A knowledge guide-based interview interaction data processing method and system

Abstract

The application provides a kind of interview interaction data processing method and system based on knowledge guidance, the method comprises: obtaining current context content in each round of conversation of current interview interaction process, based on current context content and action mapping strategy, obtain first prompt word by splicing processing;First prompt word is input to language model, at least one target action strategy used in the next round of conversation of output is obtained;Based on the language model, the interview guide information obtained and at least one target action strategy are processed, and the interview interaction question of the next round of conversation of language model output is obtained;In the case where it is determined that the current interview interaction process is completed based on the language model, the entire conversation interaction content corresponding to the current interview interaction process is obtained, and the entire conversation interaction content is analyzed to obtain the conversation interaction content analysis result.The method provided by the application can effectively improve the efficiency of interview interaction data processing, thereby greatly reducing the labor cost.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a knowledge-guided interview interaction data processing method and system. It also relates to an electronic device and a non-transitory computer-readable storage medium. Background Technology

[0002] In recent years, with the continuous development of artificial intelligence technology, various smart devices have been increasingly widely used in everyday interview scenarios. Interviews are a widely used method with a profound impact on qualitative research. The core concept of structured interviews is to ensure that the exact same questions are used in each interview and asked in the same order. This standardization ensures that answers can be reliably summarized and that sufficient comparisons can be made between different subgroups of the sample or between different survey periods. Building on structured interviews, semi-structured interviews, by breaking the constraints of fixed question sets and predefined order, ask probing questions about details that arise during the interview process, thereby revealing deeper knowledge and more profound connections while maintaining sample comparability similar to structured interviews. However, conducting semi-structured interviews requires extensive participation from experienced users, which severely limits the efficiency of data collection and processing. To automate this process, existing technologies employ task-oriented dialogue systems (TOD), which are designed to respond to user input within a predefined action space. By parsing natural language expressions into specific ontologies, the dialogue system then tracks the state and selects an action to generate a response that fulfills the intended function. However, semi-structured interviews are typically controlled by the interviewer and conducted in a question-and-answer format, aiming to extract information from the interviewee. However, dialogue-based interviews are very time-consuming and require considerable experience from the interviewer, which significantly limits the efficiency and feasibility of data collection and processing. Therefore, designing a more efficient knowledge-guided interactive data processing scheme for semi-structured interviews has become an urgent technical problem to be solved. Summary of the Invention

[0003] This invention provides a knowledge-guided interview interaction data processing method to address the shortcomings of existing interview interaction data processing schemes, which have high limitations and result in poor actual efficiency.

[0004] This invention provides a knowledge-guided interview interaction data processing method, comprising:

[0005] In each round of dialogue during the current interview interaction, the current context content is obtained, and the first prompt word is obtained by splicing the current context content and the preset action mapping strategy.

[0006] A preset language model is invoked, and the first prompt word is input into the language model to obtain at least one target action strategy to be used in the next round of dialogue output by the language model; the obtained interview guide information and the at least one target action strategy are processed based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model.

[0007] If the current interview interaction process is determined to be completed based on the language model, obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process;

[0008] The total dialogue interaction content includes at least one round of dialogue generated interview interaction questions and the historical interaction content input by the interviewee;

[0009] The language model is invoked to analyze all the dialogue interaction content to obtain the dialogue interaction content analysis results.

[0010] According to the knowledge-guided interview interaction data processing method of the present invention, before obtaining the current context content in each round of dialogue during the current interview interaction process, the method further includes:

[0011] The process involves obtaining the target question input by the interviewee, concatenating the target question with a preset question mapping strategy to obtain a second prompt word, and then calling a preset language model by inputting the second prompt word to obtain at least one inquiry question output by the language model that is associated with the target question. The question mapping strategy includes multiple question mapping relationships corresponding to each target question, and the inquiry question is a branch question or extension question based on the target question.

[0012] Obtain historical dialogue data containing the target question and at least one probing question; obtain corresponding interview guide information based on the historical dialogue data containing the target question and at least one probing question; and store the interview guide information in a database.

[0013] According to the knowledge-guided interview interaction data processing method of the present invention, the step of concatenating the target question and a preset question mapping strategy to obtain a second prompt word; invoking a preset language model, inputting the second prompt word into the language model, and obtaining at least one inquiry question output by the language model that is associated with the target question, specifically includes:

[0014] Based on the target question and the first question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a third prompt word; a preset language model is invoked, and the third prompt word is input into the language model to obtain a first analysis and processing result;

[0015] Based on the target question and the second question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a fourth prompt word; a preset language model is invoked, and the fourth prompt word is input into the language model to obtain a second analysis and processing result;

[0016] The fifth prompt word is obtained by concatenating the target question and the third question mapping relationship in the question mapping strategy; the fifth prompt word is input into the preset language model to obtain at least one inquiry question associated with the target question; wherein, the second prompt word includes the third prompt word, the fourth prompt word and the fifth prompt word; the question mapping relationship includes the first question mapping relationship, the second question mapping relationship and the third question mapping relationship.

[0017] According to the knowledge-guided interview interaction data processing method of the present invention, before processing the acquired interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model, the method further includes: obtaining interview guide information corresponding to the current interview interaction process from the database.

