Response method, device, equipment, medium and product

Abstract

The invention discloses a reply method and device, equipment, a medium and a product, and the method comprises the steps: after a question proposed by a user for a target object (such as a certain product) is obtained, the question is processed through a target model, a first answer is obtained, and the first answer is used for representing a result obtained through automatic reply for the question; the first answer is displayed to the user, so that after a feedback operation triggered by the user for the first answer is detected, if the feedback operation is used for indicating that the first answer is not matched with the question, it is determined that the answer determined by the automatic answering mode for the question is inaccurate, and the answer is not accurate. And the second answer obtained by manually answering the question is obtained and displayed, so that the answering cost can be reduced on the premise of ensuring the answering accuracy as much as possible through an automatic answering and manual correction mode.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and in particular, to a response method, apparatus, device, medium, and product. Background Art

[0002] A product refers to an object used to provide a certain service to a user, such as an application (App) or a physical device (such as a car), etc.

[0003] In addition, when a user uses a product, if the user encounters some problems, the user can feedback these problems to the relevant manager of the product in a certain way, so that the manager can feedback the answers to these problems to the user in a manual response manner. However, this manual response method has many defects, such as high cost. Summary of the Invention

[0004] To solve the above technical problems, this application provides a response method, apparatus, device, medium, and product, which is beneficial to reducing the response cost as much as possible.

[0005] To achieve the above object, the technical solutions provided in this application are as follows:

[0006] This application provides a response method, the method includes: in response to a question raised for a target object, using a target model to process the question to obtain a first answer; displaying the first answer; in response to a feedback operation triggered for the first answer indicating that the first answer does not match the question, obtaining and displaying a second answer obtained by manually responding to the question.

[0007] In a possible implementation manner, the question is raised by using a target tool, and the target tool is used to input the question into the target model; the displaying the first answer includes: in response to the target tool obtaining the first answer from the target model, displaying the first answer and a feedback component; the in response to a feedback operation triggered for the first answer indicating that the first answer does not match the question, obtaining and displaying a second answer obtained by manually responding to the question includes: in response to a feedback operation triggered through the feedback component indicating that the first answer does not match the question, obtaining and displaying a second answer obtained by manually responding to the question.

[0008] In a possible implementation manner, the problem is proposed through a session group of the target tool, and the session group is used to display at least one round of session, and the at least one round of session includes the problem; the feedback operation triggered by the feedback component is used to indicate that the first answer does not match the problem, and obtaining and displaying a second answer obtained by manually replying to the problem includes: in response to the feedback operation triggered by the feedback component being used to indicate that the first answer does not match the problem, adding the target login account of the target tool to the session group, where the target login account is used to manually reply to the problem, and displaying in the session group the second answer provided by the target login account for the problem.

[0009] In a possible implementation manner, the target tool is used to provide an instant messaging service.

[0010] In a possible implementation manner, the method further includes: in response to the second answer matching the problem, updating the target model according to the second answer and the problem.

[0011] In a possible implementation manner, the first answer is obtained by the target model performing answer prediction processing using a knowledge base, and the knowledge base includes the first answer; in response to the second answer matching the problem, updating the target model according to the second answer and the problem includes: in response to the second answer matching the problem, generating a question-and-answer pair according to the second answer and the problem; after adding the question-and-answer pair to the knowledge base, updating the target model according to the knowledge base.

[0012] In a possible implementation manner, the problem includes some or all of text, pictures, audio, and video.

[0013] The present application provides a reply device, including: a data processing unit, configured to, in response to a problem proposed for a target object, process the problem using a target model to obtain a first answer; a first display unit, configured to display the first answer; and a second display unit, configured to, in response to a feedback operation triggered for the first answer being used to indicate that the first answer does not match the problem, obtain and display a second answer obtained by manually replying to the problem.

[0014] The present application provides an electronic device, where the device includes: a processor and a memory; the memory is configured to store instructions or computer programs; and the processor is configured to execute the instructions or computer programs in the memory so that the electronic device executes the reply method provided by the present application.

[0015] The present application provides a computer-readable medium, in which instructions or a computer program are stored. When the instructions or the computer program run on a device, the device is caused to execute the response method provided by the present application.

