitment requirement semantic joint interaction coding vector and the job seeking plan - recruitment requirement semantic joint interaction coding vector. Among them, Figure 4 is a block diagram of the information push module in the talent big data platform system according to an embodiment of the present application. As Figure 4 shown, the information push module 160 includes: a comprehensive matching unit 161, configured to calculate the comprehensive matching degree between the first job seeker and the first recruiter based on the job seeking information - recruitment requirement semantic joint interaction coding vector and the job seeking plan - recruitment requirement semantic joint interaction coding vector; a push unit 162, configured to determine whether to push the job seeking information and job seeking intention plan published by the first job seeker to the first recruiter based on the comparison between the comprehensive matching degree and a preset threshold.

[0044] Specifically, in a specific example of the present application, the comprehensive matching unit 161 is configured to: input the job seeking information - recruitment requirement semantic joint interaction coding vector and the job seeking plan - recruitment requirement semantic joint interaction coding vector into an integrated decision module including a first classifier and a second classifier to obtain a first matching degree and a second matching degree. Specifically, the first classifier is used to perform feature learning on the job seeking information - recruitment requirement semantic joint interaction coding vector to accurately evaluate the matching degree between the job seeking information and the recruitment requirement according to the semantic association information contained in the job seeking information - recruitment requirement semantic joint interaction coding vector, and output the matching degree between the two, that is, the first matching degree. The second classifier is used to perform feature learning and matching probability calculation on the job seeking plan - recruitment requirement semantic joint interaction coding vector to output the second matching degree. Through this integrated decision-making method, the matching situation between the job seeker and the recruiter's requirements in the two dimensions of the current work experience and skills status and the future career plan can be effectively quantified, so as to provide a more comprehensive and accurate matching evaluation for subsequent talent recommendation.

[0045] Preferably, considering that the job application information semantic coding vector and the recruitment requirement semantic coding vector respectively represent the semantic coding features of job application information and recruitment requirements, when performing interactive responses based on the positive and negative attention fields of features, the differences in the population attributes of the semantic features encoded by different feature patterns in the source semantics will cause differences in the fairness of the reinforcement levels of the positive and negative attention fields, thus affecting the inclusiveness of the feature distribution interactive response of the job application information - recruitment requirement semantic joint interactive coding vector. The same technical problem also exists between the job intention plan and the recruitment requirement semantic coding vector, which reduces the accuracy of the first matching degree and the second matching degree obtained by inputting the job application information - recruitment requirement semantic joint interactive coding vector and the job plan - recruitment requirement semantic joint interactive coding vector into an integrated decision-making module including a first classifier and a second classifier.

[0046] Therefore, in a preferred example of the present application, when inputting the job application information - recruitment requirement semantic joint interactive coding vector and the job plan - recruitment requirement semantic joint interactive coding vector into an integrated decision-making module including a first classifier and a second classifier to obtain the first matching degree and the second matching degree, the job application information - recruitment requirement semantic joint interactive coding vector is optimized. The specific optimization process is as follows: Based on the feature association information between any two positions in the job application information - recruitment requirement semantic joint interactive coding vector, a job application information - recruitment requirement semantic joint interactive search reinforcement matrix and a job application information - recruitment requirement semantic joint interactive feedback reinforcement matrix are constructed, expressed as: Where, and represent the feature values of any two positions in the job application information - recruitment requirement semantic joint interactive coding vector, represents the activation function, represents the value at the position in the job application information - recruitment requirement semantic joint interactive search reinforcement matrix, represents the value at the position in the job application information - recruitment requirement semantic joint interactive feedback reinforcement matrix.

[0047] Based on the job application information - recruitment requirement semantic joint interactive search reinforcement matrix and the job application information - recruitment requirement semantic joint interactive feedback reinforcement matrix, the job application information - recruitment requirement semantic joint interactive coding vector is enhanced in a dual-mode driven by search and feedback self-adaptation to obtain a job application information - recruitment requirement semantic joint interactive search enhanced coding vector and a job application information - recruitment requirement semantic joint interactive feedback enhanced coding vector, expressed as: Among them, represents vector addition, represents matrix multiplication, represents the Hadamard product, represents the transpose of a vector, represents the natural exponential function, represents the job information - recruitment requirement semantic joint interactive search enhanced coding vector, represents the job information - recruitment requirement semantic joint interactive feedback enhanced coding vector.