[0018] According to the knowledge-guided interview interaction data processing method of the present invention, the step of processing the acquired interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model includes:

[0019] The language model is invoked to combine the at least one target action strategy, and corresponding inquiry questions are generated based on the interview guide information to obtain the interview interaction questions for the next round of dialogue output by the language model; wherein, the at least one target action strategy includes at least one of questioning action strategy, response action strategy, inquiry action strategy and advancement action strategy.

[0020] According to the knowledge-guided interview interaction data processing method of the present invention, the step of obtaining all dialogue interaction content output by the language model corresponding to the current interview interaction process when the completion of the current interview interaction process is determined based on the language model specifically includes:

[0021] The actual number of dialogue rounds in the current interview interaction process is detected. If the actual number of dialogue rounds reaches a preset dialogue round threshold, the current interview interaction process is determined to be completed.

[0022] After determining that the current interview interaction process has been completed, the dialogue interaction content of the current interview interaction process is summarized based on the language model to obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process.

[0023] This invention also provides a knowledge-guided interview interaction data processing system, comprising:

[0024] The prompt word acquisition module is used to obtain the current context content in each round of dialogue during the current interview interaction process, and to perform splicing processing based on the current context content and the preset action mapping strategy to obtain the first prompt word;

[0025] The interview interaction question acquisition module is used to call a preset language model, input the first prompt word into the language model, and obtain at least one target action strategy to be used in the next round of dialogue output by the language model; based on the language model, the acquired interview guide information and the at least one target action strategy are processed to obtain the interview interaction question for the next round of dialogue output by the language model.

[0026] The dialogue interaction content acquisition module is used to obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process when the completion of the current interview interaction process is determined based on the language model; wherein, the all dialogue interaction content includes interview interaction questions generated in at least one round of dialogue and the historical interaction content input by the interviewee;

[0027] The analysis module is used to call the language model to analyze all the dialogue interaction content and obtain the dialogue interaction content analysis results.

[0028] According to the knowledge-guided interview interaction data processing system of the present invention, before obtaining the current context content in each round of dialogue during the current interview interaction process, the system further includes:

[0029] The guide building module is specifically used for:

[0030] The process involves obtaining the target question input by the interviewee, concatenating the target question with a preset question mapping strategy to obtain a second prompt word, and then calling a preset language model by inputting the second prompt word to obtain at least one inquiry question output by the language model that is associated with the target question. The question mapping strategy includes multiple question mapping relationships corresponding to each target question, and the inquiry question is a branch question or extension question based on the target question.

[0031] Obtain historical dialogue data containing the target question and at least one probing question; obtain corresponding interview guide information based on the historical dialogue data containing the target question and at least one probing question; and store the interview guide information in a database.

[0032] According to the knowledge-guided interview interaction data processing system of the present invention, the step of concatenating the target question and a preset question mapping strategy to obtain a second prompt word; invoking a preset language model, inputting the second prompt word into the language model, and obtaining at least one inquiry question associated with the target question output by the language model, specifically includes:

[0033] Based on the target question and the first question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a third prompt word; a preset language model is invoked, and the third prompt word is input into the language model to obtain a first analysis and processing result;

[0034] Based on the target question and the second question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a fourth prompt word; a preset language model is invoked, and the fourth prompt word is input into the language model to obtain a second analysis and processing result;

[0035] The fifth prompt word is obtained by concatenating the target question and the third question mapping relationship in the question mapping strategy; the fifth prompt word is input into the preset language model to obtain at least one inquiry question associated with the target question; wherein, the second prompt word includes the third prompt word, the fourth prompt word and the fifth prompt word; the question mapping relationship includes the first question mapping relationship, the second question mapping relationship and the third question mapping relationship.

[0036] According to the knowledge-guided interview interaction data processing system of the present invention, before processing the acquired interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model, the system further includes: an interview guide acquisition module, used to acquire interview guide information corresponding to the current interview interaction process from the database.

[0037] According to the knowledge-guided interview interaction data processing system of the present invention, the interview interaction question acquisition module is specifically used for:

[0038] The language model is invoked to combine the at least one target action strategy, and corresponding inquiry questions are generated based on the interview guide information to obtain the interview interaction questions for the next round of dialogue output by the language model; wherein, the at least one target action strategy includes at least one of questioning action strategy, response action strategy, inquiry action strategy and advancement action strategy.

[0039] According to the knowledge-guided interview interaction data processing system of the present invention, the dialogue interaction content acquisition module is specifically used for:

[0040] The actual number of dialogue rounds in the current interview interaction process is detected. If the actual number of dialogue rounds reaches a preset dialogue round threshold, the current interview interaction process is determined to be completed.

[0041] After determining that the current interview interaction process has been completed, the dialogue interaction content of the current interview interaction process is summarized based on the language model to obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process.

[0042] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the knowledge-guided interview interaction data processing method as described in any of the preceding claims.

[0043] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the knowledge-guided interview interaction data processing method as described in any of the preceding claims.

[0044] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the knowledge-guided interview interaction data processing method as described in any of the preceding claims.