[0016] The present application provides a computer program product, which includes a computer program carried on a non-transitory computer-readable medium. The computer program includes program code for executing the response method provided by the present application.

[0017] Compared with the related art, the present application has at least the following advantages:

[0018] In the response solution provided by the present application, after obtaining a question raised by a user for a target object (such as a certain product), the target model is first used to process the question to obtain a first answer, so that the first answer is used to represent the result obtained by automatically replying to the question; then the first answer is displayed to the user. After detecting a feedback operation triggered by the user for the first answer, if the feedback operation is used to indicate that the first answer does not match the question, it is determined that the answer determined by the automatic reply method for the question is inaccurate. Therefore, the second answer obtained by manually replying to the question is obtained and displayed. In this way, it is possible to reduce the reply cost on the premise of ensuring the reply accuracy as much as possible through the automatic reply + manual correction method. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the following will briefly introduce the drawings required for use in the description of the embodiments or the related art. Obviously, the drawings in the following description are only some embodiments recorded in the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a flowchart of a response solution provided by an embodiment of the present application;

[0021] Figure 2 It is a schematic diagram of a session group provided by an embodiment of the present application;

[0022] Figure 3 It is a schematic structural diagram of a response device provided by an embodiment of the present application;

[0023] Figure 4 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] It has been found through research that after a user raises a question about a certain product, the relevant manager of the product can give a manual reply to the question. Among them, because this reply method relies on manual work, there are defects such as high cost, low efficiency, and easy to give incorrect replies under the influence of some subjective factors.

[0025] Based on the above research, in order to overcome the above problems, the present application provides a reply method, which includes: after obtaining a question raised by a user for a target object (such as a certain product), first use a target model to process the question to obtain a first answer, so that the first answer is used to represent the result obtained by automatically replying to the question; then display the first answer to the user, so that after detecting a feedback operation triggered by the user for the first answer, if the feedback operation is used to indicate that the first answer does not match the question, it is determined that the answer determined by the automatic reply method for the question is inaccurate, so obtain and display a second answer obtained by manually replying to the question. In this way, the defects caused by manual reply, such as high cost, can be overcome as much as possible while ensuring the accuracy of the reply through the automatic reply + manual correction method.

[0026] In addition, the present application does not limit the execution entity of the reply method. For example, the method can be applied to a terminal device or a server. Another example is that the method can also be implemented by means of data interaction between a terminal device and a server. Among them, the terminal device can be a smart phone, a computer, a personal digital assistant (Personal Digital Assitant, PDA), a tablet computer, etc. The server can be an independent server, a cluster server or a cloud server.

[0027] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present application.

[0028] To better understand the technical solution provided by the present application, the reply method provided by the present application will be described below in conjunction with some drawings. As Figure 1 shown, the reply method provided by the embodiment of the present application includes S1-S3 below.

[0029] S1: In response to a question raised for a target object, use a target model to process the question to obtain a first answer.

[0030] Among them, the target object is a product for providing a certain service to users. It should be noted that the implementation manner of the target object is not limited in this application. For example, it can be implemented using a certain App or a certain physical device.

[0031] In addition, for the problem raised for the target object (such as Figure 2 the problem 1 shown), this problem can be used to describe the difficulties encountered by the user when using the target object, so that the user cannot use the target object better. It should be noted that the implementation manner of this problem is not limited in this application. For example, it can be implemented using the difficulties caused by the defects existing in the target object itself (such as crashing) or the difficulties encountered by the user due to unfamiliarity with the target object.

[0032] Furthermore, the manner of raising the above problem is not limited in this application. For example, this problem can be raised using a pre-configured inquiry entry for the target object, such as an entry through an Application Programming Interface (API), a mini-program, a QR code, etc.

[0033] For another example, in some scenarios, in order to better improve real-time performance, the above problem can be raised using a conversation group (also known as a dialogue group) in any tool with instant messaging function, so that the user can obtain the solution to this problem in a timely manner through this conversation group. Among them, this conversation group is used to display some conversations (such as conversations including this problem), and the implementation manner of this conversation group is not limited in this application.