[0048] Fuse the job information - recruitment requirement semantic joint interactive search enhanced coding vector and the job information - recruitment requirement semantic joint interactive feedback enhanced coding vector to obtain an optimized job information - recruitment requirement semantic joint interactive coding vector, denoted as: Among them, and represent weight hyperparameters, represents the optimized job information - recruitment requirement semantic joint interactive coding vector.

[0049] Meanwhile, optimize the job planning - recruitment requirement semantic joint interactive coding vector in the same way to obtain an optimized job planning - recruitment requirement semantic joint interactive coding vector.

[0050] That is, in this preferred embodiment, a dual - mode enhancement mechanism for index - driven configuration and feedback - adaptive configuration is constructed through the superimposed representation tensor and differential feature mapping set generated by the cluster - integrated quantization analysis corresponding to the attribute precision hierarchical regional distribution of the job information - recruitment requirement semantic joint interactive coding vector. In this way, through the non - anchored configuration index response architecture of the cluster - integrated superposition - differential attribute topology in the open probability space, combined with the feedback signal fusion strategy, the configuration feedback redundancy caused by regionalized over - limit attributes is effectively suppressed to optimize the accurate probability regression of the feature space while realizing the feature fidelity constraint. In this way, the accuracy of the first matching degree and the second matching degree obtained by the integrated decision - making module whose input includes the first classifier and the second classifier is improved.

[0051] In a specific example of the present application, the comprehensive matching unit 161 is further configured to: calculate the mean value of the first matching degree and the second matching degree as the comprehensive matching degree. Specifically, in order to comprehensively reflect the matching situation between the job seeker and the recruiter, the present application further calculates the mean value of the first matching degree and the second matching degree as the comprehensive matching degree, so as to comprehensively consider the matching degree between the current work experience and skills of the job seeker and the recruitment requirements, as well as the compatibility between the career plan of the job seeker and the recruitment requirements, thereby providing a more comprehensive and objective matching evaluation index.

[0052] Specifically, the first matching degree reflects the direct matching degree between the job seeking information and the recruitment requirements. This part of the calculation is based on the comparison results of the hard conditions such as the existing work experience, professional skills, and educational background of the job seeker with the requirements clearly listed in the job description. For example, in a software engineer position, if the recruitment requirements state that proficiency in the Python programming language is required and at least three years of relevant work experience is needed, then a job seeker with four years of Python development experience and having participated in multiple large-scale projects will obviously obtain a higher first matching degree score. This direct comparison can quickly screen out candidates who meet the basic qualifications and lay the foundation for further in-depth analysis.

[0053] However, relying solely on the first matching degree cannot fully capture the complex relationship between the job seeker and the position. Therefore, the concept of the second matching degree is introduced, aiming to measure the compatibility between the career plan of the job seeker and the recruitment requirements, and evaluate whether the career development direction of the job seeker is consistent with the long-term goals of the enterprise. For example, a job seeker who hopes to be promoted to the technical management level and participate in multinational projects within the next five years is undoubtedly an ideal candidate for those enterprises that are expanding their international markets and technical teams.

[0054] To calculate the comprehensive matching degree, the system combines the first matching degree and the second matching degree and takes the mean value of the two as the final evaluation result. This method ensures a dual consideration of the current capabilities and future potential of the job seeker, avoiding the situation of only relying on surface conditions and ignoring the possibility of deep-level matching. In this way, even if some job seekers' current experience and skills do not fully meet the job requirements, but if they show a strong willingness to learn, growth potential, and a career plan highly consistent with the enterprise's strategic direction, they may become highly valuable potential talents. Such a comprehensive evaluation method not only improves the accuracy of job recommendations but also brings more cooperation opportunities for both enterprises and job seekers.

[0055] Specifically, the pushing unit 162 is configured to determine whether to push the job seeking information and job intention plan published by the first job seeker to the first recruiter based on the comparison between the comprehensive matching degree and a preset threshold. That is, by comparing the comprehensive matching degree with the preset threshold, it is judged whether the matching degree between the job seeker and the recruiter reaches the preset standard. If the comprehensive matching degree is higher than the preset threshold, it indicates that the job seeker has a high degree of fit with the recruiter in terms of work experience, skills, and career planning, and is a relatively suitable candidate. Thus, the job seeking information and job intention plan published by the first job seeker can be pushed to the first recruiter. On the contrary, if the comprehensive matching degree is lower than the preset threshold, it indicates that there are significant differences between the job seeker's needs and the recruiter's. In this case, further screening is required. In this way, a more comprehensive and efficient matching method is provided for talent recruitment, which helps to improve the accuracy and efficiency of talent recruitment.

