AI-based recruitment service method and device, storage medium and program product

Abstract

The embodiment of the invention provides an AI-based recruitment service method and device, a storage medium and a program product. In the embodiment of the invention, an AI evaluation model is called, and an interview evaluation report is generated for an interview video of an AI interviewer for an interview of a target job seeker in an interview room. Furthermore, under the condition that the target recruitment terminal considers that the target job seeker is qualified based on the interview evaluation report or the comprehensive evaluation score meets a preset score, calling an AI calling model to simulate a calling person to perform double calling on the target job seeking terminal and the target recruitment terminal, and according to the dialogue content of the two parties after double calling, calling the target job seeker to perform double calling on the target recruitment terminal. And creating corresponding invitation schedule information for the target job seeker and the target recruiter under the condition of determining that the two parties meet the set cooperation intention condition, so that the target job seeking end and the target recruiter can invite each other according to the respective schedule information. Therefore, the interview efficiency is improved, meanwhile, the interview objectivity is guaranteed, the two parties can achieve the cooperation intention, and the recruitment efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of computer processing technologies, and particularly to an AI-based recruitment service method, device, storage medium, and program product. Background Art

[0002] Traditional recruitment methods usually rely on manual or semi-automatic resume screening and interview processes, which are time-consuming and labor-intensive for manual interviews, resulting in high recruitment costs and low recruitment efficiency. With the rapid development of science and technology, in order to improve recruitment efficiency and save labor costs, online interview rooms have been developed and AI interviewers are used to replace human interviewers to interview job seekers, and then the interview processes of each job seeker are sorted out and sent to the recruiter, so that the human resources department of the recruiter can select job seekers suitable for the recruitment position from numerous job seekers.

[0003] However, for the above-mentioned recruitment method, on the one hand, the final selection of job seekers also depends relatively on the subjective judgment of the human resources department, which may be biased. On the other hand, since there is sometimes no timely further communication and contact between the recruiter and the candidate job seekers, the success rate of reaching a cooperation intention between the recruiter and the job seekers is relatively low, affecting the recruitment efficiency. Summary of the Invention

[0004] Multiple aspects of this application provide an AI-based recruitment service method, device, storage medium, and program product, which can improve the objectivity of the interview while improving the interview efficiency, and effectively guide further communication and contact between the recruiter and the job seekers on the basis of the interview, which is conducive to both parties reaching a cooperation intention and improving the recruitment efficiency.

[0005] An embodiment of the present application provides a recruitment service method based on AI, including: during the process of an AI conversation model based on artificial intelligence simulating an interviewer to conduct an interview with a target job seeker in an interview room, recording an interview video, where the interview video includes audio information of the interview room and an interview picture, and the audio information contains the first conversation content between the target job seeker and the interviewer; in response to an interview end operation, invoking an AI evaluation model based on artificial intelligence, generating an interview evaluation report for the target job seeker based on the first conversation content and the interview picture, and sending the interview evaluation report to the target recruitment end for the target recruiter to view the interview evaluation report, where the interview evaluation report contains a comprehensive evaluation score; in response to a first interview qualified instruction sent by the target recruitment end for the target job seeker or when the comprehensive evaluation score meets a preset score, invoking an AI call model based on artificial intelligence to simulate a caller to perform a dual call on the target job seeker end and the target recruitment end to establish an audio-video channel between the target job seeker end and the target recruitment end; collecting the second conversation content of the communication between the target job seeker end and the target recruitment end through the audio-video channel, and determining the cooperation intention information between the target job seeker and the target recruiter according to the second conversation content; when the cooperation intention information meets the set cooperation intention conditions, creating corresponding appointment schedule information for the target job seeker and the target recruiter, and sending the appointment schedule information to the target job seeker end and the target recruitment end respectively, so that the target job seeker end and the target recruitment end can invite each other according to their respective schedule information.

[0006] An embodiment of the present application further provides an electronic device, including: a memory and a processor; the memory is used to store a computer program; the processor is coupled to the memory and is used to execute the computer program to implement the steps in the above methods.

[0007] An embodiment of the present application further provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the processor is caused to implement the steps in the above methods.

[0008] An embodiment of the present application further provides a computer program product, where the computer program product includes computer programs / instructions, and when the computer programs / instructions are executed by a processor, the processor is caused to be able to implement the steps in the above methods.

[0009] In the embodiment of the present application, an interview video of an AI interviewer interviewing a target job seeker in an interview room is recorded, an AI evaluation model is called to generate an interview evaluation report, and the interview evaluation report is sent to the target recruitment side for the target recruiter to view the interview evaluation report. While improving the interview efficiency, the objectivity of the interview is ensured, and objective and accurate interview reference data can be provided for the target recruiter to conduct recruitment. Further, in response to the first interview pass instruction sent by the target recruitment side for the target job seeker or when the comprehensive evaluation score meets the preset score, an AI call model is called to simulate a call operator to conduct a dual call on the target job seeking side and the target recruitment side to establish an audio-video channel between the target job seeking side and the target recruitment side, and based on the conversation content through this audio-video channel, when it is determined that the target job seeker and the target recruiter meet the set cooperation intention conditions, corresponding appointment schedule information is created for the target job seeker and the target recruiter, so that the target job seeking side and the target recruitment side can invite each other according to their respective schedule information. Thus, on the basis of the interview, it is possible to effectively guide the recruiter and the job seeker to have a deeper communication and contact, which is conducive to both parties reaching a cooperation intention and improving the recruitment efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The illustrative embodiments and descriptions thereof of the present application are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0011] Figure 1 is a flowchart of a recruitment service method based on AI provided by an exemplary embodiment of the present application;

[0012] Figure 2a is a schematic diagram of a page provided by an exemplary embodiment of the present application;

[0013] Figure 2b is a schematic diagram of another page provided by another exemplary embodiment of the present application;

[0014] Figure 2c is a schematic diagram of yet another page provided by yet another exemplary embodiment of the present application;

[0015] Figure 2d is a schematic diagram of yet another page provided by yet another exemplary embodiment of the present application;

[0016] Figure 2e is a schematic diagram of yet another page provided by yet another exemplary embodiment of the present application;

[0017] Figure 3 is a schematic diagram of the structure of an electronic device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments of this application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of this application.

[0019] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or reject.

[0020] Continuing with the background art, in response to the technical problems mentioned in the background art that the final selection of job seekers also depends relatively on the subjective judgment of the human resources department, which may be biased, and due to the fact that sometimes there is no timely deeper communication and contact between recruiters and candidate job seekers, resulting in a relatively low success rate for recruiters and job seekers to reach a cooperation intention and affecting the recruitment efficiency. In the embodiments of this application, an interview video of the AI interviewer interviewing a target job seeker in the interview room is recorded, an AI evaluation model is called to generate an interview evaluation report, and the interview evaluation report is sent to the target recruitment end for the target recruiter to view the interview evaluation report. While improving the interview efficiency, it ensures the objectivity of the interview and can provide objective and accurate interview reference data for the target recruiter to conduct recruitment. Further, in response to the first interview qualified instruction sent by the target recruitment end for the target job seeker or when the comprehensive evaluation score meets the preset score, an AI call model is called to simulate a call operator to make a dual call to the target job seeker end and the target recruitment end to establish an audio and video channel between the target job seeker end and the target recruitment end. And based on the conversation content through this audio and video channel, when it is determined that the target job seeker and the target recruiter meet the set cooperation intention conditions, corresponding appointment schedule information is created for the target job seeker and the target recruiter, so that the target job seeker end and the target recruitment end can invite each other according to their respective schedule information. Thus, it is possible to effectively guide deeper communication and contact between recruiters and job seekers on the basis of the interview, which is conducive to both parties reaching a cooperation intention and improving the recruitment efficiency.

[0021] The following will detail a solution provided by the embodiments of this application in conjunction with the drawings.

[0022] Figure 1Schematic flowchart of the AI-based recruitment service method provided by an exemplary embodiment of the present application. As Figure 1 shown, the method includes:

[0023] 101. During the process of an AI dialogue model based on artificial intelligence simulating an interviewer to conduct an interview with a target job seeker in an interview room, record an interview video, where the interview video includes audio information and interview images of the interview room, and the audio information contains the first conversation content between the target job seeker and the interviewer;

[0024] 102. In response to the interview end operation, call an AI evaluation model based on artificial intelligence, generate an interview evaluation report for the target job seeker based on the first conversation content and the interview images, and send the interview evaluation report to the target recruitment end for the target recruiter to view the interview evaluation report, where the interview evaluation report contains a comprehensive evaluation score;

[0025] 103. In response to the first interview qualified instruction sent by the target recruitment end for the target job seeker or when the comprehensive evaluation score meets the preset score, call an AI call model based on artificial intelligence to simulate a call operator to conduct a two-way call on the target job seeker end and the target recruitment end to establish an audio-video channel between the target job seeker end and the target recruitment end;

[0026] 104. Collect the second conversation content of the communication between the target job seeker end and the target recruitment end through the audio-video channel, and determine the cooperation intention information between the target job seeker and the target recruiter according to the second conversation content;

[0027] 105. When the cooperation intention information meets the set cooperation intention conditions, create corresponding invitation schedule information for the target job seeker and the target recruiter, and send the invitation schedule information to the target job seeker end and the target recruitment end respectively, so that the target job seeker end and the target recruitment end can invite each other according to their respective schedule information.

