Interview report generation method and device based on large model

Abstract

The application provides a large model-based interview report generation method and device. An interview report template is set in advance; an interviewee information column and an interview content column of the interview report template each reserve a predetermined filling position, which is filled by a variable; voice recognition is performed on voice content recorded during an interview, and interview content in the form of recognized text is input into a large model; personnel basic information analyzed and output by the large model is taken as a value of a variable reserved in the interviewee information column, and is filled into the interviewee information column in the interview report template; and interview content points analyzed and output by the large model are taken as values of variables reserved in the interview content column, and are filled into the interview content column in the interview report template, so that an interview report corresponding to the interview is obtained. The interview report generation method of the embodiment of the application improves intelligence, avoids time and labor consumption of personnel, and improves efficiency.

Description

TECHNICAL FIELD

[0001] One or more embodiments of the present application relate to network communication technology, and particularly relate to a large model-based interview report generation method and device. BACKGROUND

[0002] In today's society, informatization is developing rapidly, and various reports often need to be generated in various industries. Staff need to spend a lot of effort to extract the required content of the report according to various forms of content, and finally generate the report.

[0003] However, the current report generation method is manually analyzed by the staff to fill in the required content of the report, so as to generate the report. For example, the staff needs to interview relevant personnel for investigation, and generate an interview report according to the interview content. It can be seen that this kind of interview report generation method has the problems of not being intelligent, time-consuming, labor-intensive, low efficiency, etc. SUMMARY

[0004] One or more embodiments of the present application describe a large model-based interview report generation method and device, which can solve at least one problem in the prior art.

[0005] According to a first aspect, a large model-based interview report generation method is provided, which comprises:

[0006] An interview report template is set in advance, including an interviewee information column, an interview content column, and a signature column, and a corresponding label is set for each column in the interviewee information column, the interview content column, and the signature column in the interview report template. In the interviewee information column and the interview content column of the interview report template, a predetermined filling position is reserved, and the filling position reserved in the interview report template is filled by a variable;

[0007] Obtain the voice content recorded in the interview process;

[0008] Perform speech recognition on the obtained voice content to obtain interview content in text form;

[0009] Input the interview content in text form into a pre-trained large model;

[0010] The basic information of the personnel analyzed and output by the large model is used as the value of the variable reserved in the interviewee information column, and the value of the variable is filled into the interviewee information column in the interview report template by using the label of the interviewee information column;

[0011] The key points of the interview content analyzed and output by the large model are used as the value of the variable reserved in the interview content column, and the value of the variable is filled into the interview content column in the interview report template by using the label of the interview content column, thereby obtaining an interview report corresponding to the interview process.

[0012] After the interview report corresponding to the interview process is obtained, further comprising:

[0013] providing the interview report corresponding to the interview process to the manager in the form of an online editable document, so that the manager confirms the interview content points in the interview report, the manager prints the interview report after confirmation, and the manager signs in the signature column of the interview report.

[0014] The training method of the large model comprises:

[0015] Set up in-set words, input the text content including the in-set words and the label of the in-set words into the large model, so as to train the large model to recognize the content of the interviewee information column; wherein the in-set words include at least one of name, gender, age, position, address and mobile phone number.

[0016] The training method of the large model comprises: obtaining at least one conversation training sample, each conversation training sample including a text form of conversation content and an interview content point as a label of the conversation content; inputting the at least one conversation training sample into the large model, so as to train the large model to intelligently analyze the text form of conversation content by using the deep learning and natural language processing capabilities of the large model, and automatically generate the interview content point.

[0017] The interview content point includes at least one of inquiry question, warning reminder and suggestion.