[0034] Moreover, the representation manner of the above problem is not limited in this application. For example, this problem includes some or all of text, pictures, audio, and video, so that this problem can more accurately and flexibly describe the difficulties encountered by the user when using the target object.

[0035] The target model refers to a machine learning model that has been pre-trained and can perform answer prediction processing for a problem, so that subsequent automatic reply can be realized with the help of this target model. It should be noted that the implementation manner of this target model is not limited in this application. For example, this target model can be implemented using a deep learning large model.

[0036] In addition, the present application does not limit the working principle of the target model. For example, the target model can be used to: perform answer prediction processing on a question by using a knowledge base pre-configured for the target model to obtain a predicted answer corresponding to the question. The knowledge base is used to record answers corresponding to some questions; moreover, the present application does not limit the knowledge base. For example, the knowledge base can at least include various Q&A documents of the target object, the standard operating procedure (SOP) of the target object, the operation manual of the target object, etc. For another example, the knowledge base can also include question-and-answer pairs obtained by manual collation, so that the knowledge base can describe as accurately and comprehensively as possible the questions that may be encountered when using the target object and their corresponding solutions.

[0037] It has been found through research that the questions provided by users for the target object may carry a large amount of useless information. Therefore, in order to avoid the interference caused by this useless information to answer prediction, the present application also provides an implementation manner of the above target model. In this manner, the target model includes a simplification network and a prediction network. The simplification network is used to simplify the question to obtain a simplification result, so that the simplification result can describe the semantic information carried by the question as concisely, clearly, and accurately as possible. The prediction network is used to perform answer prediction processing on the simplification result, such as searching for the answer that best matches the question described by the simplification result from the knowledge base corresponding to the target model.

[0038] It should be noted that the present application does not limit the implementation manner of the above simplification processing. For example, it can be implemented by using any semantic analysis processing, such as natural language processing (NLP). The NLP can be used to perform semantic analysis on the question provided by the user, extract key information and classify and summarize it to obtain the simplification result of the question, so that the simplification result can concisely, clearly, and accurately describe the semantic information carried by the question. In addition, the NLP can include processes such as sentiment analysis and topic modeling, so as to be able to accurately identify the user's concerns and main content from the question subsequently.

[0039] The first answer refers to the answer obtained by processing the above question by using the target model (such as Figure 2 the answer 1 shown), so that the first answer can represent the solution determined for the question by means of an automatic reply.

[0040] In addition, the present application does not limit the implementation manner of the above S1. For example, when the target model implements answer prediction processing based on a pre-constructed knowledge base, S1 can specifically be: in response to a question raised for the target object, the target model uses the knowledge base to perform answer prediction processing to obtain a first answer, so that the first answer can represent the answer determined from the knowledge base by the target model and most suitable for answering the question.

[0041] It can be seen that in a possible implementation manner, when the first answer is obtained by the target model using the knowledge base for answer prediction processing, the knowledge base can at least include the first answer.

[0042] For another example, when the target model includes a simplification network and a prediction network, and the target model implements answer prediction processing based on a pre-constructed knowledge base, the above S1 can specifically be: in response to a question raised for the target object, the question is input into the target model, so that the simplification network in the target model performs simplification processing on the question to obtain a simplification result (such as question 2), so that the simplification result can express the content that the user wants to express through the question in a concise, clear, and accurate manner, so that subsequently, the prediction network in the target model can find the answer in the knowledge base that best matches the simplification result as the first answer, so that the first answer can represent the answer determined from the knowledge base and most suitable for answering the question.

[0043] It can be seen that in a possible implementation manner, when the first answer is obtained by the target model using the knowledge base for answer prediction processing, the knowledge base can at least include the first answer and the question corresponding to the first answer. The question corresponding to the first answer refers to the question in the knowledge base that is closest (most similar) to the question raised by the user for the target object, and the first answer refers to the answer in the knowledge base that has a corresponding relationship with the closest question.

[0044] Based on the relevant content of S1 above, for the target object (such as a certain product), after detecting a question raised by the user for the target object, a target model with answer prediction function pre-constructed is used to process the question to obtain a first answer, so that the first answer can represent the solution determined for the question through the automatic reply method.

[0045] S2: Display the first answer.