[0028] In an interview scenario, usually each job seeker corresponds to a job seeker end, and each recruiter corresponds to a recruitment end. Any recruiter can log in to the recruitment platform through their corresponding recruitment end and post a recruitment post on the recruitment platform. For the convenience of description and distinction, this recruiter can be called the target recruiter, and the recruitment end corresponding to the target recruiter can be called the target recruitment end. Job seekers can log in to the recruitment platform through the job seeker end and browse the posted recruitment posts on the recruitment platform. For the convenience of description and distinction, this job seeker can be called the target job seeker, and the job seeker end corresponding to the target job seeker can be called the target job seeker end. When the target job seeker browses the posts posted on the recruitment platform and is interested in the recruitment positions, recruitment requirements and other information of a certain or certain posts, they can submit their resumes to the recruiter who posted the post to strive for the opportunity to interview for the corresponding recruitment position.

[0029] The embodiments of the present application do not limit the implementation forms of the recruitment side and the job-seeking side. For example, the recruitment side and the job-seeking side can be intelligent handheld devices, such as smartphones, tablets, laptops or desktop computers, etc.; for another example, the recruitment side and the job-seeking side can also be intelligent wearable devices, such as smart watches, smart bracelets, etc.; for yet another example, the recruitment side and the job-seeking side can also be various intelligent home appliances with display screens, such as smart TVs, smart large screens or smart robots, etc. In addition, various application programs can be installed on the recruitment side and the job-seeking side. The application programs can be independent running APPs, or mini-programs that rely on independent APPs to run, or web pages, and this is not limited. The application programs can be, for example, recruitment application programs or recruitment platforms.

[0030] After receiving the resume, the target recruiter can invite the target job seeker who submitted the resume to an interview. Usually, inviting the target job seeker to an interview and conducting the interview with the target job seeker after the invitation is successful are the responsibilities of the enterprise's human resources or the operator of an individual business, etc. In order to improve the interview efficiency, save manpower, and ensure the objectivity of the interview, in the embodiments of the present application, the recruitment platform can create an online interview room for the target job seeker. The interview room is used for the target recruiter to conduct an online interview with the target job seeker, and an AI dialogue model based on artificial intelligence is bound to the interview room, so as to call the AI dialogue model to simulate a customer service to invite the target job seeker to an interview, and when conducting an online interview in the interview room after the invitation is successful, call the AI dialogue model to simulate an interviewer to conduct an interview with the target job seeker.

[0031] Among them, the AI dialogue model can be a large language model (LLMs) that supports multi-modal, and has the ability of cross-modal and cross-language deep semantic understanding and generation. In the process of generating the dialogue content with the job seeker, natural language processing can be used to search in the information library and / or generate content based on semantic understanding and other at least one method to generate question information for asking the job seeker and reply information for replying to the job seeker. The large language model includes but is not limited to: models of the GLM (Generalized Linear Model) series, models of the Qwen series, etc.

[0032] It should be noted that the AI dialogue model for simulating customer service and the AI dialogue model for simulating interviewers can be the same model or different models. Each interview room can be opened only for one job seeker, or the same interview room can be repeatedly opened for different job seekers. Each AI dialogue model can be limited to use in one interview room, or can be repeatedly used in different interview rooms. When an AI dialogue model is repeatedly used in different interview rooms, the AI dialogue model can serve interview rooms with different interview times, or can serve different interview rooms with the same interview time. When the same AI dialogue model serves different interview rooms with the same interview time, the AI dialogue model can simultaneously process the conversation content with job seekers in different interview rooms.

[0033] In the embodiments of the present application, after the target job seeker enters the interview room, the AI dialogue model can simulate an interviewer to conduct an interview with the target job seeker in the interview room. In order to facilitate the evaluation of the target job seeker and facilitate the target recruiter to understand the interview details and decide whether to hire the target job seeker according to the interview details, an interview video can be recorded during the process of the AI dialogue model simulating an interviewer to conduct an interview with the target job seeker in the interview room. Among them, the interview video includes: the audio information of the interview room and the interview picture. The audio information contains the conversation content between the target job seeker and the interviewer, which can be called the first conversation content. The interview picture includes the target job seeker and the subtitle content. The subtitle content at least includes the conversation content of the interviewer. Further, the subtitle content can also include the conversation content of the target job seeker, so that the target recruiter can understand the specific conversation content of the interview only through the audio or subtitles.

[0034] In some alternative embodiments, recording an interview video during the process of the AI dialogue model simulating an interviewer to conduct an interview with the target job seeker in the interview room includes: in response to the trigger operation of the target job seeker for the interview entrance, creating an interview room and a corresponding streaming process for the target job seeking terminal corresponding to the target job seeker, and the interview room is bound with the AI dialogue model; simulating an interview interaction between the interviewer and the target job seeker in the interview room based on the AI dialogue model, and using the streaming process to receive the audio information and the interview picture sent by the target job seeking terminal; and generating the interview video according to the audio information and the interview picture when the audio information and the interview picture are received.

[0035] In practical applications, after the interview platform monitors that the target job seeker submits a resume to the target recruiter, it can call the AI dialogue model to simulate the customer service of the interview platform and send a video interview invitation to the target job seeker through the chat page. The sent video interview invitation can be a link, an interactive card, etc. For example Figure 2aAs shown, taking the video interview invitation as an interactive card as an example, the card has a card cover, and the card cover includes: an "Interview Immediately" control (i.e., the interview entrance) and a brief description of the interview requirements. The brief description of the interview requirements is such as "Please dress appropriately in a quiet environment, present a good image, and the interview content is only visible to the employer. The interview results will be out within 48 hours, and the employment process is very fast." Further, when the target job seeker agrees to the interview, the "Interview Immediately" control can be triggered to enter the interview room for an online interview. Correspondingly, in response to the triggering operation of the target job seeker on the "Interview Immediately" control, the interview platform creates an interview room for the target job seeker and binds an AI dialogue model to the interview room to provide an online interview environment for the target job seeker. It should be noted that the interview room can be created by the interview platform when receiving the resume submitted by the target job seeker, and an AI dialogue model is bound to the interview room. Or, the interview room can also be created in response to the operation of the target job seeker triggering the "Interview Immediately" control; correspondingly, if the AI dialogue model for simulating the customer service and the AI dialogue model for simulating the interviewer are the same model, the AI dialogue model can be assigned to the target job seeker first when receiving the resume submitted by the target job seeker, and in response to the triggering operation of the target job seeker on the "Interview Immediately" control, an interview room is created for the target job seeker, and the AI dialogue model is bound to the interview room; or, if the AI dialogue model for simulating the customer service and the AI dialogue model for simulating the interviewer are the same model, the interview room can be created when receiving the resume submitted by the target job seeker, and the AI dialogue model is assigned and bound to the interview room; if the AI dialogue model for simulating the customer service and the AI dialogue model for simulating the interviewer are not the same model, the AI dialogue model for simulating the customer service can be assigned to the target job seeker first when receiving the resume submitted by the target job seeker, and in response to the triggering operation of the target job seeker on the "Interview Immediately" control, an interview room is created for the target job seeker, and the AI dialogue model for simulating the interviewer is assigned and bound to the interview room.

[0036] In an optional embodiment, in order to ensure the objectivity of the interview results and provide objective and accurate interview reference data for the target recruiter to conduct recruitment, at the end of the interview in the interview room, in response to the interview end operation, an AI evaluation model based on artificial intelligence can be called to generate an interview evaluation report for the target job seeker based on the first conversation content and the interview video, and the interview evaluation report is sent to the target recruitment end for the target recruiter to view the interview evaluation report, and the interview evaluation report contains a comprehensive evaluation score.

[0037] Among them, the AI evaluation model is mainly used to comprehensively evaluate the target job seekers in the recorded interview videos. Through multimodal technology, combining natural language processing (NLP) and computer vision (CV) capabilities, it analyzes and evaluates the multi-dimensional performance information of the target job seekers, such as language expression, body movements, and emotional states, and generates a detailed evaluation report. Specifically, the AI evaluation model has multimodal analysis capabilities, deep semantic understanding capabilities, and dynamic evaluation and adaptive learning capabilities. Among them, in terms of multimodal analysis capabilities, the AI evaluation model can simultaneously process multiple modal data such as language, vision, and emotion. By analyzing the language expression, speech rate, pauses, body language, and facial expressions of the target job seekers, it can more comprehensively evaluate the communication ability and emotional stability of the target job seekers. In terms of deep semantic understanding, the AI evaluation model is based on large language model (LLMs) technology and has cross-language and cross-domain deep semantic understanding capabilities. It can accurately understand the response information of the target job seekers and evaluate its logic, professionalism, and relevance. In terms of dynamic evaluation and adaptive learning, the AI evaluation model can dynamically adjust the evaluation criteria according to different interview scenarios and job requirements. At the same time, through continuous learning and feedback mechanisms, it continuously optimizes the evaluation algorithm to meet the specific needs of different industries and enterprises. At the level of security and ethical considerations, during the evaluation process, the AI evaluation model pays attention to the security and ethical compliance of the content and avoids generating or spreading harmful, discriminatory, or inappropriate content.