[0018] According to the second aspect, a large model-based interview report generation device is provided, which comprises:

[0019] The interview report template generation module is configured to pre-set the interview report template to include the interviewee information column, the interview content column and the signature column, and set corresponding labels for each column in the interviewee information column, the interview content column and the signature column of the interview report template; pre-reserve a predetermined filling position in the interviewee information column and the interview content column of the interview report template, and fill the filling position reserved in the interview report template by variable;

[0020] The to-be-recognized content acquisition module is configured to acquire the voice content recorded in the interview process;

[0021] The conversion module is configured to perform voice recognition on the acquired voice content, thereby obtaining text form of interview content, and inputting the text form of interview content into the pre-trained large model;

[0022] The interview report generation module is configured to: take the personnel basic information analyzed and output by the large model as a value of a variable reserved in an interviewee information field, fill the value of the variable into the interviewee information field in an interview report template by using a label of the interviewee information field, take the interview content points analyzed and output by the large model as a value of a variable reserved in an interview content field, fill the value of the variable into the interview content field in the interview report template by using a label of the interview content field, and thus obtain an interview report corresponding to the interview process.

[0023] The device further comprises a docking module.

[0024] The docking module is configured to provide the interview report obtained by the interview report generation module to the manager in the form of an online editable document, so that the manager confirms the interview content points in the interview report, the manager prints the interview report after the confirmation, and the manager performs signature processing in a signature field of the interview report.

[0025] The device further comprises a large model training module.

[0026] The large model training module is configured to perform:

[0027] setting an in-set word, inputting text content including the in-set word and a label of the in-set word into the large model, so as to train the large model to recognize the content of the interviewee information field; wherein the in-set word includes at least one of a name, a gender, an age, a position, an address, and a mobile phone number;

[0028] and / or,

[0029] obtaining at least one conversation training sample, each conversation training sample including a piece of conversation content in a text form and interview content points as a label of the conversation content, inputting the at least one conversation training sample into the large model, so as to train the large model to intelligently analyze the conversation content in the text form by using deep learning and natural language processing capabilities of the large model, and automatically generate the interview content points.

[0030] The interview content points include at least one of an inquiry question, a warning reminder, and a suggestion.

[0031] According to a third aspect, a computing device is provided, including a memory and a processor, the memory stores executable code, and the processor executes the executable code to implement the method of any of the embodiments of the application.

[0032] The large model-based interview report generation method and device provided by each of the embodiments have at least the following beneficial effects:

[0033] 1、The embodiment of the application pre-sets the format of the interview report template, in order to meet the format characteristics of such template, that is, a predetermined filling position is reserved, in the embodiment of the application, the variable is filled in the filling position reserved in the interview report template, thereby solving the formatting design problem of such template, and obtaining a universal interview report template. Moreover, subsequently, only the value of the variable needs to be filled in, and then the corresponding interview content can be filled in the template. Further, in the embodiment of the application, according to the label of each column, the corresponding column position in the template can be quickly found and positioned when generating a specific interview report subsequently, thereby ensuring the generation of the interview report.

[0034] 2、The interview report generation method of the embodiment of the application uses voice-to-text technology to realize accurate transcription of the conversation content, then uses a large model to deeply mine and extract key information, and finally automatically weaves into an interview report with clear structure and prominent points. This process not only greatly improves the efficiency and accuracy of work, but also effectively reduces the manual writing burden of the staff, making the work more intelligent and standardized. The interview report generation method of the embodiment of the application improves the intelligence, avoids the time and effort of personnel, and improves the efficiency.

[0035] 3、Through the multiple technologies of the large model (speech recognition, keyword extraction, generation of official documents), the interview report is automatically generated, greatly improving the work efficiency and reducing the complexity of manual operation.

[0036] 4、Improve the accuracy and consistency of the report: the large model can follow the preset rules and standards when extracting keywords and generating reports, ensuring the accuracy and consistency of the report content. This standardized processing not only improves the quality of the report, but also makes the reports generated by different personnel comparable and traceable. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0038] Figure 1 is a flowchart of the interview report generation method based on the large model in an embodiment of the application.

[0039] Figure 2 is a flowchart of the interview report generation method based on the large model in another embodiment of the application.

[0040] Figure 3is a structural schematic diagram of a large model-based interview report generation device in an embodiment of the present application.