[0046] In the present application, after obtaining the first answer determined for the above question through the automatic reply method, the first answer is displayed to the user so that the user can check whether the first answer can solve the question, so as to determine whether it is necessary to switch to manual reply based on the user's feedback on the first answer subsequently.

[0047] S3: In response to a feedback operation triggered for the first answer indicating that the first answer does not match the question, obtain and display a second answer obtained by manually replying to the question.

[0048] Wherein, the feedback operation is triggered by the user for the first answer and is used to feedback whether the first answer can accurately solve the question raised by the user for the target object. It should be noted that the implementation manner of the above feedback operation is not limited in this application.

[0049] The second answer is obtained by a relevant manager of the target object manually replying to the above question, so that the second answer can represent the solution determined for the question by the manual reply method, thereby enabling the second answer to more accurately describe how to solve the question.

[0050] It can be seen that in a possible implementation manner, the degree of adaptation between the solution indicated by the above second answer and the "question raised for the target object" above is higher than the degree of adaptation between the solution indicated by the above first answer and the "question raised for the target object" above.

[0051] Based on the relevant content of S1 to S2 above, for the reply solution provided in this application, after obtaining the question raised by the user for the target object (such as a certain product), first use the target model to process the question to obtain a first answer, so that the first answer is used to represent the result of the automatic reply to the question; then display the first answer to the user, so that after detecting the feedback operation triggered by the user for the first answer, if the feedback operation is used to indicate that the first answer does not match the question, it is determined that the answer determined by the automatic reply method for the question is inaccurate, so obtain and display a second answer obtained by manually replying to the question. In this way, the reply cost can be reduced on the premise of ensuring the reply accuracy as much as possible through the automatic reply + manual correction method.

[0052] In addition, in order to better improve the reply effect, this application also provides a possible implementation manner of the reply method. In this manner, the reply method may include the following steps 11 - step 13.

[0053] Step 11: In response to the question raised by the user for the target object using the target tool, the target tool inputs the question into the target model, so that the target model processes the question to obtain a first answer.

[0054] Wherein, the target tool is a carrier of the target model, enabling the user to raise questions for the target object with the help of this carrier, and also enabling the target model to obtain the questions that need to be processed for answer prediction with the help of this carrier.

[0055] In addition, the present application does not limit the implementation manner of the target tool. For example, the target tool may refer to a certain social tool originally existing on the terminal device held by the user (such as any instant messaging tool), so as to effectively avoid defects caused by the difficulty in finding the problem query entrance or the need to re-download the problem query entrance, thereby reducing the difficulty of problem query.

[0056] For another example, in order to better improve the real-time performance, the target tool may be implemented using any instant messaging tool, so that the target tool can provide instant messaging services to the user, and thus the target model with the target tool as the carrier can provide real-time question-and-answer services to the user, which is conducive to quickly responding to user needs and improving the interaction efficiency.

[0057] In addition, the present application does not limit the working principle of the target tool. For example, when the target tool can provide a session group service to the user, the working principle of the target tool may be: the user asks a question about the target object through the session group of the target tool (such as Figure 2 the session group shown), for example, question 1 shown in Figure 2 so that the session group is used to display some sessions including the question (such as Figure 2 the session shown). For example, the session group can provide the conversation content between some accounts (such as user accounts + robot accounts with the function of automatically answering questions), so as to subsequently obtain the solution to the question in time with the help of the session group.

[0058] Step 12: In response to the target tool obtaining the first answer (such as Figure 2 answer 1 shown) from the target model, display the first answer and the feedback component (such as Figure 2 the two buttons of helpful and unhelpful shown), so that the user can subsequently use the feedback component to indicate whether the first answer can solve the question raised by the user for the target object.

[0059] It should be noted that the present application does not limit the implementation manner of the feedback component. For example, the feedback component may include multiple buttons (such as Figure 2 the two buttons of helpful and unhelpful shown), and different buttons are respectively used to express different contents feedback by the user for the first answer, so that the user can subsequently use different buttons to express their different views on the first answer.