[0038] In some alternative embodiments, there is an interview evaluation report template corresponding to the interview evaluation report. Before the AI dialogue model simulates the interviewer to conduct an interview with the target job seeker in the interview room, it is also possible to obtain the recruitment requirement information of the target recruiter and generate an interview evaluation report template based on the recruitment requirement information. The interview evaluation report template includes: multiple information items and evaluation items. This embodiment does not limit the multiple information items and evaluation items, which are specifically determined by the needs of the target recruiter. For example, the multiple information items include, but are not limited to, multiple basic information items of the target job seeker, such as at least one information item related to the recruitment requirements among age, gender, education level, whether full-time, whether there is relevant work experience, and entry time. The evaluation items include, but are not limited to, at least one evaluation item related to the recruitment requirements among work ability evaluation items, image evaluation items, Putonghua evaluation items, English oral evaluation items, and / or comprehensive evaluation items.

[0039] Optionally, based on the recruitment requirement information, an interview assessment report template is generated, including: obtaining the recruitment requirement information of the target recruiter; classifying the recruitment requirement information to obtain multiple pieces of recruitment requirement information; generating question information for the interview based on the multiple pieces of recruitment requirement information, where the question information includes at least one of the question information that is key - concerned, generally - concerned, and commonly - concerned. Different target recruiters and / or different recruitment positions have different degrees of attention to different question information. In other words, different target recruiters and / or different recruitment positions have different degrees of attention to different information items; and generating corresponding multiple information items and assessment items based on the multiple pieces of recruitment requirement information. Additionally, the relevant position of each information item can also be marked with the recruitment requirement information corresponding to this information item. The comprehensive assessment item is the comprehensive assessment result of the work ability assessment item, image assessment item, Putonghua assessment item, and / or oral English assessment item; correspondingly, the relevant position of each assessment item can also be marked with the recruitment requirement information of this assessment item. Further, the interview assessment report template also includes: the question information item of the interviewer and the response information item of the target job seeker. The question information item is filled with the question information of the interviewer during the interview process, and the response information item is used to fill in the response information of the target job seeker during the interview process after the interview ends. If there are many questions to be asked during the interview process, the question information that is key - concerned for this recruitment position can be selected from multiple pieces of question information as the question information in the interview assessment report template.

[0040] In other words, each information item and evaluation item included in the interview evaluation report template are adapted to the recruitment requirement information. The recruitment requirement information of different recruiters and / or different recruitment positions may be different, and the corresponding interview evaluation report templates may also be different. When the same recruiter recruits different positions, the corresponding interview evaluation report templates for different recruitment positions may also be different. For example, in the recruitment requirement information for drivers, education background is not concerned, so multiple information items in the corresponding interview evaluation report template may not include the education background information item. In the recruitment requirement information for drivers, appearance and oral English are not concerned, but work ability, Mandarin and comprehensive ability are concerned. Then the evaluation items in the interview evaluation report template for drivers include: work ability evaluation item, Mandarin evaluation item and comprehensive evaluation item. The comprehensive evaluation item for drivers is the comprehensive evaluation result of the work ability evaluation item and the Mandarin evaluation item. Another example, in the recruitment requirements for catering cleaners, information such as age, gender, education background, whether full-time, whether there is relevant work experience, and start date are concerned. Then multiple information items in the corresponding interview evaluation report template need to include information items such as age, gender, education background, whether full-time, whether there is relevant work experience, and start date. In the recruitment requirement information for catering cleaners, work ability, appearance, Mandarin, oral English and comprehensive ability are concerned. Then the evaluation items in the corresponding interview test template include: work ability evaluation item, appearance evaluation item, Mandarin evaluation item, oral English evaluation item and comprehensive evaluation item. The comprehensive evaluation item for catering cleaners is the comprehensive evaluation result of work ability, appearance, Mandarin and oral English.

[0041] Correspondingly, the AI evaluation model at least includes: work ability evaluation model, appearance evaluation model, Mandarin evaluation model, oral English evaluation model and / or comprehensive evaluation model. The types of each evaluation model included in the AI evaluation model are adapted to the evaluation information items included in the interview evaluation report template, that is, which evaluation items are included in the interview evaluation report model, and the AI evaluation model includes the evaluation models corresponding to the evaluation items included in the interview evaluation report. For each evaluation item, using the AI evaluation model adapted to each evaluation item can comprehensively improve the accuracy and efficiency of the evaluation.

[0042] To further comprehensively improve the accuracy and efficiency of the evaluation, a high degree of cooperation in evaluation can be achieved among multiple evaluation models included in the AI evaluation model. To achieve the purpose of high-degree cooperation in evaluation among multiple evaluation models, a combined evaluation model training process can be provided in this embodiment, as follows.

[0043] Each evaluation model included in the above AI evaluation model can be obtained by pre-training the corresponding initial evaluation model with the corresponding training sample set. Each sample set contains multiple training samples, and each training sample has an annotation result, where the annotation result refers to the benchmark evaluation result annotated for each training sample by manual or annotation model. Using each training sample set to train the corresponding initial evaluation model will obtain the target evaluation result corresponding to each training sample in each training sample set. Further, for any initial evaluation model, multiple initial loss functions corresponding to the initial evaluation model can be calculated according to the benchmark evaluation result and the target evaluation result of each training sample corresponding to the initial evaluation model; according to the multiple initial loss functions corresponding to the initial evaluation model, calculate the intermediate loss function corresponding to the initial evaluation model, and the intermediate loss function is the comprehensive loss function of the multiple initial loss functions of the initial evaluation model; in the case that the intermediate comprehensive loss function does not reach the set intermediate training termination condition, adjust the model parameters of the initial evaluation model until the intermediate comprehensive loss function reaches the set intermediate evaluation model training termination condition to obtain the intermediate evaluation model. When calculating the intermediate loss function according to the multiple initial loss functions of any initial evaluation model, the calculation methods include but are not limited to: weighted summation, calculating the average value, and calculating the square difference. In addition, when calculating the intermediate loss function of each initial evaluation model separately, the loss functions used by different initial evaluation models may be the same, or the loss functions used by different initial evaluation models may be different. Further, after obtaining the intermediate evaluation model corresponding to each initial evaluation model through the above training method, the loss function between multiple intermediate evaluation models can be calculated according to the intermediate loss function of each intermediate evaluation model, and in the case that the loss function between multiple intermediate evaluation models does not reach the set target training termination condition, adjust the model parameters of each intermediate evaluation model until the loss function between multiple intermediate evaluation models reaches the set model training termination condition to obtain multiple evaluation models. When calculating the loss function between multiple intermediate evaluation models according to the intermediate loss function of each intermediate evaluation model, the calculation methods include but are not limited to: weighted summation, calculating the average value, and calculating the square difference. Thus, for each initial evaluation model, on the basis of calculating the loss for each training sample in the corresponding training sample set to adjust the model parameters, a secondary loss calculation is performed on the loss function between each training sample to adjust the model parameters twice, improving the accuracy of the model parameters of each model, and thus improving the evaluation accuracy of each model. Further, on the basis of separately adjusting the model parameters of each initial evaluation model, the joint loss function between the models is calculated, and the model parameters of each model are jointly adjusted, which not only further improves the accuracy of each model parameter, but also improves the accuracy of the comprehensive evaluation.

[0044] Taking the AI evaluation models including the work ability evaluation model, the image evaluation model, the Mandarin evaluation model, the spoken English evaluation model, and the comprehensive evaluation model as examples, the joint training process of the above AI evaluation models will be described in detail.