[0041] Figure 4 is a structural schematic diagram of a large model-based interview report generation device in another embodiment of the present application. DETAILED DESCRIPTION

[0042] First of all, it should be noted that the terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0043] It should be understood that the term "and / or" used herein is only to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " herein generally represents an "or" relationship between the front and rear associated objects.

[0044] Figure 1 is a flowchart of a large model-based interview report generation method in an embodiment of the present application. Referring to Figure 1 In an embodiment of the present application, the large model-based interview report generation method comprises:

[0045] Step 101: Pre-set an interview report template.

[0046] The specific process of this step 101 can include: pre-setting the interview report template to include an interviewee information column, an interview content column and a signature column, and setting a corresponding label for each column in the interviewee information column, the interview content column and the signature column in the interview report template; pre-reserving a predetermined filling position in the interviewee information column and the interview content column of the interview report template, and filling the filling position reserved in the interview report template through a variable, thereby obtaining a general interview report template.

[0047] The interview report template generally includes three core parts: the interviewee information column, the interview content column and the signature column. In the interviewee information column, a corresponding position is reserved for subsequent filling of relevant information; the interview content column is used to record the specific exchange content during the interview; and the signature column is designed to be filled by hand to ensure its authenticity and seriousness, so the signature column does not need to be pre-filled with content during the template preparation stage.

[0048] Step 103: Obtain the voice content recorded during the interview.

[0049] Step 105: voice recognition is performed on the acquired voice content, so as to obtain the interview content in the form of text.

[0050] For example, during the interview process, in order to ensure the accuracy and integrity of the information, usually audio recording is performed. These audio recordings are processed by voice recognition technology, and are efficiently converted into text form. Through advanced technology, the clarity of the recording can be ensured, so that the converted text content is accurate and correct.

[0051] Step 107: input the interview content in the form of text into a pre-trained large model.

[0052] Step 109: the personnel basic information analyzed and output by the large model is taken as the value of the variable reserved in the interviewee information column, and the value of the variable is filled into the interviewee information column in the interview report template by using the label of the interviewee information column.

[0053] Step 111: the interview content points analyzed and output by the large model are taken as the value of the variable reserved in the interview content column, and the value of the variable is filled into the interview content column in the interview report template by using the label of the interview content column, so as to obtain the interview report corresponding to the interview process.

[0054] For the above steps 107, 109 and 111, from the converted interview content in the form of text, the key information in the text needs to be further extracted using the keyword extraction capability of the large model, such as basic personnel information including name, age, gender, and interview content points such as inquiry questions, warning reminders, and suggestions. For example, for step 111, the extracted inquiry questions, warning reminders, and suggestions are intelligently analyzed by using the deep learning and natural language processing capabilities of the large model, and the interview content conforming to the specification, with clear logic and accurate content, is automatically generated. These generated contents are automatically filled into the interview content column.

[0055] According to the flowchart shown in Figure 1 As can be seen from the flowchart, in the embodiment of the present application, the format of the interview report template is pre-set, in order to meet the format characteristics of such template, that is, a predetermined filling position needs to be reserved, in the embodiment of the present application, the variable is filled in the filling position reserved in the interview report template, so as to solve the formatting design problem of such template, and a general interview report template is obtained. Moreover, in the subsequent, only the value of the variable needs to be filled in, and then the corresponding interview content can be filled in the template. Further, in the embodiment of the present application, according to the label of each column, the corresponding column position in the template can be quickly found and positioned when generating a specific interview report, so as to ensure the generation of the interview report.

[0056] The interview report generation method of the embodiment of the present application uses voice-to-text technology to achieve accurate transcription of the conversation content, then uses a large model to deeply mine and extract key information, and finally automatically weaves a structured and highlighted interview report. This process not only greatly improves the efficiency and accuracy of work, but also effectively reduces the manual writing burden of staff, making the work more intelligent and standardized. The interview report generation method of the embodiment of the present application improves the intelligence, avoids the time and effort of personnel, and improves the efficiency.