[0056] Specifically, the comprehensive matching degree is a comprehensive evaluation index that reflects the matching degree between the current work experience and skills of the job seeker and the recruitment requirements, as well as the degree of fit between the career plan and the long-term development goals of the enterprise. When the comprehensive matching degree is higher than the preset threshold, it indicates that the job seeker is highly consistent with the recruiter's needs in multiple aspects and has the potential to become a suitable candidate. For example, if a job seeker has rich project management experience and excellent technical capabilities, and their personal career plan is in line with the enterprise's development vision, then such candidates are obviously more likely to meet the recruitment requirements. At this time, the system will push the job seeking information and job intention plan to the recruiter for further communication and evaluation.

[0057] On the contrary, if the comprehensive matching degree is lower than the preset threshold, it indicates that there are significant differences between the job seeker's needs and the recruiter's. This situation may occur when the job seeker has certain professional skills but fails to fully align with the job requirements in terms of work experience or career development direction. For example, a job seeker may be good at programming languages but lack the necessary industry experience or their future career plan does not conform to the enterprise's long-term strategy. In this case, directly pushing the job seeking information may lead to unsatisfactory matching results, so further screening is required to find more suitable candidates.

[0058] In this process, the setting of the preset threshold is crucial and directly affects the quality and efficiency of the matching recommendation. Different positions and industries may have different matching standards, so the threshold also needs to be flexibly adjusted to adapt to the specific situation. For some technology-intensive positions, such as software development and data analysis, the importance of work experience and skill matching may be higher; while for management or creative positions, more attention is paid to the degree of fit of the career plan. By reasonably setting the threshold, the system can ensure that the most suitable candidate can be found for each position, while avoiding unnecessary information overload and improving the recruitment efficiency.

[0059] In addition, by comparing the overall matching degree with a preset threshold, the system can filter out job seekers who clearly do not meet the requirements at an early stage, thereby reducing the time cost and energy consumption of the recruiter. This not only helps improve the overall efficiency of the recruitment process but also lays a solid foundation for subsequent interview and selection processes. When the job seekers pushed by the system have been strictly screened, meet the basic qualification requirements and have a high matching degree, the recruiter can focus more on in-depth understanding of the candidate's capabilities and potential rather than spending a lot of time screening resumes.

[0060] In summary, the talent big data platform system based on the embodiments of the present application is elucidated. It uses natural language processing technology based on deep learning to semantically analyze the job seeking information and job intention plans released by job seekers, as well as the recruitment requirements released by recruiters, in order to extract the semantic feature representations of job seeking information, job intention plans, and recruitment requirement information respectively. Furthermore, by semantically jointly analyzing the job seeking information and job intention plans of job seekers with the recruitment requirements respectively, the matching degree of job seekers with the recruiter's requirements in terms of work experience skills and career planning is revealed, thereby achieving a more comprehensive talent matching recommendation. In this way, the accuracy and efficiency of talent matching can be improved, meeting the enterprise's demand for talents and at the same time helping to improve the job success rate of job seekers.

[0061] The basic principles of the present invention have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, benefits, effects, etc. mentioned in the present invention are only examples and not limitations. It cannot be considered that these advantages, benefits, effects, etc. are essential for each embodiment of the present invention. In addition, the specific details of the above embodiments are only for illustrative purposes and for ease of understanding, rather than limitations. The above details do not limit the present invention to necessarily adopt the above specific details for implementation.

[0062] In the above embodiments, the descriptions of each embodiment have their own emphases. For parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiments described above are only illustrative. For example, the unit division is only a logical function division, and there can be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0063] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-described exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, in any respect, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced by the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

[0064] In addition, it is obvious that the term "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0065] Finally, it should be noted that the above description has been given for purposes of illustration and description. In addition, the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A talent big data platform system, characterized in that, Including: A job seeker information extraction module for extracting the job seeking information and job intention plan released by a first job seeker from the job seeking module; A recruitment requirement information extraction module for extracting the recruitment requirements released by a first recruiter from the recruitment module; A semantic encoding module for respectively performing semantic encoding on the job seeking information, job intention plan released by the first job seeker, and the recruitment requirements released by the first recruiter to obtain a job seeking information semantic encoding vector, a job intention plan semantic encoding vector, and a recruitment requirement semantic encoding vector; A job seeking information - recruitment requirement semantic joint analysis module for performing a job seeking information - recruitment requirement semantic joint analysis driven by bidirectional attention on the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job seeking information - recruitment requirement semantic joint interaction encoding vector; A job seeking plan - recruitment requirement semantic joint analysis module for performing a job seeking plan - recruitment requirement semantic joint analysis driven by bidirectional attention on the job intention plan semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job seeking plan - recruitment requirement semantic joint interaction encoding vector; An information push module for determining whether to push the job seeking information and job intention plan released by the first job seeker to the first recruiter based on the job seeking information - recruitment requirement semantic joint interaction encoding vector and the job seeking plan - recruitment requirement semantic joint interaction encoding vector.

