Information collection method and device, server, medium and computer program product

Abstract

The embodiment of the invention relates to an information collection method and device, a server, a medium and a computer program product. The method comprises the following steps: receiving at least two rounds of response messages sent by a terminal; the response message is input into the terminal by the user according to an investigation question sent to the terminal by the server; summarizing and processing the at least two rounds of response messages by using a preset deep learning model to obtain summarized information of the user response, and sending the summarized information to the terminal; and receiving feedback information for the summary information sent by the terminal, and obtaining target information required by information acquisition according to the feedback information, the summary information and the at least two rounds of response messages. By adopting the method, the accuracy of the information collected when the user is investigated and researched can be improved.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of artificial intelligence technology, and in particular, to an information collection method, device, server, medium, and computer program product. Background Art

[0002] In order to understand the user experience and improve product performance, it is necessary to survey users. Traditional survey methods, such as questionnaires and interviews, have problems such as low efficiency, small survey scope, and unstable data quality. Therefore, using chatbots to survey users as a survey tool has become the current trend in user surveys.

[0003] Traditional chatbots present survey questions from a preset database to users and use the information entered by users as survey results.

[0004] However, traditional chatbots have the problem of low accuracy of the information collected. Summary of the invention

[0005] The embodiments of the present disclosure provide an information collection method, device, server, medium and computer program product, which can be used to improve the accuracy of information collected when conducting surveys on users.

[0006] In a first aspect, an embodiment of the present disclosure provides an information collection method, the method comprising:

[0007] Receiving at least two rounds of response messages sent by the terminal; the response messages are input by the user into the terminal according to the survey questions sent to the terminal by the server;

[0008] Summarizing at least two rounds of the response messages using a preset deep learning model to obtain summary information of the user's responses, and sending the summary information to the terminal;

[0009] receiving feedback information for the summary information sent by the terminal, and obtaining target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages.

[0010] In a second aspect, an embodiment of the present disclosure provides an information collection device, the device comprising:

[0011] A receiving module, used for receiving at least two rounds of response messages sent by the terminal; the response messages are input by the user into the terminal according to the survey questions sent to the terminal by the server;

[0012] A processing module, used to summarize and process at least two rounds of the response messages using a preset deep learning model, obtain summary information of the user's response, and send the summary information to the terminal;

[0013] The acquisition module is used to receive feedback information sent by the terminal for the summary information, and obtain target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages.

[0014] In a third aspect, an embodiment of the present disclosure provides a server, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the first aspect when executing the computer program.

[0015] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in the first aspect above.

[0016] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program, which implements the method described in the first aspect when executed by a processor.

[0017] The information collection method, device, server, medium and computer program product provided by the embodiments of the present disclosure include: the server sends at least two rounds of survey questions to the terminal, the user inputs a response message into the terminal according to the survey questions sent by the server to the terminal, the terminal sends at least two rounds of response messages to the server, the server summarizes and processes the at least two rounds of response messages using a preset deep learning model to obtain summary information of the user's responses, and sends the summary information to the terminal; after the server receives feedback information for the summary information sent by the terminal, it obtains target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages, determines the target information in combination with the feedback information, thereby improving the accuracy of the obtained target information, and summarizes and processes the at least two rounds of response messages using a deep learning model, so that the obtained summary information is relatively accurate, thereby improving the accuracy of the feedback information, thereby further improving the accuracy of the target information. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 An application environment diagram of an information collection method in an embodiment;

[0019] Figure 2 is a flowchart of an information collection method in one embodiment;

[0020] Figure 3 is a flowchart of an information collection method in another embodiment;

[0021] Figure 4 is a flowchart of an information collection method in another embodiment;

[0022] Figure 5is a flowchart of an information collection method in another embodiment;

[0023] Figure 6 is a flowchart of an information collection method in another embodiment;

[0024] Figure 7 is a flowchart of an information collection method in another embodiment;

[0025] Figure 8 is a structural block diagram of an information collection device in one embodiment;

[0026] Fig. 9 FIG. 4 is a diagram showing the internal structure of an electronic device in one embodiment. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure more clear, the embodiments of the present disclosure are further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present disclosure and are not used to limit the embodiments of the present disclosure.