[0057] After obtaining the interview report corresponding to the interview process in the above step 111, the method of the embodiment of the present application can further include:

[0058] Step 113: providing the interview report corresponding to the interview process in the form of an online editable document to the management personnel, so that the management personnel can confirm the key points of the interview content in the interview report, the management personnel prints the interview report after confirmation, and the management personnel signs in the signature column of the interview report.

[0059] Through the processing of the above step 113, the generated interview content is formed into an online editable document. The staff will carefully check and modify this document to ensure the accuracy and completeness of its content. After confirming that there is no error, they can download this interview report and directly use it for subsequent work and decision-making.

[0060] According to Figure 1 As shown in the flowchart, in the embodiment of the present application, a large model needs to be pre-trained.

[0061] In an embodiment of the present application, the training process of the large model includes: setting in-set words, inputting text content including the in-set words and labels of the in-set words into the large model, so as to train the large model to recognize the content of the interviewee information column; wherein the in-set words include at least one of name, gender, age, position, address and mobile phone number.

[0062] For the content of the interviewee information column, it is relatively fixed, and in the subsequent interview process, the staff can be required to "cooperate" with the speech recognition system and try to say "in-set words" such as name, gender, age, etc. For the conversation in the interview content column, first, a large number of conversation audios are collected as a speech data set to ensure the generalization ability of the large model; then, the speech data set is preprocessed, including removing silent segments, segmenting long audios, and distinguishing speech and non-speech, etc., to reduce noise and redundancy in the training process of the large model; finally, data enhancement and transfer learning techniques are used to train the large model. Through the training process, the large model can more accurately obtain various information required in the interviewee information column from the text content.

[0063] In another embodiment of the present application, the training method of the large model comprises:

[0064] Obtaining at least one interview training sample, each of which includes a segment of text-form interview content and an interview content gist as a label of the interview content; inputting the at least one interview training sample into the large model to train the large model to intelligently analyze the text-form interview content by using the deep learning and natural language processing capabilities of the large model and automatically generate the interview content gist.

[0065] In an embodiment of the present application, the interview content gist includes at least one of an inquiry question, a warning reminder, and a suggestion. For example, for step 111 described above, for the interview communication content such as inquiry questions, warning reminders, and suggestions, the large model can classify and extract interview topic information, simplify the interview content, and extract core content.

[0066] Figure 2 is a flowchart of the interview report generation method based on the large model in another embodiment of the present application. Referring to Figure 2 , the method comprises:

[0067] (1) Audio to text: First, advanced speech recognition technology is used to accurately convert the audio content in the interview process into text format.

[0068] (2) Extract keywords: Next, the converted text is analyzed in depth, and through professional keyword extraction technology, basic personnel information keywords (such as name, gender, department, etc.) and interview content summary keywords (such as interview focus, problem feedback, and suggestion measures, etc.) are accurately identified and classified.

[0069] (3) Fill in the information column: Then, the extracted basic personnel information keywords are directly filled into the interviewer information column of the interview report template.

[0070] (4) Generate interview content: Using the powerful document generation capability, the interview content summary keywords are logically integrated and language optimized to generate a segment of clear content and standard format interview content.

[0071] (5) Fill in the interview content column: The generated interview content is accurately filled into the interview content column of the interview report template.

[0072] (6) Perfect and print the report: Finally, the staff carefully checks and perfects the generated interview report to ensure the accuracy and completeness of the report content. After confirming that there is no error, the report is printed out for subsequent work and decision-making.

[0073] According to Figure 1 , Figure 2As can be seen from the process, the present invention significantly reduces the workload of staff by efficiently converting the content of the conversation into an interview report. Specifically, this process utilizes advanced speech recognition, keyword extraction of large models, and official document generation technologies to automatically convert the recorded content of the interview process into text form. Subsequently, through intelligent analysis and processing, key information is extracted and automatically filled into the interview report template to generate an interview report with a complete structure and accurate content. In this way, staff no longer need to manually organize the recording content and write reports, which greatly saves time and energy and achieves a significant improvement in work efficiency.