[0045] First, obtain the initial work ability evaluation model, initial image evaluation model, initial Mandarin evaluation model, initial spoken English evaluation model, and initial comprehensive evaluation model, as well as the training sample sets of the initial work ability evaluation model, the training sample set of the initial image evaluation model, the training sample set of the initial Mandarin evaluation model, the training sample set of the initial spoken English evaluation model, and the training sample set of the initial comprehensive evaluation model. Further, obtain the training sample sets of the initial work ability evaluation model, the training sample set of the initial image evaluation model, the training sample set of the initial Mandarin evaluation model, the training sample set of the initial spoken English evaluation model, and the training sample set of the initial comprehensive evaluation model, and train the corresponding initial evaluation models respectively to obtain the target evaluation results corresponding to each training sample in each training sample set. Further, for any initial evaluation model, multiple initial loss functions corresponding to any initial evaluation model can be calculated according to the benchmark evaluation results and target evaluation results of each training sample corresponding to any initial evaluation model, so as to obtain multiple initial loss functions corresponding to the initial work ability evaluation model, the initial image evaluation model, the initial Mandarin evaluation model, the initial spoken English evaluation model, and the initial comprehensive evaluation model respectively.Further, based on multiple first initial loss functions corresponding to the initial work ability evaluation model, calculate the first intermediate state loss function corresponding to the initial work ability evaluation model. In the case where the first intermediate state loss function does not reach the set training termination condition for the first intermediate state evaluation model, adjust the model parameters of the initial work ability evaluation model until the first intermediate state loss function reaches the set training termination condition for the first intermediate state work ability evaluation model, so as to obtain the first intermediate state work ability evaluation model; based on multiple second initial loss functions corresponding to the initial image evaluation model, calculate the second intermediate state loss function corresponding to the initial image evaluation model. In the case where the second intermediate state loss function does not reach the set training termination condition for the second intermediate state image evaluation model, adjust the model parameters of the initial image evaluation model until the second intermediate state loss function reaches the set training termination condition for the second intermediate state image evaluation model, so as to obtain the second intermediate state image evaluation model; based on multiple third initial loss functions corresponding to the initial Putonghua evaluation model, calculate the third intermediate state loss function corresponding to the initial Putonghua evaluation model. In the case where the third intermediate state loss function does not reach the set training termination condition for the third intermediate state Putonghua evaluation model, adjust the model parameters of the initial Putonghua evaluation model until the third intermediate state loss function reaches the set training termination condition for the third intermediate state Putonghua evaluation model, so as to obtain the third intermediate state Putonghua evaluation model; based on multiple fourth initial loss functions corresponding to the initial spoken English evaluation model, calculate the fourth intermediate state loss function corresponding to the initial spoken English evaluation model. In the case where the fourth intermediate state loss function does not reach the set training termination condition for the fourth intermediate state spoken English evaluation model, adjust the model parameters of the initial spoken English evaluation model until the fourth intermediate state loss function reaches the set training termination condition for the fourth intermediate state spoken English evaluation model, so as to obtain the fourth intermediate state spoken English evaluation model; based on multiple fifth initial loss functions corresponding to the initial comprehensive evaluation model, calculate the fifth intermediate state loss function corresponding to the initial comprehensive evaluation model. In the case where the fifth intermediate state loss function does not reach the set training termination condition for the fifth intermediate state comprehensive evaluation model, adjust the model parameters of the initial comprehensive evaluation model until the fifth intermediate state loss function reaches the set training termination condition for the fifth intermediate state comprehensive evaluation model, so as to obtain the fifth intermediate state comprehensive evaluation model. Further, based on the first intermediate state loss function, the second intermediate state loss function, the third intermediate state loss function, the fourth intermediate state loss function, and the fifth intermediate state loss function, calculate the intermediate state comprehensive loss function between the five intermediate state evaluation models. In the case where the intermediate state comprehensive loss function does not reach the set training termination condition, adjust the model parameters of each intermediate state evaluation model until the intermediate state comprehensive loss function reaches the set training termination condition for each intermediate state evaluation model, so as to obtain the work ability evaluation model, the image evaluation model, the Putonghua evaluation model, the spoken English evaluation model, and the comprehensive evaluation model.

[0046] In some alternative embodiments, after obtaining multiple evaluation models included in the AI evaluation model through the above-mentioned joint training method, in response to the end-of-interview operation, the AI evaluation model is called, and based on the first conversation content and the interview video, an interview evaluation report for the target job applicant is generated, including: inputting the audio information, the interview evaluation report template, and the recruitment requirement information into the work ability evaluation model, extracting information adapted to multiple information items based on the multiple information items and the first conversation content, calculating the first matching degree between the information adapted to the multiple information items and the recruitment requirement information, and calculating the work ability evaluation score of the target job applicant based on the first matching degree; and / or inputting the interview video and the recruitment requirement information into the image evaluation model, extracting the appearance feature information of the target job applicant, calculating the second matching degree between the appearance feature information and the recruitment requirement information, and calculating the image evaluation score of the target job applicant based on the second matching degree; and / or inputting the audio information into the Mandarin evaluation model, extracting the Mandarin audio features, and calculating the Mandarin evaluation score of the target job applicant based on the Mandarin audio features; and / or inputting the audio information into the spoken English evaluation model, extracting the spoken English audio features, and calculating the spoken English evaluation score of the target job applicant based on the spoken English audio features; inputting the work ability evaluation score, the image evaluation score, the Mandarin evaluation score, and / or the spoken English evaluation score into the comprehensive evaluation model, calculating the comprehensive evaluation score of the target job applicant; filling each evaluation score and each piece of information into the interview evaluation report template respectively to obtain the interview evaluation report of the target job applicant.

[0047] In some alternative embodiments, when the work ability assessment items are included in the interview assessment report template, the work ability assessment model is called to calculate the work ability assessment score of the target job applicant. The work ability assessment model includes: an information extraction network layer, a matching degree calculation network layer, and a work ability assessment network layer; the audio information, the interview assessment report template, and the recruitment requirement information are input into the work ability assessment model, and based on multiple information items and the first conversation content, the information adapted to the multiple information items is extracted, and the first matching degree between the information adapted to the multiple information items and the recruitment requirement information is calculated, and the work ability assessment score of the target job applicant is calculated based on the first matching degree, including: inputting the audio information, the interview assessment report template, and the recruitment requirement information into the information extraction network layer, extracting multiple work ability information items included in the interview assessment report template, extracting the first conversation content included in the audio information, extracting the requirement information adapted to each work ability information item included in the recruitment requirement information, and extracting the information adapted to each work ability information item from the first conversation content based on the multiple work ability information items; inputting the information adapted to each work ability information item and the corresponding recruitment requirement information into the matching degree calculation network layer to determine the degree of fit between the information adapted to each work ability information item and the corresponding recruitment requirement information; calculating the first matching degree between the information adapted to any work ability information item and the corresponding recruitment requirement information according to each degree of fit; inputting each first matching degree into the work ability assessment network layer, and calculating the work ability assessment score of the target job applicant according to the first matching degree corresponding to the information adapted to each work ability information item. When calculating the work ability assessment score of the target job applicant according to the first matching degree corresponding to the information adapted to each work ability information item, methods such as weighted summation, averaging, and calculating the square difference can be used for calculation.

[0048] In some alternative embodiments, when an image evaluation item is included in the interview evaluation report template, an image evaluation model is called to calculate the image evaluation score of the target job applicant. The image evaluation model includes: a feature information extraction network layer, a matching degree calculation network layer, and an image evaluation network layer; the interview video and the recruitment requirement information are input into the image evaluation model to extract the appearance feature information of the target job applicant, calculate the second matching degree between the appearance feature information and the recruitment requirement information, and calculate the image evaluation score of the target job applicant based on the second matching degree, including: inputting the audio information, the interview video, the interview evaluation report template, and the recruitment requirement information into the information extraction network layer, extracting multiple image information items included in the interview evaluation report template, extracting the first image information related to the image from the first conversation content included in the audio information, extracting the second image information related to the image of the target job applicant from the interview video, extracting the requirement information adapted to each image information item included in the recruitment requirement information, and determining the information adapted to each image information item from the first image information and the second image information based on the multiple image information items. The first image information and the second image information may be duplicate information. The first image information and the second image information may be height, weight, facial features, etc.; inputting the information adapted to each image information item and the corresponding recruitment requirement information into the matching degree calculation network layer to determine the degree of fit between the information adapted to each image information item and the corresponding recruitment requirement information; calculating the second matching degree between the information adapted to any image information item and the corresponding recruitment requirement information according to each degree of fit; inputting each second matching degree into the image evaluation network layer, and calculating the image evaluation score of the target job applicant according to the second matching degree corresponding to the information adapted to each image information item. When calculating the image evaluation score of the target job applicant according to the second matching degree corresponding to the information adapted to each image information item, calculation methods such as weighted summation, mean value calculation, and square difference calculation can be used.

[0049] In some optional embodiments, when the interview evaluation report template includes a Mandarin evaluation item, a Mandarin evaluation model is called to calculate the Mandarin evaluation score of the target job applicant. The Mandarin evaluation model includes: a feature information extraction network layer, a matching degree calculation network layer, and a Mandarin evaluation network layer; the audio information is input into the Mandarin evaluation model, the Mandarin audio features are extracted, and the Mandarin evaluation score of the target job applicant is calculated based on the Mandarin audio features, including: inputting the audio information, the interview evaluation report template, and the recruitment requirement information into the information extraction network layer, extracting multiple Mandarin information items contained in the interview evaluation report template, extracting the Chinese dialogue information of the target job applicant from the first dialogue content contained in the audio information, and extracting the Mandarin information items that are matched with each Mandarin information item contained in the recruitment requirement information. , and based on multiple Mandarin information items, determine the information that matches each Mandarin information item from the Chinese dialogue information of the target job applicant; input the information that matches each Mandarin information item and the corresponding recruitment requirement information into the matching degree calculation network layer to determine the degree of fit between the information that matches each Mandarin information item and the corresponding recruitment requirement information; calculate the third matching degree between the information that matches any Mandarin information item and the corresponding recruitment requirement information according to each degree of fit; calculate the Mandarin assessment score of the target job applicant according to the third matching degree corresponding to the information that matches each Mandarin information item. When calculating the Mandarin assessment score of the target job applicant according to the third matching degree corresponding to the information that matches each Mandarin information item, weighted summation, mean, square difference, etc. can be used for calculation.

[0050] In some alternative embodiments, when an oral English assessment item is included in the interview assessment report template, an oral English assessment model is invoked to calculate the Mandarin assessment score of the target job applicant. The audio information is input into the oral English assessment model to extract English oral audio features, and the oral English assessment score of the target job applicant is calculated based on the English oral audio features, including: inputting the audio information, the interview assessment report template, and the recruitment requirement information into the information extraction network layer, extracting multiple oral English information items included in the interview assessment report template, extracting the English conversation information of the target job applicant from the first conversation content included in the audio information, extracting the requirement information adapted to each oral English information item included in the recruitment requirement information, and determining the information adapted to each oral English information item from the English conversation information of the target job applicant based on the multiple oral English information items; inputting the information adapted to each oral English information item and the corresponding recruitment requirement information into the matching degree calculation network layer to determine the degree of fit between the information adapted to each oral English information item and the corresponding recruitment requirement information; calculating the fourth matching degree between the information adapted to any oral English information item and the corresponding recruitment requirement information according to each degree of fit; and calculating the oral English assessment score of the target job applicant according to the fourth matching degrees corresponding to the information adapted to each oral English information item. When calculating the oral English assessment score of the target job applicant according to the fourth matching degrees corresponding to the information adapted to each oral English information item, methods such as weighted summation, averaging, and calculating the squared difference can be used for calculation.