[0028] First, before specifically introducing the technical solution of the embodiment of the present disclosure, the technical background or technical evolution context based on the embodiment of the present disclosure is introduced. Generally, when conducting a survey on users, for example, conducting a survey on online car-hailing drivers, conducting a survey on shared bicycle users, etc., the current technical background is: traditional questionnaires or inquiry methods have problems such as low efficiency, small survey scope, and unstable data quality. For example, when using traditional questionnaires for surveys, the survey questions are relatively simple, and the collected user information has low accuracy; or, when using the inquiry method for surveys, one-to-one inquiries are usually conducted, resulting in low survey efficiency. Therefore, a chat robot is introduced for surveying users. The existing chat robot usually displays the survey questions in the preset database to the user in sequence through an interactive interface. Specifically, after the user enters a response message for the current survey question, the chat robot displays the next survey question in the preset database to the user through the interactive interface until the user enters the response message of the survey question. However, the data collected by the existing chat robot still has the problem of low accuracy, and the chat robot in the prior art has low flexibility in responding to users, and the user experience is poor. How to improve the accuracy of information collected when conducting user surveys has become a difficult problem that needs to be solved urgently.

[0029] The technical solutions involved in the embodiments of the present disclosure are introduced below in combination with the scenarios to which the embodiments of the present disclosure are applied.

[0030] The information collection method provided by the embodiment of the present disclosure can be applied to Figure 1In the application environment shown. The terminal 102 communicates with the server 104 through the network. The server 104 receives at least two rounds of response messages sent by the terminal 102, sends summary information to the terminal 102, and receives feedback information for the summary information sent by the terminal 102. The terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices, and the server 104 can be implemented by an independent server or a server cluster consisting of multiple servers.

[0031] In one embodiment, Figure 2 As shown, a method for collecting information is provided, which is applied to Figure 1 The server in is used as an example to illustrate, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0032] S201, receiving at least two rounds of response messages sent by the terminal; the response messages are input by the user into the terminal according to the survey questions sent to the terminal by the server.

[0033] In an embodiment of the present application, the server sends at least two rounds of survey questions to the terminal. Each time a round of survey questions is sent, the terminal returns to the server a response message input by the user in the terminal according to the survey questions of that round, and the response message corresponds to the survey question one by one.

[0034] For example, taking the example of the server receiving two rounds of response messages sent by the terminal, the server sends the first round of survey questions to the terminal, the terminal displays the first round of survey questions to the user through the interactive interface, the user inputs the first round of response messages corresponding to the first round of survey questions into the terminal through the interactive interface, and the terminal sends the first round of response messages to the server; after receiving the first round of response messages, the server sends the second round of survey questions to the terminal, the terminal displays the second round of survey questions to the user through the interactive interface, the user inputs the second round of response messages corresponding to the second round of survey questions into the terminal through the interactive interface, and the terminal sends the second round of response messages to the server. The interactive interface may be in the application of the terminal, or the interactive interface may be a World Wide Web (web) page.

[0035] Optionally, the user can enter text information or voice information in the terminal's interactive interface. If the user enters text information in the interactive interface, the terminal sends the text information as a response message to the server; if the user enters voice information in the interactive interface, the terminal converts the voice information into text information through voice recognition technology, and then sends the text information as a response message to the server.

[0036] S202: Use a preset deep learning model to summarize at least two rounds of response messages to obtain summary information of user responses, and send the summary information to the terminal.

[0037] In an embodiment of the present application, the preset deep learning model is a pre-trained model that can summarize and process information. The deep learning model can extract features from at least two rounds of response messages, obtain feature information of at least two rounds of response messages, and then generate coherent sentences based on the feature information, thereby obtaining summary information of the user's response, and sending the summary information to the terminal.