[0045] The knowledge-guided interview interaction data processing method provided by this invention obtains the current context content in each round of dialogue during the current interview interaction process, and performs concatenation processing based on the current context content and action mapping strategy to obtain a first prompt word; inputs the first prompt word into a language model to obtain at least one target action strategy to be used in the next round of dialogue; processes the obtained interview guide information and at least one target action strategy based on the language model to obtain the interview interaction question for the next round of dialogue output by the language model; when the language model determines that the current interview interaction process is complete, obtains all dialogue interaction content corresponding to the current interview interaction process, and analyzes all dialogue interaction content to obtain dialogue interaction content analysis results; it can effectively improve the efficiency of interview interaction data processing, thereby significantly reducing labor costs. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0047] Figure 1 This is a flowchart illustrating the knowledge-guided interview interaction data processing method provided by the present invention.

[0048] Figure 2 This is an interactive diagram corresponding to the guide building module provided by existing technology.

[0049] Figure 3 This is an interactive diagram corresponding to the dialogue module provided by the present invention.

[0050] Figure 4 This is a schematic diagram of system calls in the hardware environment provided by the present invention.

[0051] Figure 5 This is a schematic diagram of the operation of the knowledge-guided interview interaction data processing system provided by the present invention.

[0052] Figure 6 This is a schematic diagram of the structure of the knowledge-guided interview interaction data processing system provided by the present invention.

[0053] Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0055] The following is combined with Figures 1-7 This invention describes the knowledge-guided interview interaction data processing method and system of the present invention, and provides a detailed description of its embodiments.

[0056] The following is a detailed description of embodiments of the knowledge-guided interview interaction data processing method described in this invention. For example... Figure 1 As shown, this is a flowchart illustrating the knowledge-guided interview interaction data processing method provided by the present invention. The specific implementation process includes the following steps:

[0057] Step 101: Obtain the current context content in each round of dialogue during the current interview interaction process, and perform splicing processing based on the current context content and the preset action mapping strategy to obtain the first prompt word.

[0058] In this embodiment of the invention, before obtaining the current context content in each round of dialogue during the current interview interaction, an interview guide needs to be pre-constructed. Specifically, firstly, multiple target questions (i.e., the main questions of the interview) input by interviewees are obtained. Based on the target questions and a preset question mapping strategy, a concatenation process is performed to obtain a second prompt word. A preset language model is invoked, and the second prompt word is input into the language model to obtain at least one inquiry question associated with the target question output by the language model. The question mapping strategy includes multiple question mapping relationships corresponding to each target question. The inquiry question is a branch question or extension question associated with the target question, explored based on the target question. The language model is a preset Large Language Model (LLM), which refers to a deep learning model trained using a large amount of text data, capable of generating natural language text or understanding the meaning of language text. Furthermore, historical dialogue data containing the target question and the at least one probing question can be obtained. This historical dialogue data may also contain other dialogue data besides the target question and the probing question. Based on obtaining the historical dialogue data containing the target question and the at least one probing question, corresponding interview guide information can be obtained (that is, the historical dialogue data containing the target question and the at least one probing question can be packaged into an interview guide to obtain interview guide information). The interview guide information or interview guide can be stored in a database.

[0059] In the process of concatenating the target question and a preset question mapping strategy to obtain a second prompt word; and then calling a preset language model and inputting the second prompt word into the language model to obtain at least one inquiry question associated with the target question, the concatenation process can be performed based on the target question and a first question mapping relationship in the question mapping strategy to obtain a third prompt word; and then calling a preset language model and inputting the third prompt word into the language model to obtain a first analysis result. Concatenation processing is performed based on the target question and a second question mapping relationship in the question mapping strategy to obtain a fourth prompt word; and then calling a preset language model and inputting the fourth prompt word into the language model to obtain a second analysis result. Concatenation processing is performed based on the target question and a third question mapping relationship in the question mapping strategy to obtain a fifth prompt word; and then calling a preset language model and inputting the fifth prompt word into the language model to obtain at least one inquiry question associated with the target question. The second prompt word includes the third, fourth, and fifth prompt words; and the question mapping relationship includes the first, second, and third question mapping relationships. Specifically, the first question mapping relationship can be a mapping relationship between the target question and the prompt words of the first step; the second question mapping relationship can be a mapping relationship between the target question and the prompt words of the second step; the third question mapping relationship can be a mapping relationship between the target question and the prompt words of the third step. Additionally, the question mapping relationship may also include a fourth question mapping relationship. The fourth question mapping relationship can be a mapping relationship between the target question and the prompt words of the fourth step. Wherein, for example... Figure 2 As shown, the prompt for the first step can be "understand the main question" or "understand the main question." "Understanding the main question" allows the knowledge model to explain the meaning and function of each target question, making the model's subsequent output more specific and relevant. The prompt for the second step can be "predict the direction of the conversation." "Predicting the direction of the conversation" allows the knowledge model to predict various possible scenarios during the interview based on the main question. The prompt for the third step can be "generate probing questions." "Generate probing questions" allows the knowledge model to design multiple probing questions to be asked in different situations, referencing the possible directions of the interview. The prompt for the fourth step can be "propose quantitative indicators." "Propose quantitative indicators" allows the knowledge model to output corresponding quantitative indicators for each interviewee. These quantitative indicators are used in the post-interview analysis and processing.