[0051] In some alternative embodiments, when a comprehensive assessment item is included in the interview assessment report template, a comprehensive assessment model is invoked to calculate the comprehensive assessment score of the target job applicant. The comprehensive assessment model includes: a comprehensive assessment network layer; inputting the work ability assessment score, the image assessment score, the Mandarin assessment score, and / or the oral English assessment score into the comprehensive assessment model to calculate the comprehensive assessment score of the target job applicant, including: inputting the work ability assessment score, the image assessment score, the Mandarin assessment score, and / or the oral English assessment score into the comprehensive assessment network layer to perform a comprehensive calculation on the work ability assessment score, the image assessment score, the Mandarin assessment score, and / or the oral English assessment score to obtain the comprehensive assessment score of the target job applicant. Among them, the comprehensive assessment calculation includes but is not limited to: weighted summation, averaging, and calculating the squared difference.

[0052] Further, after obtaining the information corresponding to each information item included in the interview assessment report and the assessment scores corresponding to each assessment item, the assessment scores and each item of information can be filled into the interview assessment report template respectively to obtain the interview assessment report of the target job applicant, so as to facilitate the target recruiter to view the interview assessment report and determine whether the target job applicant meets the requirements based on the objective and accurate assessment results given in the interview assessment report. Whether the target job applicant meets the requirements indicates whether the target recruiter intends to communicate further with the target job applicant.

[0053] Further optionally, the interview assessment report template further includes: a display area for the interview video; filling each assessment score and each piece of information into the interview assessment report template respectively to obtain the interview assessment report of the target job seeker, and further including: filling the interview video into the interview video display area to obtain the interview assessment report of the target job seeker, so as to facilitate the target recruiter to view the interview video and determine whether the target job seeker meets the requirements based on the objective and accurate assessment results given by the interview video and the interview assessment report. Whether it meets the requirements indicates whether the target recruiter is interested in further in-depth communication with the target job seeker.

[0054] Figure 2bShown is an exemplary diagram of an interview assessment report for a catering cleaner, which does not limit the style and layout of the interview assessment report of this application. At the top of the interview assessment report in this diagram is the display area for the interview video, where the interview video during the interview process is shown. Below the display area for the interview video is the assessment result display area, which shows the interview results of the target job applicant. The assessment results include: the name of the target job applicant, the applied position (cleaner), the comprehensive assessment score (80 points), the matching degree with the recruitment position (high matching degree), whether it is suitable for the recruitment position (suitable), and the employment-related information of the target job applicant. The employment-related information can be, for example, "This applicant is 35 years old. Although they have no catering cleaning experience, they have domestic cleaning experience. They can work full-time long-term and can start working as early as the end of March. It is recommended to contact by phone as soon as possible." Further, below the assessment result display area is the key attention information item display area, which shows multiple basic information items of the target job applicant. In each basic information item, the requirements of the target recruiter for this information item and the corresponding information of the target job applicant are shown. For example, the target recruiter's requirement for age is 16 - 45 years old, and the target job applicant's age is 35 years old; the target recruiter's requirement for gender is not limited to male or female, and the target job applicant's gender is male; the target recruiter's requirement for education level is not limited, and the target job applicant's education level is high school; the target recruiter's requirement for the position attribute is full-time, and the target job applicant can work full-time; the target recruiter's requirement for work experience is not limited, and the target job applicant actually has work experience; the target recruiter's requirement for the employment time is to start working at any time, and the target job applicant can also start working at any time, and so on. Further, below the key attention information item display area is the general attention information item display area, which shows other basic information items of the target job applicant. In each other basic information item, the requirements of the target recruiter for this information item and the corresponding information of the target job applicant are shown. For example, the target recruiter's requirement for height is 165 - 200 cm, and the target job applicant's height is 170 cm; the target recruiter's requirement for work experience is that those with cleaning work experience are preferred, and the target job applicant actually has no catering cleaning experience; the target recruiter's requirement for appearance and temperament is regular features, and the target job applicant has regular features. Further, below the general attention information item display area is the interview details information item display area, which shows interview details information items (which can be regarded as ordinary attention information items). Each interview details information item shows the conversation information during the interview process. For example, the interviewer asks, "Please give a brief self-introduction, including your age, height, and weight." The target job applicant answers, "I am 35 years old, 1.7 m tall, and weigh 65 kg."; Another example is that the interviewer asks, "Do you want to find a part-time or full-time job? Can you work long-term?" The target job applicant answers, "I want to find a long-term full-time job next month." and so on.

[0055] Continuing with the above optional embodiments, in another optional embodiment, in order to ensure the objectivity of the interview results and provide objective and accurate interview reference data for the target recruiter to conduct recruitment, an AI evaluation model can also be called in real time during the interview process to conduct an evaluation based on the real-time conversation content and interview video. As the interview progresses, the evaluation results will be continuously updated, and until the end of the interview, an interview evaluation report for the target job seeker will be generated.

[0056] Based on this, before interviewing the target job seeker, the interview question information can be classified by question type according to the recruitment requirement information of the target recruiter to obtain multiple question information corresponding to each question type. In other words, the interview question information is classified by question type according to each evaluation item in the interview evaluation report template corresponding to the target recruiter to obtain multiple question information corresponding to each question type. The multiple question information corresponding to the same question type can be centrally asked to the target job seeker during the interview to quickly and centrally obtain the response information made by the target job seeker to the question information of the same question type. Thus, during the interview, after obtaining the response information corresponding to the multiple question information of any question type, the corresponding AI evaluation model can be immediately triggered to generate the evaluation result of the evaluation item corresponding to this question type according to the conversation information corresponding to this question type. Then, as the interview ends, the evaluation results corresponding to each evaluation item to be evaluated will be obtained, and an on-the-spot evaluation report for the target job seeker can be generated according to each piece of information and the evaluation results.

[0057] Further optionally, the interview evaluation report includes: a qualified control, which can be referred to as the first interview qualified control. Figure 2c and 2d For two exemplary implementation manners of the first interview qualified control, Figure 2c in the first one, the first interview qualified control is fixedly displayed at the bottom of the screen in a sunken manner. As the interview evaluation report is scrolled up for display, the first interview qualified control does not scroll along. Figure 2dThe first interview qualified control is displayed above the interview assessment report in the form of a pop-up window. In the pop-up window, voice and text prompts can also be used to guide the target recruiter to make a choice on whether the target job seeker is suitable for the recruitment position. After the target recruiter views the interview assessment report, if the target job seeker deems the target job seeker qualified, the qualified control can be triggered, and an interview qualified instruction can be generated, which can be referred to as the first interview qualified instruction. Then, in response to the first interview qualified instruction sent by the target recruitment end for the target job seeker, an AI call model based on artificial intelligence is called to simulate a call operator to conduct a two-way call between the target job seeking end and the target recruitment end to establish a communication channel between the target job seeking end and the target recruitment end. By automatically triggering the two-way call process, the efficiency of subsequent communication is improved. Especially when the target recruiter deems the target job seeker suitable, it is more conducive to enhancing the cooperation intention between both parties. Additionally, the interview assessment report also includes: an unqualified control, which can be referred to as the first interview unqualified control. The first unqualified control can have the same display method as the first qualified control. After the target recruiter views the interview assessment report, if the target job seeker deems the target job seeker unqualified, the unqualified control can be triggered, and an interview unqualified instruction can be generated, which can be referred to as the first interview unqualified instruction. Then, in response to the first interview unqualified instruction sent by the target recruitment end for the target job seeker, the call of the AI call model to simulate a call operator to conduct a two-way call between the target job seeking end and the target recruitment end is abandoned.

[0058] In some alternative embodiments, in response to a first interview qualification instruction sent by a target recruitment side to a target job seeker, an AI call model based on artificial intelligence is called to simulate a call operator to conduct a dual call between the target job seeker side and the target recruitment side to establish a communication channel therebetween, including: receiving the first interview qualification instruction sent by the target recruitment side to the target job seeker, where the first interview qualification instruction is generated by a triggering operation of a target recruiter on a first interview qualification control; in the case where it is not monitored that the target recruitment side actively calls the target job seeker side, in response to the first interview qualification instruction, calling the AI call model to simulate a call operator to call the target recruitment side; in the case where the target recruitment side answers this call and indicates a willingness to talk to the target job seeker side, calling the intelligent call model to simulate a call operator to transfer this call to the target job seeker side to establish a communication channel between the target job seeker side and the target recruitment side. It should be noted that after the interview test report is sent to the target recruitment side, if the target recruitment side does not view the interview evaluation report within a certain time interval and does not trigger the first interview qualification control, the AI dual call model can judge whether the target job seeker is qualified according to the comprehensive evaluation score, and in the case of qualification, automatically trigger the dual call process. In this dual call, since it is not detected that the target recruiter views the interview test report and / or the target recruitment side actively calls the target job seeker side, there may be a situation where the target recruiter changes the intention to cooperate with the target job seeker. Therefore, when conducting the dual call, the target recruitment side can be called first to obtain the permission of the target recruitment side to agree to the dual call. In the case where the target job seeker side agrees to establish a connection with the target recruitment side, then the target job seeker side is called to establish a communication channel between the target job seeker side and the target recruitment side, thereby avoiding the probability of an invalid call to the target job seeker side due to the change of the target recruiter's cooperation intention. However, it is not limited to this. In this call, the target job seeker side can also be called first. In the case where the target job seeker side agrees to establish a connection, then the target recruitment side is called to establish a communication channel between the target job seeker side and the target recruitment side.