[0038] Optionally, the deep learning model can be a large language model LLM. The LLM model is a deep learning model trained using a large amount of text data. It can generate natural language text or understand the meaning of language text, and can handle a variety of natural language tasks, such as text classification, question and answer, and dialogue.

[0039] S203, receiving feedback information for the summary information sent by the terminal, and obtaining target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages.

[0040] The feedback information on the summary information is used to indicate whether the user determines whether the summary information is correct.

[0041] In an embodiment of the present application, after the terminal receives the summary information sent by the server, the summary information is displayed to the user through an interactive interface. The terminal can obtain feedback information for the summary information through user clicks or inputs, and send the feedback information to the server. Furthermore, the server obtains the target information required for information collection based on the feedback information, the summary information and at least two rounds of response messages.

[0042] Optionally, feature extraction may be performed on the feedback information, summary information and at least two rounds of response messages respectively to obtain feature information corresponding to the feedback information, feature information corresponding to the summary information and feature information corresponding to at least two rounds of response messages, and data cleaning and other processing may be performed on each feature information to obtain processed feature information, and the processed feature information may be used as the target information required for information collection.

[0043] In the above information collection method, the server sends at least two rounds of survey questions to the terminal, the user inputs the answer message into the terminal according to the survey questions sent by the server to the terminal, the terminal sends at least two rounds of answer messages to the server, the server summarizes and processes the at least two rounds of answer messages using a preset deep learning model, obtains summary information of the user's answers, and sends the summary information to the terminal. After the server receives the feedback information sent by the terminal for the summary information, it obtains the target information required for information collection according to the feedback information, the summary information and at least two rounds of answer messages, determines the target information in combination with the feedback information, thereby improving the accuracy of the obtained target information, and uses the deep learning model to summarize and process at least two rounds of answer messages, so that the obtained summary information is more accurate, thereby improving the accuracy of the feedback information, thereby further improving the accuracy of the target information.

[0044] In the above scenario of receiving at least two rounds of response messages sent by the terminal, the deep learning model generates the next round of survey questions based on the response messages returned by the terminal. In one embodiment, the deep learning model includes a dynamic questioning model, such as Figure 3 As shown, the above S201 includes:

[0045] S301, sending a survey question to a terminal, and receiving a response message returned by the terminal.

[0046] In an embodiment of the present application, the server sends survey questions to the terminal. The survey questions may be first-round survey questions or survey questions of other rounds. The terminal collects response messages input by the user according to the survey questions and sends the response messages to the server.

[0047] For example, in the scenario of surveying online ride-hailing drivers, the survey question may be "Are you satisfied with the current rating mechanism for online ride-hailing drivers?"; in the scenario of surveying shared bicycle users, the survey question may be "Are you satisfied with the comfort of the shared bicycle you used in your most recent riding experience?"

[0048] S302, generating the next round of survey questions according to the response message through the dynamic questioning model, and sending the next round of survey questions to the terminal.

[0049] Among them, the dynamic question model can generate natural language text based on known data.

[0050] In an embodiment of the present application, the dynamic questioning model can analyze the response message to understand the meaning of the response message, and further generate the next round of survey questions based on the meaning of the response message. The server sends the next round of survey questions to the terminal.

[0051] For example, in a scenario where a survey is conducted on users of shared bicycles, the survey question may be "Are you satisfied with the comfort of the shared bicycle you used in your most recent riding experience?" If the response message is "Not satisfied," the next round of survey questions may be "Which part of the shared bicycle do you think affects the comfort?" If the response message is "The seat is poorly comfortable," the next round of survey questions may be "Which specific aspect of the seat affects the comfort? For example, the seat height or the softness or hardness of the seat."

[0052] S303, receiving a next round of response message returned by the terminal in response to the next round of survey questions.

[0053] In an embodiment of the present application, the user can input a next round of response messages for the next round of survey questions in the interactive page of the terminal, and the terminal sends the next round of response messages to the server.