[0060] It should be noted that although the interviewer will make adjustments according to the actual progress of the semi-structured interview, such adjustments still need to be made within or at least around a predefined question framework. This predefined question framework is called the interview guide, which contains interview guide information. Therefore, the interview guide plays a key role in the semi-structured interview. The interview guide should contain two layers of open-ended questions: (1) main questions, which are broad topics of interest and are used to guide the overall direction of the interview dialogue, and therefore need to be pre-built in the interview guide; (2) probing questions, which are derived from the main questions and are designed to explore specific points that are particularly valuable in the discussion, and are generated by the guide building module. For example: Input a series of main questions {Mi}, and the guide building module (

[0061] The Guide Construction Module (GCM) generates multiple inquiry questions {Pi, j} for each of the main problems. That is:

[0062]

[0063] Here, GCM is a mapping from a principal problem Mi to multiple corresponding inquiry problems Pi,j.

[0064] The interview guide consists of the main input questions and the generated inquiry questions:

[0065]

[0066] This generation involves bridging two differences between the two layers of questions. (1) General vs. Specific: The main questions establish the framework of the interview, while the probing questions must delve into the details of each main question, requiring a thorough understanding of them; (2) Expected vs. Actual: The main questions outline the expected interview questions, while the probing questions must cover potentially valuable points that may arise during the interview, requiring prediction of the actual process. This invention begins with the provided main questions, such as... Figure 2 As shown. Specifically, in the multi-round dialogue, the knowledge model is guided to (1) understand the main issues, resolving the first difference (i.e., general vs. specific), (2) predict the potential direction of the interview, resolving the second difference (i.e., expectation vs. reality), and finally (3) call the knowledge model to generate at least one inquiry question for each main issue. In addition, the present invention also designs an additional step, namely (4) indicator design, which calls the knowledge model to generate quantitative indicators for each interviewee for subsequent analysis of the interview.

[0067] like Figure 4As shown, in the specific process of constructing or generating the interview guide: (1) Through client (front-end) interaction, guide the user to input the main questions (i.e. target questions) involved in the interview in a given way. (2) Combine the main questions and use the template designed in advance through prompting engineering to generate prompt words containing four steps, guide the big language model to achieve the following functions: understand the main questions: let the big language model explain the meaning and role of each main question, so that the subsequent output of the model is more specific and relevant; predict the direction of the conversation: let the big language model predict various situations that may occur in the interview based on the main questions; generate inquiry questions: let the big language model design inquiry questions to be raised in different situations with reference to the possible direction of the interview; propose quantitative indicators: let the big language model generate quantitative indicators for each interviewee for subsequent analysis of the interview. (3) Integrate the output of the big language model in the above steps to obtain the interview guide.

[0068] Before performing this step, interview guide information corresponding to the current interview interaction process can be obtained from the database.

[0069] Step 102: Invoke the preset language model, input the first prompt word into the language model, and obtain at least one target action strategy to be used in the next round of dialogue output by the language model; process the obtained interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model.

[0070] In this embodiment of the invention, the language model is invoked to combine the at least one target action strategy, and corresponding inquiry questions are generated based on the interview guide information to obtain the interview interaction questions for the next round of dialogue output by the language model; wherein, the at least one target action strategy includes at least one of questioning action strategy, response action strategy, inquiry action strategy, and advancement action strategy.

[0071] The questioning action strategy can be to send questions to the interviewee; the response action strategy can be to provide answers to the interviewee; the inquiry action strategy can be to send inquiring questions to the interviewee; and the advancement action strategy can be to send questions to the interviewee.

[0072] The questioning strategy is as follows: Figure 3 The action strategy for asking questions, marked in blue in the image. Figure 3 The blue-labeled message "Thank you for participating in the interview. Do you often use large models?" output by the knowledge model (i.e., the agent) on the left is one type of dialogue content corresponding to the questioning action strategy; the response action strategy is... Figure 3 The action strategy for the response action marked in red in the text. Figure 3 The red-marked "Indeed, large models play an increasingly important role in our lives" output by the knowledge model (i.e., the agent) on the left is one type of dialogue content corresponding to the response action strategy; the inquiry action strategy is... Figure 3 The orange markers in the text indicate the action strategies for probing actions. Figure 3 The orange-marked message "Can you tell me some specific scenarios in which you use the large model?" output by the knowledge model (i.e., the agent) on the left is one type of dialogue content corresponding to the inquiry action strategy; the advancing action strategy is... Figure 3 The purple-marked propulsion action strategy Figure 3 The purple-marked message output by the knowledge model (i.e., the agent) on the left, "We just mentioned reading books and writing papers. Now, let's discuss some education-related topics," represents one type of dialogue content corresponding to the action strategy.

[0073] Step 103: If the current interview interaction process is determined to be completed based on the language model, obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process; wherein, the all dialogue interaction content includes interview interaction questions generated in at least one round of dialogue and the historical interaction content input by the interviewee.

[0074] In this embodiment of the invention, the actual number of dialogue rounds in the current interview interaction process is detected. If the actual number of dialogue rounds reaches a preset dialogue round threshold (e.g., 10 rounds), the current interview interaction process is determined to be complete. After determining that the current interview interaction process is complete, the dialogue interaction content of the current interview interaction process is summarized based on the language model to obtain all dialogue interaction content corresponding to the current interview interaction process output by the language model.