[0059] For example, during this dual call process, when calling the target recruitment side by calling the AI call model to simulate a call operator, after the target recruitment side answers this call, the call operator's words can be "Well, hello, boss. Boss, I'm from ** company. I have a candidate who sent you an interview video and you haven't watched it yet. Let me briefly introduce his situation to you. He is xx years old, {gender}, {height}. Do you think it's suitable? If it is, I'll help you contact him." When calling the target job seeker side by calling the AI call model to simulate a call operator in the case where the target recruitment side indicates a willingness to talk to the target job seeker side, the call words can be "Hey, hello. I received your interview video and think you're very good. I want to have a deeper chat. I'm from xx company, recruiting for the xx position. Do you have time now?"

[0060] Among them, the AI call model can also be another large language model with call functions and supporting multi-modal (optimization technology based on LLMs), which has the ability of cross-modal and cross-language deep semantic understanding and generation. In the process of generating the conversation content with job seekers, natural language processing can be adopted to search in the information database and / or generate content creation based on semantic understanding and other at least one method to generate question information for asking job seekers and reply information for replying to job seekers. The large language model includes but is not limited to: optimized models of the GLM series, optimized models of the Qwen series, etc.

[0061] Further optionally, the AI call model is also obtained by training the initial AI call model with a large number of training samples. In order to comprehensively improve the objectivity of the interview, the accuracy of the interview evaluation report, and the success rate of reaching a cooperation intention in the double call, when conducting model training, the initial AI dialogue model, multiple initial AI evaluation models, and the initial AI call model can be jointly trained to obtain the trained AI dialogue model, AI evaluation model, and AI call model, so as to use the trained AI dialogue model, AI evaluation model, and AI call model to perform the corresponding operations above. For the process of jointly training the initial AI dialogue model, multiple initial AI evaluation models, and the initial AI call model, reference can be made to the relevant description of the joint training of multiple evaluation models in the above embodiments, which will not be elaborated here.

[0062] Further optionally, before calling the AI call model based on artificial intelligence to simulate a call operator to perform a double call on the target job seeker side and the target recruiter side, in order to further improve the subsequent communication efficiency between the two parties, it can also be determined whether the current situation meets the double call conditions. For example, the double call conditions can include: whether the time interval between the time when the target recruiter sends the interview evaluation report and the current time is greater than the first time threshold. The first time threshold can be several minutes, several hours, etc., such as 24 hours. If the time is too long, it may affect the effectiveness of the interview. Another example is that the double call conditions can also include: whether the time when the target recruiter triggers the appropriate control exceeds the second time threshold. The purpose of this condition is to reserve time for the target recruiter or the target job seeker to take the initiative to contact the other party. If the target recruiter or the target job seeker takes the initiative to contact the other party, the process of calling the AI call model for a double call can be saved, thereby reducing the data processing volume and the double call cost. The second time threshold can be several minutes, several hours, etc.

[0063] Based on this, before calling the AI call model based on artificial intelligence to simulate a call operator to make a dual call to the target job seeker side and the target recruitment side, it is possible to determine whether the time interval between the time of sending the interview assessment report to the target recruitment side and the current time is greater than the first time threshold; if not, determine whether it is detected that the target recruitment side actively calls the target job seeker side; if not, then perform the operation of calling the AI call model based on artificial intelligence to simulate a call operator to make a dual call to the target job seeker side and the target recruitment side to establish a communication channel between the target job seeker side and the target recruitment side when the comprehensive assessment score meets the first score threshold. The first score threshold refers to the passing score value of the comprehensive assessment, and the passing score value can be 60 points, 70 points, 80 points, etc. The passing scores of different recruitment positions may be the same or different; if so, it means that the target recruiter and the target job seeker have spontaneously conducted a phone call, and the dual call process is abandoned.

[0064] Further optionally, the dual call corresponds to a callable time period and a non-callable time period. The callable time period refers to the time period during which a dual call can be made, and the non-callable time period refers to the time period during which a dual call cannot be made. Then the dual call condition may further include: whether the current time is within the callable time period. For example, from 10:00 to 11:00 am and from 2:00 to 5:00 pm from Monday to Friday every week are callable time periods, and other time periods are non-callable time periods. Based on this, calling the AI call model based on artificial intelligence to simulate a call operator to make a dual call to the target job seeker side and the target recruitment side includes: judging whether the current time point is within the callable time period according to the callable time period; if not, putting the contact information of the target recruitment side and the target job seeker side into the to-do pool. The to-do pool is associated with a re-call time point and an automatic dual call process. The re-call time point is a time point within the callable time period. For example, in the case where from 10:00 to 11:00 am and from 2:00 to 5:00 pm from Monday to Friday every week are callable time periods, the re-call time point can be 10:30 am and 3:00 pm of the next day; when reaching the re-call time point, calling the AI call model based on artificial intelligence to simulate a call operator to make a dual call based on the contact information of the target recruitment side and the target job seeker side to establish an audio-visual channel between the target job seeker side and the target recruitment side.

[0065] The above embodiments describe the dual call process based on the interview assessment report in the case where the interview assessment report of the target job seeker can be generated. The following embodiments will describe the dual call process based on the interview video in the case where the interview assessment report of the target job seeker cannot be generated.

[0066] In some alternative embodiments, in the case where the interview assessment report of the target job seeker cannot be generated, an interview pass control, which can be called the second interview pass control, can be added to the interview video, and the interview video with the second interview pass control added is sent to the target recruitment side corresponding to the target recruiter.

[0067] Correspondingly, after the target recruiter watches the interview video and deems the target job seeker qualified, the second interview qualified control can be triggered, and a second interview qualified instruction will be generated accordingly. Figure 2e This is an exemplary implementation of the second interview qualified control. The second interview video includes the second interview qualified control, which is displayed above the interview video. Further, in response to the second interview qualified instruction, it is determined whether the time interval between the time when the interview video is sent to the target recruitment end and the time when the interview qualified instruction is received is greater than the second time threshold. The second interview qualified instruction is generated by the target recruiter's triggering operation on the second interview qualified control. If not, after a specified duration from receiving the second interview qualified instruction, an AI call model based on artificial intelligence is called to simulate a call operator to conduct a dual call to the target job seeker end and the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end. By automatically triggering the dual call process, the efficiency of subsequent communication is improved. Especially when the target recruiter deems the target job seeker suitable, it is more conducive to increasing the likelihood of both parties reaching a cooperation intention. The settings of the second time threshold and the specified duration can refer to the settings of the first time threshold and the second time threshold respectively, which will not be elaborated here.

[0068] Optionally, calling an AI call model based on artificial intelligence to simulate a call operator to conduct a dual call to the target job seeker end and the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end includes: calling the AI call model to simulate a call operator to call the target job seeker end; when the target job seeker end answers this call and expresses the willingness to talk to the target recruitment end, calling the intelligent call model to simulate a call operator to transfer this call to the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end. During this dual call process, since it is monitored that the dual call is made when the target recruiter deems the target job seeker qualified, there is no need to first call the target recruitment end to obtain the permission to call the target recruitment end. The AI call model can first call the target job seeker end. When the target job seeker end agrees to establish a connection with the target recruitment end, then call the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end. Thus, the probability of making an ineffective call to the target recruitment end due to a failed call to the target job seeker end can be reduced. However, this is not limited to this. In this call, it is also possible to first call the target recruitment end. When the target recruitment end agrees to establish a connection, then call the target job seeker end to establish a communication channel between the target job seeker end and the target recruitment end.

[0069] It should be noted that in the case of generating an interview evaluation report for the target job seeker and not being able to generate an interview evaluation report for the target job seeker, the call order when using the AI dual call model for dual call can be the same or different, and this embodiment does not make a limitation on this.

[0070] For example, during this dual-call process, when the AI call model is invoked to simulate a caller calling the target job seeker, after the target job seeker answers this call, the caller's speech can be "Hey, hello. I've received your interview video and think you're really good. I'd like to have a deeper chat. I'm from xx company, recruiting for the xx position. Do you have time now?" When the target recruiter indicates willingness to talk to the target job seeker, when the AI call model is invoked to simulate a caller calling the target job seeker, the call speech can be "Hey, boss, I'm Xiao Zhang from 58. The candidate you thought was quite suitable after watching the video today, xx years old, {gender}, {height}, is waiting for your call. Shall I connect you?"

[0071] Further, in the case where the interview evaluation report of the target job seeker cannot be generated, an interview unqualified control, which can be called the second interview unqualified control, can be added to the interview video, and the interview video with the second interview unqualified control added is sent to the target recruitment end corresponding to the target recruiter. Correspondingly, after the target recruiter watches the interview video and deems the target job seeker unqualified, the second interview unqualified control can be triggered, and a second interview unqualified instruction will be generated accordingly, thereby abandoning the subsequent dual-call process.