[0054] Optionally, each survey process includes multiple groups of survey questions and answers, and each group of survey questions and answers includes at least two rounds of survey questions and response messages. The deep learning model can determine whether to end this group of survey questions and answers based on the meaning of each round of response messages. If this group of survey questions and answers is ended, the deep learning model is used to summarize and process at least two rounds of response messages to obtain summary information of user responses, and send the summary information to the terminal; if the survey questions and answers are not ended, the next round of survey questions will continue to be generated based on the previous round of response messages.

[0055] In the above embodiment, the dynamic questioning model generates the next round of survey questions based on the response message. Compared with determining the survey questions from a preset question library in the prior art, the survey questions generated by the dynamic questioning model are more closely matched to the user and have higher accuracy, thereby improving the accuracy of the target information.

[0056] In the above scenario where the dynamic questioning model generates the next round of survey questions based on the answer message and sends the next round of survey questions to the terminal, the dynamic questioning model generates the speech information and the next round of survey questions based on the answer message. Figure 4 As shown, the above S302 includes:

[0057] S401, generating speech information and next round of survey questions according to the response message through a dynamic questioning model.

[0058] Among them, the speech information is used to provide feedback on the response message.

[0059] In an embodiment of the present application, the dynamic questioning model can analyze the response message to understand the meaning of the response message, generate corresponding speech information for the response message based on the meaning of the response message, and generate the next round of survey questions based on the meaning of the response message.

[0060] For example, when the survey question in the previous round is "How important is it to you that artificial intelligence systems prioritize safety and do not cause physical or mental harm?", if the response message to the survey question is "This is important", the speech information generated by the dynamic questioning model may be "I understand that safety is very important to you", and the next round of survey questions generated by the dynamic questioning model may be "When it comes to artificial intelligence systems, can you provide any specific examples of emotional or physical harm that you are worried about?"

[0061] S402, sending the speech information and the next round of survey questions to the terminal.

[0062] In an embodiment of the present application, the server sends the speech information and the next round of survey questions generated by the dynamic questioning model to the terminal, and the terminal displays the speech information and the next round of survey questions to the user through an interactive interface.

[0063] In the above embodiment, the dynamic questioning model generates speech information and the next round of survey questions, and sends the speech information and the next round of survey questions to the terminal, and uses the speech information to provide feedback on the user's response information in the previous round, thereby improving the user's sense of use when being surveyed.

[0064] In the above scenario of sending survey questions to the terminal and receiving the response message returned by the terminal, the first round of survey questions are determined according to a preset database. Figure 5 As shown, the above S301 includes:

[0065] S501, determining the first round of research questions according to a preset database; the database includes a plurality of alternative research questions.

[0066] In the embodiment of the present application, the researcher can form multiple candidate survey questions in advance according to the survey scenario, historical survey data, surveyed users and other parameters, store the multiple candidate survey questions in the server, and the server encapsulates the multiple candidate survey questions to form a database. The server determines the first round of survey questions in the database. For example, the server can use the candidate survey questions as the first round of survey questions for each group of survey questions and answers in turn according to the order of the candidate survey questions in the database.

[0067] S502, sending the first round of survey questions to the terminal.

[0068] In an embodiment of the present application, the server sends the first round of survey questions to the terminal, so that the terminal can display the first round of survey questions to the user through an interactive interface.

[0069] In the above embodiment, the first round of survey questions are determined based on a preset database, and the research angles of the next round of survey questions can be determined based on the first round of survey questions, so that when the deep learning model generates the next round of survey questions based on the response messages corresponding to the determined survey questions, it is more in line with the purpose of this survey, thereby making the generated next round of survey questions more accurate.

[0070] The following will describe in detail the process of summarizing and processing at least two rounds of response messages using a preset deep learning model to obtain summary information of the user's response. In one embodiment, the deep learning model also includes an information checking model, and the above S202 includes: inputting at least two rounds of response messages into the information checking model for summary processing, and obtaining summary information of the user's response output by the information checking model.