[0075] It should be noted that, in the dialogue module processing, the main advantage of semi-structured interviews lies in allowing the interviewer to focus on the planned route while still giving them autonomy to explore relevant ideas that arise during the interview. However, these advantages also present challenges for even experienced human interviewers in controlling the pace, namely, maintaining a balance between two conflicting aspects: (1) following pre-built interview guidelines; and (2) exploring details that arise during the actual process to obtain additional information. This requirement necessitates fine-grained control over the behavior of the interview agent. Following the well-known state-action-reward reinforcement learning paradigm, the dialogue in the current interview is constructed as a multi-turn dialogue in the current context content (i.e., context), action, and information process. Here, the current context content refers to the natural language dialogue history alternating between the interviewee (interviewee) and the language model (i.e., the large language model invoked in the interview system). i When taking turns, use QiThis indicates the questions asked in this round of the interview system (i.e., the interview interaction questions generated in each round of dialogue), and uses... Ai Indicates the respondents' opinion Qi The responses (i.e., the historical interaction content input by the interviewee), the current context content (i.e.) ) can be represented as:

[0076]

[0077] In this embodiment of the invention, an action (i.e., a target action strategy) refers to the behavior of the agent (i.e., the language model invoked by the interview system) in the next question (i.e., the interview interaction question in the next round of dialogue) given the current context. By defining action types and the conditions applicable to each action, the behavior of the language model can be finely tuned to maximize the information to be collected. Specifically, each question posed by the language model... Qi (i.e., interview questions) can all be broken down into multiple actions in sequence. i , j (i.e., at least one target action strategy):

[0078]

[0079] Based on the current context content Select a series of actions (Action) i , j The process is a mapping, called a pre-defined action mapping strategy. P ):

[0080]

[0081] In this embodiment of the invention, adjusting the behavior of the agent (i.e., the interviewee) according to the above definition involves defining an action space and establishing a strategy (i.e., an action mapping strategy). For the action space, in order to simultaneously satisfy two conflicting aspects of a semi-structured interview, two actions are defined for each aspect. Figure 3 An example is used to illustrate this. Strategies are encoded as principles in corresponding cue words (such as action mapping strategies), which specify behaviors and applicable conditions through a set of natural language guidelines.

[0082] like Figure 4 As shown, in the specific process of the dialogue module: (1) It is necessary to pre-select and summarize classic literature and existing interview data, and obtain the strategy P used by the language model through existing prompting engineering technology. (2) In each round of dialogue in the current interview interaction process, the language model is called, and the interview guide information and the current context content of the interview guide are input into the language model. (3) The language model processes the acquired interview guide information and the current context content, and selects at least one specific action based on the interview guide information and the current context content. i , j (i.e., at least one target action strategy), further by Action i , j Combined into questions Qi (i.e., the interview interaction questions for the next round of dialogue). (4) Call upon existing text-to-speech technology, such as Figure 4 Data transformation engines in existing technologies generate questions Qi The voice is transmitted to the client. (5) Client display Qi , and play the audio. (6) Obtain the interviewee's input (i.e., the interviewee's historical interaction content) through client interaction: (a) Call the existing automatic speech recognition technology to recognize and convert the interviewee's voice input into text, (b) concatenate the converted voice input with the user's text input as the interviewee's response. Qi The answer Ai (i.e., the historical interaction content input by the interviewee). (6) Repeat the question-and-answer process of (2)-(4) above until the language model determines that the interview has been completed (i.e., determines that the current interview interaction process has been completed).

[0083] Step 103: Call the language model to analyze all the dialogue interaction content and obtain the dialogue interaction content analysis results.

[0084] In this embodiment of the invention, the original output of the dialogue module in step 102 includes all dialogue interaction content (i.e., a series of questions and answers). Without analysis, this output cannot be effectively utilized in various interview application scenarios. This invention can analyze all dialogue interaction content from both qualitative and quantitative dimensions. In the qualitative dimension, this invention can automatically summarize the dialogue information. Similar to an analysis memo, the summary contains key points and main findings about the interview. In the quantitative dimension, as the data scale increases, qualitative analysis becomes difficult and often leads to a loss of generality in the research. Therefore, a large language model is used to analyze the interview and obtain scores for the indicators proposed in the final stage of guideline construction. The explanation of the scores is generated along with the scores to enhance credibility. The specific implementation process of this invention is as follows: Input the complete dialogue history (i.e., all dialogue interaction content) obtained from the dialogue module, embed it into a preset template, and generate corresponding qualitative and quantitative analysis prompts. During the process of calling and guiding the large language model to generate: Qualitative analysis results, including a summary of the dialogue and main conclusions; quantitative analysis results, analyzed according to each quantitative analysis indicator proposed in the guideline construction module, providing scores and reasons for the scores.

[0085] In addition, in this embodiment of the invention, after calling the language model to analyze all the dialogue interaction content and obtaining the dialogue interaction content analysis results, the dialogue interaction content and its corresponding dialogue interaction content analysis results can be stored in the database so that the dialogue interaction content can be read from the database or the dialogue interaction content analysis results can be obtained from the database and sent to the client for display in the form of charts.

[0086] The knowledge-guided interview interaction data processing method provided by this invention obtains the current context content in each round of dialogue during the current interview interaction process, and performs concatenation processing based on the current context content and action mapping strategy to obtain a first prompt word; inputs the first prompt word into a language model to obtain at least one target action strategy to be used in the next round of dialogue; processes the obtained interview guide information and at least one target action strategy based on the language model to obtain the interview interaction question for the next round of dialogue output by the language model; when the language model determines that the current interview interaction process is complete, obtains all dialogue interaction content corresponding to the current interview interaction process, and analyzes all dialogue interaction content to obtain dialogue interaction content analysis results; it can effectively improve the efficiency of interview interaction data processing, thereby significantly reducing labor costs.