[0072] The technical solutions provided by the embodiments of the present application record the interview video during the interview of the target job seeker by the AI interviewer in the interview room, invoke the AI evaluation model to generate an interview evaluation report, and send the interview evaluation report to the target recruitment end for the target recruiter to view the interview evaluation report. While improving the interview efficiency, it ensures the objectivity of the interview and can provide objective and accurate interview reference data for the target recruiter to conduct recruitment. Further, in response to the first interview qualified instruction sent by the target recruitment end for the target job seeker or when the comprehensive evaluation score meets the preset score, the AI call model is invoked to simulate a caller to conduct a dual call between the target job seeker end and the target recruitment end to establish an audio-video channel between the target job seeker end and the target recruitment end, and based on the conversation content through this audio-video channel, when it is determined that the target job seeker and the target recruiter meet the set cooperation intention conditions, corresponding appointment schedule information is created for the target job seeker and the target recruiter, so that the target job seeker end and the target recruitment end can invite each other according to their respective schedule information to reach cooperation. Thus, on the basis of the interview, it can effectively guide the recruiter and the job seeker to have a deeper communication and contact, which is conducive to both parties reaching a cooperation intention and improving the recruitment efficiency.

[0073] Figure 3 It is a schematic structural diagram of an electronic device provided for an exemplary embodiment of the present application. As Figure 3As shown in the figure, it includes: a memory 30a and a processor 30b; the memory 30a is used to store a computer program; the processor 30b is coupled to the memory 30a and is used to execute the computer program to implement the following steps:

[0074] During the process of an AI dialogue model based on artificial intelligence simulating an interviewer to conduct an interview with a target job seeker in an interview room, an interview video is recorded. The interview video includes audio information and an interview picture of the interview room. The audio information contains the first conversation content between the target job seeker and the interviewer. In response to an interview end operation, an AI evaluation model based on artificial intelligence is called. Based on the first conversation content and the interview picture, an interview evaluation report for the target job seeker is generated, and the interview evaluation report is sent to the target recruitment side for the target recruiter to view the interview evaluation report. The interview evaluation report contains a comprehensive evaluation score. In response to a first interview pass instruction sent by the target recruitment side for the target job seeker or when the comprehensive evaluation score meets a preset score, an AI call model based on artificial intelligence is called to simulate a caller to conduct a two-way call between the target job seeker side and the target recruitment side to establish an audio and video channel between the target job seeker side and the target recruitment side. The second conversation content of the communication between the target job seeker side and the target recruitment side through the audio and video channel is collected, and based on the second conversation content, the cooperation intention information between the target job seeker and the target recruiter is determined. When the cooperation intention information meets the set cooperation intention conditions, corresponding appointment schedule information is created for the target job seeker and the target recruiter, and the appointment schedule information is sent to the target job seeker side and the target recruitment side respectively, so that the target job seeker side and the target recruitment side can invite each other according to their respective schedule information.

[0075] In some alternative embodiments, during the process of an AI dialogue model based on artificial intelligence simulating an interviewer to conduct an interview with a target job seeker in an interview room, recording the interview video includes: in response to a trigger operation of the target job seeker for the interview entry, an interview room and a corresponding streaming process are created for the target job seeker side corresponding to the target job seeker. The interview room is bound with an AI dialogue model; based on the AI dialogue model, simulate an interview interaction between the interviewer and the target job seeker in the interview room, and use the streaming process to receive the audio information and the interview picture sent by the target job seeker side; in the case of receiving the audio information and the interview picture, generate an interview video according to the audio information and the interview picture.

[0076] Further optionally, before the AI conversation model based on artificial intelligence simulates an interviewer to conduct an interview for a target job seeker in an interview room, it is also possible to obtain the recruitment requirement information of the target recruiter; based on the recruitment requirement information, generate an interview evaluation report template, which includes multiple information items and evaluation items. The evaluation items include: work ability evaluation item, image evaluation item, Putonghua ability evaluation item, and / or comprehensive evaluation item. The comprehensive evaluation item is the comprehensive evaluation result of the work ability evaluation item, image evaluation item, and Putonghua ability evaluation item.

[0077] Correspondingly, the AI evaluation model at least includes: a work ability evaluation model, an image evaluation model, a Putonghua evaluation model, an English oral evaluation model, and / or a comprehensive evaluation model; in response to the interview end operation, call the AI evaluation model based on artificial intelligence, and based on the first conversation content and the interview video, generate an interview evaluation report for the target job seeker, including: inputting the audio information, the interview evaluation report template, and the recruitment requirement information into the work ability evaluation model, extracting the information adapted to the multiple information items based on the multiple information items and the first conversation content, calculating the first matching degree between the information adapted to the multiple information items and the recruitment requirement information, and calculating the work ability evaluation score of the target job seeker based on the first matching degree; and / or, inputting the interview video and the recruitment requirement information into the image evaluation model, extracting the appearance feature information of the target job seeker, calculating the second matching degree between the appearance feature information and the recruitment requirement information, and calculating the image evaluation score of the target job seeker based on the second matching degree; and / or, inputting the audio information into the Putonghua evaluation model, extracting the Putonghua audio features, and calculating the Putonghua evaluation score of the target job seeker based on the Putonghua audio features; and / or, inputting the audio information into the English oral evaluation model, extracting the English oral audio features, and calculating the English oral evaluation score of the target job seeker based on the English oral audio features; inputting the work ability evaluation score, the image evaluation score, the Putonghua evaluation score, and / or the English oral evaluation score into the comprehensive evaluation model, calculating the comprehensive evaluation score of the target job seeker; filling each evaluation score and each item of information into the interview evaluation report template respectively to obtain the interview evaluation report of the target job seeker.

[0078] Further optionally, the interview evaluation report template also includes: a display area for the interview video; filling each evaluation score and each item of information into the interview evaluation report template respectively to obtain the interview evaluation report of the target job seeker, and further includes: filling the interview video into the interview video display area to obtain the interview evaluation report of the target job seeker.

[0079] In some alternative embodiments, the interview assessment report includes: a first interview qualification control; in response to a first interview qualification instruction sent by a target recruitment end to a target job seeker, an AI call model based on artificial intelligence is called to simulate a call operator to conduct a two-way call between the target job seeker end and the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end, including: receiving the first interview qualification instruction sent by the target recruitment end to the target job seeker, where the first interview qualification instruction is generated by a triggering operation of the target recruiter on the first interview qualification control; in the case where it is not monitored that the target recruitment end actively calls the target job seeker end, in response to the qualification instruction, an AI call model based on artificial intelligence is called to simulate a call operator to call the target recruitment end; in the case where the target recruitment end answers this call and indicates a willingness to talk to the target job seeker end, the intelligent call model is called to simulate a call operator to transfer this call to the target job seeker end to establish a communication channel between the target job seeker end and the target recruitment end.

[0080] Further optionally, before calling an AI call model based on artificial intelligence to simulate a call operator to conduct a two-way call between the target job seeker end and the target recruitment end, it further includes: determining whether the time interval between the time when the interview assessment report is sent to the target recruitment end and the current time is greater than a first time threshold; if not, determining whether it is monitored that the target recruitment end actively calls the target job seeker end; if not, then in the case where the comprehensive assessment score meets the first score threshold, the operation of calling an AI call model based on artificial intelligence to simulate a call operator to conduct a two-way call between the target job seeker end and the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end is performed.

[0081] Further optionally, in the case where the interview assessment report of the target job seeker cannot be generated, it further includes: adding a second interview qualification control to the interview video and sending the interview video with the second interview qualification control added to the target recruitment end corresponding to the target recruiter; in response to the second interview qualification instruction, determining whether the time interval between the time when the interview video is sent to the target recruitment end and the time when the interview qualification instruction is received is greater than a second time threshold, where the second interview qualification instruction is generated by a triggering operation of the target recruiter on the second interview qualification control; if not, then after a specified duration of receiving the second interview qualification instruction, an AI call model based on artificial intelligence is called to simulate a call operator to conduct a two-way call between the target job seeker end and the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end.

[0082] In some alternative embodiments, calling an AI call model based on artificial intelligence to simulate a call operator to make a dual call to a target job-seeking terminal and a target recruitment terminal includes: judging whether the current time point is within the callable time period according to the callable time period; if not, putting the contact information of the target recruitment terminal and the target job-seeking terminal into a to-do pool, and the to-do pool is associated with a make-up call time point, and the make-up call time point is a time point within the callable time period; when the make-up call time point is reached, calling an AI call model based on artificial intelligence to simulate a call operator to make a dual call based on the contact information of the target recruitment terminal and the target job-seeking terminal, so as to establish an audio-video channel between the target job-seeking terminal and the target recruitment terminal.

[0083] Further, as Figure 3 shown, the server further includes: other components such as a communication component 30c, a display 30d, a power supply component 30e, and an audio component 30f. Figure 3 Only some components are schematically shown in Figure 3 the figure, which does not mean that the electronic device only includes

[0084] The detailed implementation manners and beneficial effects of the electronic device provided in the embodiments of the present application have been described in detail in the foregoing embodiments, and will not be elaborated herein.

[0085] An exemplary embodiment of the present application further provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the processor is caused to implement the steps in the foregoing method embodiments.

[0086] An exemplary embodiment of the present application further provides a computer program product, the computer program product includes a computer program / instructions, and when the computer program / instructions are executed by a processor, the processor is caused to be able to implement the steps in the foregoing method embodiments.