[0071] In an embodiment of the present application, the preset deep learning model includes an information checking model, which takes at least two rounds of response messages as input to the information checking model. The information checking model analyzes and processes the at least two rounds of response messages, extracts key information from the at least two rounds of response messages, and generates summary information that can be understood by the user based on the key information.

[0072] Optionally, the summary information may also include inquiry information, for example, the inquiry information may be “Do you agree with this view?”

[0073] In the above embodiment, the information checking model generates summary information according to at least two rounds of response messages. The summary information generated by the information checking model has high accuracy. Applying the summary information in the survey process can improve the accuracy of the target information.

[0074] The process of obtaining target information required for information collection based on feedback information, summary information and at least two rounds of response messages is described in detail below. In one embodiment, the above S203 includes: if the feedback information indicates that the user determines that the summary information is correct, obtaining target information based on the summary information and at least two rounds of response messages.

[0075] In the embodiment of the present application, the terminal sends the feedback information input by the user to the server, and when the feedback information indicates that the user determines that the summary information is correct, the target message is obtained according to the summary information and at least two rounds of response messages. Optionally, the summary information and at least two rounds of response messages can be determined as the target message; or the summary information and at least two rounds of response messages can be merged and the merged information can be determined as the target information.

[0076] Exemplarily, when the feedback information indicates that the user has determined that the summary information is correct, the feedback information may be information such as agreement, confirmation, or correctness.

[0077] In another embodiment, the above S203 includes: if the feedback information indicates that the user determines that the summary information is wrong, generating the next round of survey questions based on at least two rounds of response messages and the deep learning model, and sending the next round of survey questions to the terminal.

[0078] In an embodiment of the present application, if the feedback information indicates that the user has determined that the summary information is incorrect, it means that the accuracy of at least two rounds of response messages collected by the server is low, and it is necessary to continue to survey the user. Therefore, based on at least two rounds of response messages and the deep learning model, the next round of survey questions is generated, and the next round of survey questions is sent to the terminal to obtain the user's response message to the next round of survey questions until the user confirms that the summary information is correct.

[0079] Exemplarily, when the feedback information indicates that the user determines that the summary information is wrong, the feedback information may be information such as disagreement, error, inaccuracy, etc.

[0080] In the above embodiment, when the user confirms that the summary information is correct, the target information is obtained based on the existing summary information and at least two rounds of response messages. When the user confirms that the summary information is wrong, survey questions are continued to be generated to ask the user to obtain more information, thereby improving the accuracy of the summary information generated next time.

[0081] The training process of the deep learning model will be described in detail below. Figure 6 As shown, the above method also includes:

[0082] S601, obtaining survey data of a historical period.

[0083] The survey data for the historical period includes questionnaires and / or interview survey data for the historical period.

[0084] In an embodiment of the present application, the survey data of the historical period may be pre-stored in the server, or the server may send a data acquisition request to other electronic devices so that the other electronic devices send the survey data of the historical period to the server.

[0085] S602, using the survey data of the historical period to supervise the initial learning model and obtain a deep learning model.

[0086] In an embodiment of the present application, survey data of a historical period is used as sample data of an initial learning model, and the initial learning model is supervised and trained using the sample data to obtain a trained deep learning model, so that the sentence form in the text information generated by the deep learning model is in a survey form.

[0087] Optionally, the initial learning model is an LLM model, and the LLM model is fine-tuned and trained using survey data from a historical period to obtain a trained LLM model, so that the trained LLM model better matches the survey scenario.

[0088] In the above embodiment, the initial learning model is trained based on the survey data of the historical period to obtain a deep learning model, so that the survey questions and summary information generated by the deep learning model are more accurate.

[0089] After obtaining the target information, the target information is quantitatively analyzed and / or qualitatively analyzed. Figure 7 As shown, the above method also includes:

[0090] S701, performing quantitative analysis on the target information to obtain the quantitative analysis result of the target information.