[0074] One embodiment of the present invention proposes a device for generating an interview report based on a large model, see Figure 3 , the device comprises:

[0075] The interview report template generation module 301 is configured to pre-set the interview report template to include an interviewer information column, an interview content column, and a signature column, and to set corresponding labels for each of the interviewer information column, the interview content column, and the signature column in the interview report template; to pre-set predetermined filling spaces in the interviewer information column and the interview content column of the interview report template, and to fill the reserved filling spaces in the interview report template using variables;

[0076] The to-be-recognized content acquisition module 302 is configured to acquire the voice content recorded during the interview process;

[0077] The conversion module 303 is configured to perform speech recognition on the speech content obtained by the recognition content acquisition module 302 to obtain the textual content of the interview, and input the textual content of the interview into the pre-trained large model;

[0078] The interview report generation module 304 is configured to use the basic information of the personnel analyzed and output by the large model as the value of the variable reserved in the interviewer information column, and use the label of the interviewer information column to fill the value of the variable into the interviewer information column in the interview report template; use the key points of the interview content analyzed and output by the large model as the value of the variable reserved in the interview content column, and use the label of the interview content column to fill the value of the variable into the interview content column in the interview report template, thereby obtaining an interview report corresponding to the interview process.

[0079] In one embodiment of the apparatus of the present invention, the apparatus further comprises a docking module 401;

[0080] The docking module 401 is configured to provide the interview report obtained by the interview report generation module 304 to the manager in the form of an online editable document, so that the manager confirms the interview content points in the interview report, the manager prints the interview report after confirmation, and the manager performs signature processing in the signature column of the interview report.

[0081] In an embodiment of the device, the device further comprises a large model training module.

[0082] The large model training module is configured to perform the following: setting an in-set word, inputting text content including the in-set word and a label of the in-set word into the large model, so as to train the large model to recognize the content of the interviewee information column; wherein the in-set word includes at least one of a name, a gender, an age, a position, an address, and a mobile phone number.

[0083] And / or,

[0084] The large model training module is configured to perform the following: obtaining at least one conversation training sample, each conversation training sample including a piece of conversation content in the form of text and an interview content point as a label of the conversation content; inputting the at least one conversation training sample into the large model, so as to train the large model to intelligently analyze the conversation content in the form of text by using the deep learning and natural language processing capabilities of the large model, and automatically generate the interview content point.

[0085] It should be noted that each of the above devices is usually implemented on a server side, and can be arranged on an independent server, or a combination of part or all of the devices can be arranged on the same server. The server can be a single server or a server cluster composed of multiple servers. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system. Each of the above devices can also be implemented on a computer terminal with strong computing power.

[0086] An embodiment of the present application provides a computer readable storage medium, which stores a computer program, and when the computer program is executed in a computer, the computer executes the method in any one of the embodiments of the specification.

[0087] An embodiment of the present application provides a computing device, comprising a memory and a processor, the memory stores executable code, and when the processor executes the executable code, the method in any one of the embodiments of the specification is implemented.

[0088] It can be understood that the structural schematic of the embodiments of the present application does not constitute a specific limitation on the device of the embodiments of the present application. In other embodiments of the specification, the above-mentioned device can include more or less components than the schematic, or combine certain components, or split certain components, or different component arrangement. The components shown can be implemented in hardware, software or a combination of software and hardware.

[0089] Each of the embodiments of the present application is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiments.

[0090] Those skilled in the art should be aware that functions described in the above one or more examples can be realized by hardware, software, a plug-in or any combination thereof. When realized by software, the functions can be stored in a computer readable medium or transmitted as one or more instructions or codes on a computer readable medium.

[0091] The above specific embodiments further detail the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made on the basis of the technical solutions of the present application shall be included in the protection scope of the present application.