[0087] The knowledge-guided interview interaction data processing system provided by this invention is described below. The knowledge-guided interview interaction data processing system described below can be referred to in conjunction with the knowledge-guided interview interaction data processing method described above. References Figure 6 The diagram shown is a structural schematic of the knowledge-guided interview interaction data processing system provided by this invention. The knowledge-guided interview interaction data processing system of this invention specifically includes the following parts:

[0088] The prompt word acquisition module 601 is used to obtain the current context content in each round of dialogue during the current interview interaction process, and to perform splicing processing based on the current context content and a preset action mapping strategy to obtain the first prompt word.

[0089] The interview interaction question acquisition module 602 is used to call a preset language model, input the first prompt word into the language model, and obtain at least one target action strategy to be used in the next round of dialogue output by the language model; based on the language model, the acquired interview guide information and the at least one target action strategy are processed to obtain the interview interaction question for the next round of dialogue output by the language model.

[0090] The dialogue interaction content acquisition module 603 is used to obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process when the completion of the current interview interaction process is determined based on the language model; wherein, the all dialogue interaction content includes interview interaction questions generated in at least one round of dialogue and the historical interaction content input by the interviewee.

[0091] The analysis module 604 is used to call the language model to analyze all the dialogue interaction content and obtain the dialogue interaction content analysis results.

[0092] According to the knowledge-guided interview interaction data processing system of the present invention, before obtaining the current context content in each round of dialogue during the current interview interaction process, the system further includes:

[0093] The guide building module is specifically used for:

[0094] The process involves obtaining the target question input by the interviewee, concatenating the target question with a preset question mapping strategy to obtain a second prompt word, and then calling a preset language model by inputting the second prompt word to obtain at least one inquiry question output by the language model that is associated with the target question. The question mapping strategy includes multiple question mapping relationships corresponding to each target question, and the inquiry question is a branch question or extension question based on the target question.

[0095] Obtain historical dialogue data containing the target question and at least one probing question; obtain corresponding interview guide information based on the historical dialogue data containing the target question and at least one probing question; and store the interview guide information in a database.

[0096] According to the knowledge-guided interview interaction data processing system of the present invention, the step of concatenating the target question and a preset question mapping strategy to obtain a second prompt word; invoking a preset language model, inputting the second prompt word into the language model, and obtaining at least one inquiry question associated with the target question output by the language model, specifically includes:

[0097] Based on the target question and the first question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a third prompt word; a preset language model is invoked, and the third prompt word is input into the language model to obtain a first analysis and processing result;

[0098] Based on the target question and the second question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a fourth prompt word; a preset language model is invoked, and the fourth prompt word is input into the language model to obtain a second analysis and processing result;

[0099] The fifth prompt word is obtained by concatenating the target question and the third question mapping relationship in the question mapping strategy; the fifth prompt word is input into the preset language model to obtain at least one inquiry question associated with the target question; wherein, the second prompt word includes the third prompt word, the fourth prompt word and the fifth prompt word; the question mapping relationship includes the first question mapping relationship, the second question mapping relationship and the third question mapping relationship.

[0100] According to the knowledge-guided interview interaction data processing system of the present invention, before processing the acquired interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model, the system further includes: an interview guide acquisition module, used to acquire interview guide information corresponding to the current interview interaction process from the database.

[0101] According to the knowledge-guided interview interaction data processing system of the present invention, the interview interaction question acquisition module is specifically used for:

[0102] The language model is invoked to combine the at least one target action strategy, and corresponding inquiry questions are generated based on the interview guide information to obtain the interview interaction questions for the next round of dialogue output by the language model; wherein, the at least one target action strategy includes at least one of questioning action strategy, response action strategy, inquiry action strategy and advancement action strategy.

[0103] According to the knowledge-guided interview interaction data processing system of the present invention, the dialogue interaction content acquisition module is specifically used for:

[0104] The actual number of dialogue rounds in the current interview interaction process is detected. If the actual number of dialogue rounds reaches a preset dialogue round threshold, the current interview interaction process is determined to be completed.

[0105] After determining that the current interview interaction process has been completed, the dialogue interaction content of the current interview interaction process is summarized based on the language model to obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process.

[0106] In embodiments of the present invention, such as Figure 5 The dialogue module shown includes the prompt word acquisition module 601, the interview interaction question acquisition module 602, and the dialogue interaction content acquisition module 603. For example... Figure 5 In this invention, interviews are created in an interview system (i.e., a knowledge-guided interview interaction data processing system), key questions are uploaded, interviewees log in to the system to participate in the current interview interaction, the system analyzes the interview data, and provides it to researchers. Internally, the database is implemented using Python's sqlite3. Externally, the LLM API is a large model interface, and this interview system uses the GPT-4 large model; TTS (Text-To-Speech) is a text-to-speech module; and ASR (Automatic Speech Recognition) is a speech-to-text module. The interview system is LM-Interview. This invention addresses the aforementioned problems by utilizing a knowledge-guided language model. LM-Interview solves these problems through three modules: a guide construction module, a dialogue module, and an analysis module. Qualitative users can utilize the system to build interview guidelines before the interview, then collect large amounts of data through LLM-driven interviews without manual labor. Finally, insights are gained from the system's analysis of the interview data to improve analytical efficiency. A method for automatically conducting semi-structured interviews using a knowledge-guided language model is proposed. LM-Interview has been implemented to support the entire process of designing, conducting, and analyzing semi-structured interviews for qualitative users.