[0091] In the embodiment of the present application, the target information can be quantitatively analyzed using a preset model or function. For example, if the target information includes numerical data, such as the daily working hours of online car-hailing drivers, the preset function can be used to calculate the average daily working hours of multiple online car-hailing drivers in the target information to obtain the average daily working hours of online car-hailing drivers; or, if the target information includes words such as satisfied and correct that can express the user's subjective emotions, the deep learning model can be used to obtain the keywords in the target information.

[0092] S702, performing qualitative analysis on the target information to obtain a qualitative analysis result of the target information.

[0093] In the embodiment of the present application, a preset model or method can be used to perform a qualitative analysis on the target information to obtain a qualitative analysis result of the target information.

[0094] Exemplarily, when the target information includes user suggestions on the use of shared bicycles, the target information can be processed by data cleaning, word segmentation, etc. to obtain processed target information. Furthermore, feature extraction is performed on the processed target information using a preset feature extraction model to obtain key information in the user's suggestions on the use of shared bicycles.

[0095] It should be noted that the quantitative analysis and qualitative analysis of the target information in the above embodiment is one scenario of analyzing the target information. In other scenarios, only the quantitative analysis or only the qualitative analysis may be performed on the target information.

[0096] In the above embodiments, the target information is subjected to quantitative analysis and / or qualitative analysis, so that the user needs can be understood and improvement suggestions can be put forward according to the quantitative analysis results and / or qualitative analysis results.

[0097] An embodiment of the present disclosure is introduced below in conjunction with a specific information collection scenario. The method includes the following steps:

[0098] S1, obtaining survey data of historical periods; using the survey data of historical periods to supervise the initial learning model and obtain a deep learning model; wherein the deep learning model includes a dynamic questioning model and an information checking model.

[0099] S2, sending survey questions to the terminal and receiving a response message returned by the terminal; the response message is input by the user into the terminal according to the survey questions sent to the terminal by the server; wherein, when the survey questions are the first round of survey questions, the first round of survey questions are determined according to a preset database; the database includes multiple alternative survey questions; the first round of survey questions are sent to the terminal.

[0100] S3, generating speech information and the next round of survey questions according to the response message through the dynamic questioning model; sending the speech information and the next round of survey questions to the terminal.

[0101] S4, receiving the next round of response messages returned by the terminal in response to the next round of survey questions.

[0102] S5, inputting at least two rounds of response messages into the information checking model for summary processing, obtaining summary information of the user responses output by the information checking model, and sending the summary information to the terminal.

[0103] S6, receiving feedback information regarding the summary information sent by the terminal. If the feedback information indicates that the user determines that the summary information is correct, obtaining the target information based on the summary information and at least two rounds of response messages; if the feedback information indicates that the user determines that the summary information is wrong, generating the next round of survey questions based on at least two rounds of response messages and a deep learning model, and sending the next round of survey questions to the terminal.

[0104] S7, performing quantitative analysis on the target information to obtain a quantitative analysis result of the target information; and / or performing qualitative analysis on the target information to obtain a qualitative analysis result of the target information.

[0105] In the above information collection method, the server sends at least two rounds of survey questions to the terminal, the user inputs the answer message into the terminal according to the survey questions sent by the server to the terminal, the terminal sends at least two rounds of answer messages to the server, the server summarizes and processes the at least two rounds of answer messages using a preset deep learning model, obtains summary information of the user's answers, and sends the summary information to the terminal. After the server receives the feedback information sent by the terminal for the summary information, it obtains the target information required for information collection according to the feedback information, the summary information and at least two rounds of answer messages, determines the target information in combination with the feedback information, thereby improving the accuracy of the obtained target information, and uses the deep learning model to summarize and process at least two rounds of answer messages, so that the obtained summary information is more accurate, thereby improving the accuracy of the feedback information, thereby further improving the accuracy of the target information.

[0106] It should be understood that although Figure 2-7 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-7 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0107] In one embodiment, Figure 8 As shown, an information collection device is provided, comprising: a receiving module 11, a processing module 12 and a first acquisition module 13, wherein:

[0108] The receiving module 11 is used to receive at least two rounds of response messages sent by the terminal; the response messages are input by the user into the terminal according to the survey questions sent to the terminal by the server.