[0087] The foregoing memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read only memory (EEPROM), an erasable programmable read only memory (EPROM), a programmable read only memory (PROM), a read only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disc.

[0088] The above communication component is configured to facilitate communication between the device where the communication component is located and other devices in a wired or wireless manner. The device where the communication component is located can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G / LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wide Band (UWB) technology, Bluetooth (BT) technology and other technologies.

[0089] The above display includes a screen, and the screen can include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes and gestures on the touch panel. The touch sensors can sense not only the boundaries of touch or swipe actions, but also detect the duration and pressure associated with the touch or swipe operations.

[0090] The above power supply component provides power for various components of the device where the power supply component is located. The power supply component can include a power management system, one or more power supplies, and other components associated with generating, managing and distributing power for the device where the power supply component is located.

[0091] The above audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC), and when the device where the audio component is located is in an operation mode, such as a call mode, a recording mode and a voice recognition mode, the microphone is configured to receive external audio signals. The received audio signals can be further stored in a memory or sent via the communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals.

[0092] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM (Compact Disc Read-Only Memory), optical memory, etc.) that contain computer-usable program code.

[0093] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or the combination of blocks.

[0094] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or the combination of blocks.

[0095] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or the combination of blocks.

[0096] In a typical configuration, a computing device includes one or more processors (Central Processing Unit, CPU), an input / output interface, a network interface, and a memory.

[0097] The memory may include non-permanent memory in the form of computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory is an example of computer-readable media.

[0098] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0099] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or apparatus comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or apparatus. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising the element.

[0100] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. An AI-based recruitment service method, characterized in that: include: The AI ​​dialogue model based on artificial intelligence simulates an interviewer interviewing a target job applicant in an interview room, and records an interview video, wherein the interview video includes audio information and an interview screen of the interview room, and the audio information includes a first dialogue content between the target job applicant and the interviewer; In response to the interview end operation, calling an AI evaluation model based on artificial intelligence, generating an interview evaluation report for the target job seeker based on the first conversation content and the interview screen, and sending the interview evaluation report to the target recruitment end for the target recruiter to view the interview evaluation report, wherein the interview evaluation report includes a comprehensive evaluation score; In response to the first interview qualified instruction sent by the target recruitment terminal to the target job seeker or when the comprehensive assessment score meets the preset score, calling the AI ​​call model based on artificial intelligence to simulate the caller to make a double call to the target job seeker terminal and the target recruitment terminal, so as to establish an audio and video channel between the target job seeker terminal and the target recruitment terminal; Collecting a second conversation content between the target job seeker and the target recruiter through the audio and video channel, and determining cooperation intention information between the target job seeker and the target recruiter based on the second conversation content; When the cooperation intention information satisfies the set cooperation intention conditions, corresponding invitation schedule information is created for the target job seeker and the target recruiter, and the invitation schedule information is sent to the target job seeker and the target recruiter respectively, so that the target job seeker and the target recruiter can invite each other according to their respective schedule information.

2. The method according to claim 1, characterized in that The AI ​​dialogue model based on artificial intelligence simulates the interviewer's interview with the target job applicant in the interview room and records the interview video, including: In response to a trigger operation of a target job seeker on an interview entrance, an interview room and a corresponding streaming process are created for a target job seeker terminal corresponding to the target job seeker, wherein the interview room is bound to the AI ​​dialogue model; Based on the AI ​​dialogue model, the interviewer is simulated to interact with the target job seeker in the interview room, and the audio information and the interview screen sent by the target job seeker are received by using the streaming process; When the audio information and the interview screen are received, the interview video is generated according to the audio information and the interview screen.

3. The method according to claim 1, characterized in that Before the AI ​​dialogue model based on artificial intelligence simulates the interviewer to interview the target job applicant in the interview room, the method further includes: Obtaining recruitment requirement information of the target recruiter; Based on the recruitment requirement information, an interview evaluation report template is generated, wherein the interview evaluation report template includes multiple information items and evaluation items, wherein the evaluation items include: a work ability evaluation item, an image evaluation item, a Mandarin ability evaluation item, and / or a comprehensive evaluation item, wherein the comprehensive evaluation item is a comprehensive evaluation result of the work ability evaluation item, the image evaluation item, and the Mandarin ability evaluation item; Accordingly, the AI ​​evaluation model at least includes: a work ability evaluation model, an image evaluation model, a Mandarin evaluation model, an English oral evaluation model and / or a comprehensive evaluation model; in response to the interview end operation, the AI ​​evaluation model based on artificial intelligence is called to generate an interview evaluation report for the target job applicant based on the first dialogue content and the interview screen, including: Inputting the audio information, the interview evaluation report template and the recruitment requirement information into a work ability evaluation model, extracting information adapted to the multiple information items based on the multiple information items and the first conversation content, calculating a first matching degree between the information adapted to the multiple information items and the recruitment requirement information, and calculating a work ability evaluation score of the target job seeker based on the first matching degree; and / or Inputting the interview screen and the recruitment requirement information into the image assessment model, extracting the appearance feature information of the target job applicant, calculating a second matching degree between the appearance feature information and the recruitment requirement information, and calculating the image assessment score of the target job applicant based on the second matching degree; and / or Inputting the audio information into the Mandarin assessment model, extracting Mandarin audio features, and calculating the Mandarin assessment score of the target job applicant based on the Mandarin audio features; and / or Inputting the audio information into the spoken English assessment model, extracting spoken English audio features, and calculating the spoken English assessment score of the target job applicant based on the spoken English audio features; Inputting the work ability assessment score, the image assessment score, the Mandarin assessment score and / or the spoken English assessment score into the comprehensive assessment model to calculate the comprehensive assessment score of the target job seeker; Fill each assessment score and each information into the interview assessment report template to obtain the interview assessment report of the target job applicant.

4. The method according to claim 3, characterized in that The interview evaluation report template also includes: a display area for the interview video; filling each evaluation score and each information into the interview evaluation report template to obtain the interview evaluation report of the target job seeker, and also includes: The interview video is filled into the interview video display area to obtain an interview evaluation report of the target job applicant.

5. The method according to claim 1, characterized in that The interview evaluation report includes: a first interview qualified control; in response to the first interview qualified instruction sent by the target recruitment end to the target job seeker, calling an AI call model based on artificial intelligence to simulate a caller to perform a double call on the target job seeker end and the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end, including: Receiving a first interview pass instruction sent by the target recruitment end to the target job seeker, where the first interview pass instruction is generated by the triggering operation of the target recruiter on the first interview pass control; In the case where the target recruitment end is not detected to actively call the target job-seeking end, in response to the qualified instruction, calling an AI call model based on artificial intelligence to simulate a caller calling the target recruitment end; When the target recruiting end answers the call and expresses willingness to talk with the target job-seeking end, the AI ​​calling model S is called to simulate the calling person to transfer the call to the target job-seeking end, so as to establish a communication channel between the target job-seeking end and the target recruiting end.

6. The method according to claim 1, characterized in that Before calling the AI ​​call model based on artificial intelligence to simulate the caller to make a double call to the target job seeker terminal and the target recruitment terminal, it also includes: Determine whether the time interval between the time of sending the interview evaluation report to the target recruitment end and the current time is greater than a first time threshold; If not, determining whether the target recruitment terminal actively calls the target job-seeking terminal; If not, when the comprehensive assessment score meets the first score threshold, an AI call model based on artificial intelligence is called to simulate the caller making a double call to the target job seeker and the target recruiter to establish a communication channel between the target job seeker and the target recruiter.

7. The method according to claim 1, characterized in that In the event that the interview assessment report of the target job applicant cannot be generated, it also includes: Adding a second qualified interview control to the interview video, and sending the interview video with the second qualified interview control added to a target recruitment terminal corresponding to the target recruiter; In response to a second interview pass instruction, determining whether a time interval between a time when the interview video is sent to the target recruiter and a time when the interview pass instruction is received is greater than a second time threshold, wherein the second interview pass instruction is generated by a trigger operation of the target recruiter on the second interview pass control; If not, after receiving the specified time of the second interview qualified instruction, the AI ​​call model based on artificial intelligence is called to simulate the caller to make a double call to the target job seeker end and the target recruitment end to establish a communication channel between the target job seeker end and the target recruitment end.

8. The method according to any one of claims 1 to 7, characterized in that: Calling an AI call model based on artificial intelligence to simulate a caller to make a double call to the target job seeker and the target recruiter, including: According to the callable time period, determining whether the current time point is within the callable time period; If not, the contact information of the target recruitment terminal and the target job-seeking terminal is put into an agent pool, and the agent pool is associated with a supplementary call time point, and the supplementary call time point is a time point within the available call time period; When the supplementary call time point is reached, an AI call model based on artificial intelligence is called to simulate the caller to make a double call based on the contact information of the target recruitment end and the target job seeker end, so as to establish an audio and video channel between the target job seeker end and the target recruitment end.

9. An electronic device, characterized in that: include: Memory and processor; The memory is used to store a computer program; the processor is coupled to the memory and is used to execute the computer program to implement the steps in the method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor is caused to implement the steps in the method according to any one of claims 1 to 8.

11. A computer program product, characterized in that The computer program product comprises a computer program / instruction, and when the computer program / instruction is executed by a processor, the processor is enabled to implement the steps in any one of the methods of claims 1-8.

Images