[0060] It should also be noted that the implementation manner of step 12 in this application is not limited. For example, when the above problem is proposed through the conversation group of the target tool, step 12 may specifically be: in response to the target tool obtaining the first answer from the target model, display the first answer and the feedback component in the conversation group, so that subsequent users can not only view the first answer in the conversation group in a timely manner, but also use the feedback component to indicate in a timely manner whether the first answer can solve the problem, which is beneficial to improving the response efficiency.

[0061] For another example, in some scenarios, when the above problem is proposed through the conversation group of the target tool, step 12 may specifically include: in response to the target tool obtaining the first answer from the target model, display the first answer, the problem corresponding to the first answer (or the simplified result obtained by simplifying the "problem proposed for the target object"), and the feedback component in the conversation group, so that users can view the first answer and the problem used when querying the first answer through the conversation group, so that subsequent relevant managers of the target object can better analyze the reason why the first answer is inaccurate (such as the reason that the problem found in the knowledge base is incorrect) based on these contents. This is beneficial to focusing on optimizing the part of the target model related to this reason (such as the simplified network) during the subsequent model update process, thereby improving the model performance.

[0062] Step 13: In response to a feedback operation triggered by the above feedback component (such as Figure 2 the operation of clicking the "not helpful" button as shown) indicating that the first answer does not match the above problem, obtain and display the second answer obtained by manually answering the problem (such as Figure 2 Answer 2 as shown).

[0063] It should be noted that the implementation manner of step 13 in this application is not limited.

[0064] In addition, in order to better reduce the work pressure of the relevant managers of the target object, this application also provides a possible implementation manner of step 13. In this manner, when the above problem is proposed through the conversation group of the target tool, step 13 may specifically include: in response to a feedback operation triggered by the above feedback component indicating that the first answer does not match the above problem, add the target login account of the target tool (such as the account used by the manager to log in to the target tool) to the conversation group. The target login account is used to manually answer the problem, and display the second answer provided by the target login account for the problem in the conversation group. The target login account refers to the account used by the relevant manager of the target object to log in to the target tool.

[0065] It can be seen that when querying questions by means of a conversation group, if it is detected that the user does not approve of the first answer given by the automatic reply method, the login account corresponding to the relevant manager of the target object in the target tool can be directly added to the conversation group, so that the manager can timely view the questions raised by the user for the target object and the first answer by means of the conversation group, and enable the manager to manually reply to the question on the basis of these contents. In this way, it can be realized that only when it is determined that the automatic reply method is inaccurate, the manager needs to participate in the reply process of the question, which is beneficial to reducing the work pressure of the manager.

[0066] Based on the relevant content of the above steps 11 to 13, the present application provides a reply solution based on a certain social tool (such as an instant messaging tool), so that users can query questions about a certain product by means of the social tool originally held on their terminal devices, which is beneficial to improving the user experience.

[0067] In addition, in order to better improve the automatic reply effect, the present application also provides a possible implementation manner of the above reply method. In this manner, the reply method may at least include step 21 below.

[0068] Step 21: In response to the second answer matching the above question, update the target model according to the second answer and the question, so that the updated target model can learn that the answer most suitable for answering the question is the second answer, thereby enabling the updated target model to have better performance.

[0069] It should be noted that the present application does not limit the implementation manner of the above step 21. For example, it can be implemented by adopting any model training method.

[0070] Based on the relevant content of the above step 21, for the pre-trained target model, when using the target model to execute some question reply tasks, it can automatically collect the questions raised by the user for the target object and their answers, and perform some processing on these collected data (such as cleaning processing for removing useless data and standardization processing for restricting the data format, etc.) to obtain new data, so that the new data can be passed to the target model subsequently, enabling the target model to use these new data for continuous learning and training. In this way, it can be realized that the parameters in the target model are automatically adjusted based on the changes in the questions raised by the user, so that the adjusted parameters can adjust the divergence degree of answering questions to an as reasonable state as possible, thereby enabling the target model with adjusted parameters to have better reply performance, which is beneficial to improving the model adaptability and accuracy.

[0071] It should be noted that the implementation manner of the step of "performing some processing on the collected data" in this application is not limited. For example, specific algorithms or machine learning models with corresponding functions can be used for implementation.