[0109] The processing module 12 is used to summarize at least two rounds of response messages using a preset deep learning model, obtain summary information of the user's response, and send the summary information to the terminal.

[0110] The first acquisition module 13 is used to receive feedback information for the summary information sent by the terminal, and obtain target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages.

[0111] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0112] In one embodiment, the receiving module 11 includes: a sending unit, a generating unit and a receiving unit, wherein:

[0113] The sending unit is used to send survey questions to the terminal and receive response messages returned by the terminal.

[0114] The generating unit is used to generate the next round of survey questions according to the answer message through the dynamic questioning model, and send the next round of survey questions to the terminal.

[0115] The receiving unit is used to receive the next round of response messages returned by the terminal for the next round of survey questions.

[0116] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0117] In one embodiment, the generation unit is used to generate speech information and next round of survey questions according to the response message through a dynamic questioning model; and send the speech information and next round of survey questions to the terminal.

[0118] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0119] In one embodiment, the sending unit is used to determine the first round of survey questions according to a preset database; the database includes a plurality of alternative survey questions; and send the first round of survey questions to the terminal.

[0120] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0121] In one embodiment, the processing module 12 includes: a summarizing unit, which is used to input at least two rounds of response messages into the information checking model for summarizing processing, and obtain summary information of the user response output by the information checking model.

[0122] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0123] In one embodiment, the first acquisition module 13 includes: a first acquisition unit, configured to obtain target information according to the summary information and at least two rounds of response messages if the feedback information indicates that the user determines that the summary information is correct.

[0124] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0125] In one embodiment, the first acquisition module 13 includes: a second acquisition unit, which is used to generate the next round of survey questions based on at least two rounds of response messages and a deep learning model if the feedback information indicates that the user determines that the summary information is wrong, and send the next round of survey questions to the terminal.

[0126] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0127] In one embodiment, the above device further includes: a second acquisition module and a training module, wherein:

[0128] The second acquisition module is used to obtain survey data for historical periods.

[0129] The training module is used to use historical survey data to supervise the initial learning model and obtain a deep learning model.

[0130] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0131] In one embodiment, the above device further includes: a first analysis module and a second analysis module, wherein:

[0132] The first analysis module is used to perform quantitative analysis on the target information to obtain the quantitative analysis result of the target information.

[0133] The second analysis module is used to perform qualitative analysis on the target information to obtain qualitative analysis results of the target information.

[0134] The information collection device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effects are similar, which will not be repeated here.

[0135] For the specific definition of the information collection device, please refer to the definition of the information collection method above, which will not be repeated here. Each module in the above-mentioned information collection device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the electronic device in the form of hardware, or can be stored in the memory of the electronic device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0136] Fig. 9 FIG. 1 is a block diagram of a server 1400 according to an exemplary embodiment. Fig. 9The server 1400 includes a processing component 1420, which further includes one or more processors, and a memory resource represented by a memory 1422 for storing instructions or computer programs, such as applications, that can be executed by the processing component 1420. The application stored in the memory 1422 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 1420 is configured to execute instructions to perform the above-mentioned information collection method.

[0137] The server 1400 may also include a power supply component 1424 configured to perform power management of the device 1400, a wired or wireless network interface 1426 configured to connect the device 1400 to a network, and an input / output (I / O) interface 1428. The server 1400 may operate based on an operating system stored in the memory 1422, such as Windows 14 14erverTM, Mac O14 XTM, UnixTM, LinuxTM, FreeB14DTM or the like.

[0138] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 1304 including instructions, and the instructions can be executed by a processor 1320 of the electronic device 1300 to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0139] In an exemplary embodiment, a storage medium including instructions is also provided, such as a memory 1422 including instructions, which can be executed by a processor of the server 1400 to perform the above method. The storage medium can be a non-transitory computer-readable storage medium, for example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0140] In an exemplary embodiment, a computer program product is also provided, and when the computer program is executed by a processor, the above method can be implemented. The computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, part or all of the above method can be implemented in whole or in part according to the process or function described in the embodiment of the present disclosure.