Claims

1. A method for generating an interview report based on a large model, characterized in that: The method includes: The interview report template is pre-set to include an interviewer information column, an interview content column, and a signature column, and corresponding labels are set for each of the interviewer information column, the interview content column, and the signature column in the interview report template; predetermined filling positions are reserved in the interviewer information column and the interview content column of the interview report template, and the reserved filling positions in the interview report template are filled in through variables; Access the audio content recorded during the interview; Perform voice recognition on the acquired voice content to obtain the interview content in text form; Input the text-based interview content into the pre-trained large model; The basic information of the personnel analyzed and output by the large model is used as the value of the variable reserved in the interviewer information column. The value of the variable is filled into the interviewer information column in the interview report template using the label of the interviewer information column. The key points of the interview content analyzed and output by the large model are used as the value of the variable reserved in the interview content column. The label of the interview content column is used to fill the value of the variable into the interview content column in the interview report template, thereby obtaining an interview report corresponding to the interview process.

2. The method according to claim 1, characterized in that After obtaining the interview report corresponding to the interview process, the method further includes: The interview report corresponding to the interview process is provided to the manager in the form of an online editable document so that the manager can confirm the key points of the interview content in the interview report. After confirmation, the manager prints the interview report and signs in the signature column of the interview report.

3. The method according to claim 1, characterized in that The training method of the large model includes: Set the words in the set, and input the text content including the words in the set and the labels of the words in the set into the large model to train the large model's ability to recognize the content of the interviewer information column; wherein the words in the set include at least one of name, gender, age, position, address, and mobile phone number.

4. The method according to claim 1, wherein The training method of the large model includes: obtaining at least one conversation training sample, each conversation training sample including a text-based conversation content and interview content key points as labels for the conversation content; inputting the at least one conversation training sample into the large model to train the large model to use the deep learning and natural language processing capabilities of the large model to intelligently analyze the text-based conversation content and automatically generate interview content key points.

5. The method according to claim 1, wherein The main points of the interview content include: asking questions, warnings, and making suggestions.

6. The device for generating interview reports based on a large model is characterized in that: The device includes: The interview report template generation module is configured to pre-set the interview report template to include an interviewer information column, an interview content column, and a signature column, and set corresponding labels for each of the interviewer information column, the interview content column, and the signature column in the interview report template; pre-reserved filling positions are reserved in the interviewer information column and the interview content column of the interview report template, and the reserved filling positions in the interview report template are filled in through variables; A module for obtaining content to be recognized is configured to obtain the voice content recorded during the interview process; A conversion module is configured to perform speech recognition on the acquired speech content to obtain the interview content in text form, and input the text form of the interview content into a pre-trained large model; The interview report generation module is configured to use the basic information of the personnel analyzed and output by the large model as the value of the variable reserved in the interviewer information column, and use the label of the interviewer information column to fill the value of the variable into the interviewer information column in the interview report template; use the key points of the interview content analyzed and output by the large model as the value of the variable reserved in the interview content column, and use the label of the interview content column to fill the value of the variable into the interview content column in the interview report template, thereby obtaining an interview report corresponding to the interview process.

7. The device according to claim 6, characterized in that The apparatus further includes a docking module; The docking module is configured to provide the interview report obtained by the interview report generation module to the management personnel in the form of an online editable document, so that the management personnel can confirm the key points of the interview content in the interview report, print the interview report after confirmation, and sign the interview report in the signature column.

8. The device according to claim 6, characterized in that The device further includes: a large model training module; The large model training module is configured to execute: Setting a set of words, and inputting text content including the words and labels of the words into the large model to train the large model's ability to recognize the content of the interviewee information column; wherein the words in the set include at least one of name, gender, age, position, address, and mobile phone number; and / or, At least one conversation training sample is obtained, each conversation training sample including a textual conversation content and interview key points as labels for the conversation content; the at least one conversation training sample is input into the large model to train the large model to utilize the deep learning and natural language processing capabilities of the large model to intelligently analyze the textual conversation content and automatically generate interview key points.

9. The device according to claim 6, characterized in that The main points of the interview content include: asking questions, warnings, and making suggestions.

10. A computing device comprising a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method according to any one of claims 1 to 5 is implemented.

Images