2. The talent big data platform system according to claim 1, characterized in that The semantic encoding module is used for: Using a semantic encoder based on the BERT model to respectively perform semantic encoding on the job seeking information, the job intention plan, and the recruitment requirements to obtain the job seeking information semantic encoding vector, the job intention plan semantic encoding vector, and the recruitment requirement semantic encoding vector.

3. The talent big data platform system according to claim 2, wherein The job seeking information - recruitment requirement semantic joint analysis module includes: A bidirectional attention calculation unit for calculating the bidirectional attention scores between the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector to obtain a job seeking information - recruitment requirement semantic association forward attention score field and a job seeking information - recruitment requirement semantic association reverse attention score field; A bidirectional attention balance field construction unit for constructing a job seeking information - recruitment requirement semantic association positive - negative bidirectional attention balance field based on the job seeking information - recruitment requirement semantic association forward attention score field and the job seeking information - recruitment requirement semantic association reverse attention score field; A semantic mapping interaction analysis unit for performing semantic mapping interaction analysis on the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector based on the job seeking information - recruitment requirement semantic association positive - negative bidirectional attention balance field to obtain the job seeking information - recruitment requirement semantic joint interaction encoding vector.

4. The talent big data platform system according to claim 3, characterized in that The bidirectional attention calculation unit is used for: Performing a homography projection transformation on the job seeking information semantic encoding vector and the recruitment requirement semantic encoding vector respectively to obtain a job seeking information semantic homography projection encoding vector and a recruitment requirement semantic homography projection encoding vector; Calculate the forward attention score field of the semantic homography projection encoding vector of the job application information relative to the semantic homography projection encoding vector of the recruitment requirement to obtain the positive attention score field of the job application information - recruitment requirement semantic association; Calculate the reverse attention score field of the semantic homography projection encoding vector of the recruitment requirement relative to the semantic homography projection encoding vector of the job application information to obtain the reverse attention score field of the job application information - recruitment requirement semantic association.

5. The talent big data platform system according to claim 4, characterized in that The bidirectional attention balance field construction unit is used for: Input the positive attention score field of the job application information - recruitment requirement semantic association and the reverse attention score field of the job application information - recruitment requirement semantic association into the attention balance modulation module based on the convolutional neural network to obtain the positive and negative bidirectional attention balance field of the job application information - recruitment requirement semantic association.

6. The talent big data platform system according to claim 5, wherein, The semantic mapping interaction analysis unit is used for: Map the semantic homography projection encoding vector of the job application information and the semantic homography projection encoding vector of the recruitment requirement to the positive and negative bidirectional attention balance field of the job application information - recruitment requirement semantic association respectively to obtain the attention modulation encoding vector of the semantic homography projection of the job application information and the attention modulation encoding vector of the semantic homography projection of the recruitment requirement; Calculate the element-wise division between the attention modulation encoding vector of the semantic homography projection of the job application information and the attention modulation encoding vector of the semantic homography projection of the recruitment requirement to obtain the semantic joint interaction encoding vector of the job application information - recruitment requirement.

7. The talent big data platform system according to claim 6, characterized in that, The information push module includes: The comprehensive matching unit is used for calculating the comprehensive matching degree between the first job applicant and the first recruiter based on the semantic joint interaction encoding vector of the job application information - recruitment requirement and the semantic joint interaction encoding vector of the job search plan - recruitment requirement; The push unit is used for determining whether to push the job application information and job intention plan released by the first job applicant to the first recruiter based on the comparison between the comprehensive matching degree and a preset threshold.

8. The talent big data platform system according to claim 7, characterized in that, The comprehensive matching unit is used for: Input the semantic joint interaction encoding vector of the job application information - recruitment requirement and the semantic joint interaction encoding vector of the job search plan - recruitment requirement into the integrated decision module including the first classifier and the second classifier to obtain the first matching degree and the second matching degree; Calculate the mean of the first matching degree and the second matching degree as the comprehensive matching degree.

Images