[0141] It should be noted that the solutions described in this specification and the embodiments, if they involve the processing of personal information, will be processed on the premise of having a legal basis (such as obtaining the consent of the subject of personal information, or being necessary for the performance of a contract, etc.), and will only be processed within the scope of regulations or agreements. If a user refuses to process personal information other than the necessary information for basic functions, it will not affect the user's use of basic functions.

[0142] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the embodiments of the present disclosure can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0143] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] The above-described embodiments only express several implementation methods of the embodiments of the present disclosure, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the embodiments of the present disclosure, and these all belong to the protection scope of the embodiments of the present disclosure. Therefore, the protection scope of the patent of the embodiments of the present disclosure shall be subject to the attached claims.

Claims

1. An information collection method, characterized in that: The method comprises: Receiving at least two rounds of response messages sent by the terminal; the response messages are input by the user into the terminal according to the survey questions sent to the terminal by the server; Summarizing and processing at least two rounds of the response messages using a preset deep learning model to obtain summary information of the user's responses, and sending the summary information to the terminal; receiving feedback information for the summary information sent by the terminal, and obtaining target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages.

2. The method according to claim 1, characterized in that The deep learning model includes a dynamic questioning model, and the at least two rounds of response messages sent by the receiving terminal include: Sending a survey question to the terminal and receiving a response message returned by the terminal; Generate the next round of survey questions according to the response message through the dynamic questioning model, and send the next round of survey questions to the terminal; Receive a next round of response messages returned by the terminal in response to the next round of survey questions.

3. The method according to claim 2, characterized in that Generating the next round of survey questions according to the response message by using the dynamic questioning model, and sending the next round of survey questions to the terminal, includes: Generate speech information and the next round of survey questions according to the response message through the dynamic questioning model; The speech information and the next round of survey questions are sent to the terminal.

4. The method according to claim 2, characterized in that: The sending of the survey question to the terminal includes: Determining the first round of survey questions based on a preset database; the database includes a plurality of alternative survey questions; The first round of survey questions is sent to the terminal.

5. The method according to any one of claims 2 to 4, characterized in that: The deep learning model also includes an information checking model, and the method of summarizing and processing at least two rounds of the response messages using the preset deep learning model to obtain summary information of the user's response includes: At least two rounds of response messages are input into the information checking model for summary processing, and summary information of the user response output by the information checking model is obtained.

6. The method according to any one of claims 1 to 4, characterized in that: The obtaining target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages includes: If the feedback information indicates that the user determines that the summary information is correct, the target information is obtained according to the summary information and at least two rounds of response messages.

7. The method according to claim 6, characterized in that The method further comprises: If the feedback information indicates that the user determines that the summary information is wrong, a next round of survey questions is generated based on at least two rounds of response messages and the deep learning model, and the next round of survey questions is sent to the terminal.

8. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Obtain survey data for historical periods; The survey data of the historical period is used to perform supervised training on the initial learning model to obtain the deep learning model.

9. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Performing quantitative analysis on the target information to obtain a quantitative analysis result of the target information; And / or, performing a qualitative analysis on the target information to obtain a qualitative analysis result of the target information.

10. An information collection device, characterized in that: The device comprises: A receiving module, used for receiving at least two rounds of response messages sent by the terminal; the response messages are input by the user into the terminal according to the survey questions sent to the terminal by the server; A processing module, used to summarize and process at least two rounds of the response messages using a preset deep learning model, obtain summary information of the user's response, and send the summary information to the terminal; The acquisition module is used to receive feedback information sent by the terminal for the summary information, and obtain target information required for information collection according to the feedback information, the summary information and at least two rounds of response messages.

11. A server 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 computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

12. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

13. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

Images