[0107] like Figure 4The diagram illustrates how the various hardware architectures collaborate. In the guide construction module's processing: (a) the user accesses the guide construction module's page; (b) the front-end (i.e., the client) sends a GET request to the / create address to retrieve the front-end page and styles; (c) the user fills in the main questions and sends a POST request to the / create address; (d) the back-end (i.e., the server) calls the language model to generate the interview guide; (e) the data (i.e., the interview guide or interview guide information) is inserted into the database.

[0108] In the dialogue module processing: (f) The interviewee (i.e., the interviewee) visits the interview page. The front end sends a GET request to the / interview address to obtain the front end page and style. The back end creates an entry in the database to record the dialogue data (i.e., all dialogue interaction content). (g) The interviewee inputs the answer (i.e., the interaction content entered by the interviewee) and sends a POST request to the / interview address; (h) The interview system raises a question: (1) The back end retrieves the corresponding dialogue data from the database; (2) The back end calls the language model to generate the next question (i.e., the interview interaction question for the next round of dialogue); (3) The back end calls the speech-to-text service to generate the voice of the question; (4) The back end writes to the database to update the dialogue data; (5) The back end sends the generated question and voice to the front end for display; (6) Repeat (g) and (h) until the interview dialogue interaction is completed.

[0109] During the analysis module processing: (1) After the interviewee completes the interview, the analysis module is automatically executed, calling the language model to analyze the dialogue data (i.e., all dialogue interaction content), and storing the analysis (i.e., the analysis results of dialogue interaction content) in the database; (2) The researcher (i.e. the user) accesses the analysis page, and the front end sends a GET request to the / analyze address to obtain the page content and style; (3) The back end queries the database, reads the stored analysis data, and returns it to the front end for display.

[0110] The knowledge-guided interview interaction data processing system provided by this invention obtains the current context content in each round of dialogue during the current interview interaction process, and performs concatenation processing based on the current context content and action mapping strategy to obtain a first prompt word; inputs the first prompt word into a language model to obtain at least one target action strategy to be used in the next round of dialogue; processes the obtained interview guide information and at least one target action strategy based on the language model to obtain the interview interaction question for the next round of dialogue output by the language model; when the language model determines that the current interview interaction process is complete, it obtains all dialogue interaction content corresponding to the current interview interaction process, and analyzes all dialogue interaction content to obtain dialogue interaction content analysis results; it can effectively improve the efficiency of interview interaction data processing, thereby significantly reducing labor costs.

[0111] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7 As shown, the electronic device may include: a processor 701, a communication interface 704, a memory 702, and a communication bus 703, wherein the processor 701, the communication interface 704, and the memory 702 communicate with each other through the communication bus 703. The processor 701 can call logical instructions in the memory 702 to execute a knowledge-guided interview interaction data processing method. This method includes: obtaining the current context content in each round of dialogue during the current interview interaction process; performing concatenation processing based on the current context content and a preset action mapping strategy to obtain a first prompt word; calling a preset language model, inputting the first prompt word into the language model, and obtaining at least one target action strategy to be used in the next round of dialogue output by the language model; processing the obtained interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction question for the next round of dialogue output by the language model; when the current interview interaction process is determined to be complete based on the language model, obtaining all dialogue interaction content corresponding to the current interview interaction process output by the language model; wherein, the all dialogue interaction content includes interview interaction questions generated in at least one round of dialogue and historical interaction content input by the interviewee; and calling the language model to analyze the all dialogue interaction content to obtain dialogue interaction content analysis results.

[0112] Furthermore, the logical instructions in the aforementioned memory 702 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0113] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the knowledge-guided interview interaction data processing method provided by the above methods. The method includes: obtaining the current context content in each round of dialogue in the current interview interaction process; performing concatenation processing based on the current context content and a preset action mapping strategy to obtain a first prompt word; calling a preset language model, inputting the first prompt word into the language model, and obtaining at least one target action strategy to be used in the next round of dialogue output by the language model; processing the obtained interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction question for the next round of dialogue output by the language model; when it is determined based on the language model that the current interview interaction process is completed, obtaining all dialogue interaction content output by the language model corresponding to the current interview interaction process; wherein, the all dialogue interaction content includes the interview interaction question generated in at least one round of dialogue and the historical interaction content input by the interviewee; calling the language model to analyze the all dialogue interaction content to obtain the dialogue interaction content analysis result.

[0114] In another aspect, this application also provides a computer-readable storage medium, which includes a stored program, wherein the program, when running, executes the knowledge-guided interview interaction data processing method provided by the above methods. This method includes: obtaining current context content in each round of dialogue during the current interview interaction process; performing concatenation processing based on the current context content and a preset action mapping strategy to obtain a first prompt word; invoking a preset language model, inputting the first prompt word into the language model, and obtaining at least one target action strategy to be used in the next round of dialogue output by the language model; processing the obtained interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction question for the next round of dialogue output by the language model; when the current interview interaction process is determined to be complete based on the language model, obtaining all dialogue interaction content corresponding to the current interview interaction process output by the language model; wherein the all dialogue interaction content includes interview interaction questions generated in at least one round of dialogue and historical interaction content input by the interviewee; and invoking the language model to analyze the all dialogue interaction content to obtain dialogue interaction content analysis results.

[0115] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0116] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A knowledge-guided interview interaction data processing method, characterized in that, include: In each round of dialogue during the current interview interaction, the current context content is obtained, and the first prompt word is obtained by splicing the current context content and the preset action mapping strategy. Invoke a preset language model, input the first prompt word into the language model, and obtain at least one target action strategy to be used in the next round of dialogue output by the language model; Based on the language model, the acquired interview guide information and the at least one target action strategy are processed to obtain the interview interaction questions for the next round of dialogue output by the language model. If the current interview interaction process is determined to be completed based on the language model, obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process; The total dialogue interaction content includes at least one round of dialogue generated interview interaction questions and the historical interaction content input by the interviewee; The language model is invoked to analyze all the dialogue interaction content to obtain the dialogue interaction content analysis results.

2. The knowledge-guided interview interaction data processing method according to claim 1, characterized in that, Before obtaining the current context content in each round of dialogue during the current interview interaction, it also includes: The process involves obtaining the target question input by the interviewee, concatenating the target question with a preset question mapping strategy to obtain a second prompt word, and then calling a preset language model by inputting the second prompt word to obtain at least one inquiry question output by the language model that is associated with the target question. The question mapping strategy includes multiple question mapping relationships corresponding to each target question, and the inquiry question is a branch question or extension question based on the target question. Obtain historical dialogue data containing the target question and at least one probing question; obtain corresponding interview guide information based on the historical dialogue data containing the target question and at least one probing question; and store the interview guide information in a database.

3. The knowledge-guided interview interaction data processing method according to claim 2, characterized in that, The second prompt word is obtained by concatenating the target question and a preset question mapping strategy. A preset language model is invoked, and the second prompt word is input into the language model to obtain at least one inquiry question associated with the target question output by the language model, specifically including: Based on the target question and the first question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a third prompt word; a preset language model is invoked, and the third prompt word is input into the language model to obtain a first analysis and processing result; Based on the target question and the second question mapping relationship in the question mapping strategy, a concatenation process is performed to obtain a fourth prompt word; a preset language model is invoked, and the fourth prompt word is input into the language model to obtain a second analysis and processing result; The fifth prompt word is obtained by concatenating the target question and the third question mapping relationship in the question mapping strategy; the fifth prompt word is input into the preset language model to obtain at least one inquiry question associated with the target question; wherein, the second prompt word includes the third prompt word, the fourth prompt word and the fifth prompt word; the question mapping relationship includes the first question mapping relationship, the second question mapping relationship and the third question mapping relationship.

4. The knowledge-guided interview interaction data processing method according to claim 2, characterized in that, Before processing the acquired interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model, the method further includes: obtaining interview guide information corresponding to the current interview interaction process from the database.

5. The knowledge-guided interview interaction data processing method according to claim 1, characterized in that, The process of processing the acquired interview guide information and the at least one target action strategy based on the language model to obtain the interview interaction questions for the next round of dialogue output by the language model includes: The language model is invoked to combine the at least one target action strategy, and corresponding inquiry questions are generated based on the interview guide information to obtain the interview interaction questions for the next round of dialogue output by the language model; wherein, the at least one target action strategy includes at least one of questioning action strategy, response action strategy, inquiry action strategy and advancement action strategy.

6. The knowledge-guided interview interaction data processing method according to claim 1, characterized in that, When the current interview interaction process is determined to be complete based on the language model, obtaining all dialogue interaction content output by the language model corresponding to the current interview interaction process specifically includes: The actual number of dialogue rounds in the current interview interaction process is detected. If the actual number of dialogue rounds reaches a preset dialogue round threshold, the current interview interaction process is determined to be completed. After determining that the current interview interaction process has been completed, the dialogue interaction content of the current interview interaction process is summarized based on the language model to obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process.

7. A knowledge-guided interview interaction data processing system, characterized in that, include: The prompt word acquisition module is used to obtain the current context content in each round of dialogue during the current interview interaction process, and to perform splicing processing based on the current context content and the preset action mapping strategy to obtain the first prompt word; The interview interaction question acquisition module is used to call a preset language model, input the first prompt word into the language model, and obtain at least one target action strategy to be used in the next round of dialogue output by the language model. Based on the language model, the acquired interview guide information and the at least one target action strategy are processed to obtain the interview interaction questions for the next round of dialogue output by the language model. The dialogue interaction content acquisition module is used to obtain all dialogue interaction content output by the language model corresponding to the current interview interaction process when the completion of the current interview interaction process is determined based on the language model; wherein, the all dialogue interaction content includes interview interaction questions generated in at least one round of dialogue and the historical interaction content input by the interviewee; The analysis module is used to call the language model to analyze all the dialogue interaction content and obtain the dialogue interaction content analysis results.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the knowledge-guided interview interaction data processing method as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the knowledge-guided interview interaction data processing method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the knowledge-guided interview interaction data processing method as described in any one of claims 1 to 6.

Images