[0072] In addition, in order to better improve the automation effect, this application also provides a possible implementation manner of the above step 21. In this manner, when the first answer is obtained by the target model using the knowledge base for answer prediction processing, step 21 can specifically be: in response to the second answer matching the above question, generating a question-answer pair based on the second answer and the question, so that the question-answer pair includes the question and the second answer, so that after adding the question-answer pair to the knowledge base, the target model is updated based on the knowledge base, so that the updated target model can learn how to better use the knowledge base for automatic answering.

[0073] It can be seen that for a pre-trained target model, when using the target model to perform some question-answering tasks, it can automatically collect the questions and answers raised by users for the target object, and perform some processing on these collected data to obtain new data; then, use these new data to update the knowledge base configured for the target model in real time, so that the updated knowledge base includes these new data, so that the updated knowledge base can more accurately and comprehensively describe the answers to some questions, and then make the target model trained based on the updated knowledge base have better automatic answering performance, which is beneficial to improving the response speed and efficiency of question answering. For ease of understanding, an example is given below.

[0074] As an example, for the target model, after constructing the knowledge base using some text content of the target object (such as Q&A documents, SOPs, operation manuals, and text-type Q&A pairs obtained by manual collation), the target model is trained using the knowledge base so that the trained target model can better use the knowledge base for question-answering processing, so that some question-answering tasks can be performed using the target model later; however, when the questions provided by the user for the target object encountered by the target model include non-text data (such as pictures, videos, audio, etc.), since the target model only has better performance on text data, but has relatively poor performance on non-text data, the knowledge base and the target model can be updated using the non-text data and the answers given for the non-text data by manual answering, so that the target model can not only have better performance on text data, but also have better performance on non-text data, so that the automatic answering implemented based on the target model has better answering effects.

[0075] Based on the relevant content of the above response method, the present application provides a product intelligent response solution for analyzing user questions based on a large deep learning model (such as the above-mentioned target model), and this solution includes the features shown in (1)-(6) below.

[0076] (1) Collection and processing of user questions, specifically: first, automatically collect questions (and answers) raised by users for the product through various channels, such as integrated API interfaces, data warehouse bottom tables, and user question feedback group messages (such as some conversations in the session group), to ensure the timeliness and comprehensiveness of the data; then, clean and standardize the collected information, such as removing noise and unifying the format, etc., for subsequent use of these data to optimize the target model. It can be seen that the present application greatly improves the efficiency of processing user questions and minimizes the need for manual responses by automatically collecting and processing user questions and their answers.

[0077] (2) Model training and continuous learning, specifically: establish a dynamic update mechanism, regularly transfer the newly collected data to the target model to achieve continuous learning and training, so that based on this mechanism, it can automatically adjust the model parameters according to the changes in user questions (such as changes from text data to non-text data, etc.) to improve the adaptability and accuracy of the model. In addition, the present application automatically transfers the newly collected data to this model to ensure that the model can update its knowledge base in real time, so as to provide more accurate answers in subsequent interactions. In this way, this process reduces manual intervention and improves the response speed and efficiency of question answering. Moreover, the model can accurately summarize user questions to ensure that the questions can be structured and output using concise, clear, and accurate descriptions subsequently.

[0078] (3) The target model in the present application uses natural language processing technology to perform semantic analysis on user questions to extract key information and classify and summarize them. In this way, this process includes processing such as sentiment analysis and topic modeling, so that this process can accurately identify user concerns and main questions from user questions. It can be seen that accurately summarizing user questions using the target model effectively improves the ability to understand user needs and questions.

[0079] (4) Manual intervention and optimization method, specifically: the present application provides a manual review interface, so that this manual review interface can provide a relatively friendly operation interface for relevant managers of the target object (such as a certain product), enabling them to view the automatically generated answers to user questions and manually correct or optimize the answers according to the actual situation. In this way, the accuracy and practicality of the final answer can be ensured. It can be seen that the present application combines the advantages of manual response and automatic response achieved by machine learning to achieve a more efficient question answering solution.

[0080] (5) Automatic intelligent Q&A robot, specifically: This application identifies the questions raised by users through a certain social tool (such as chat applications, social media), and uses the target model to automatically reply to obtain answers, realizing the real-time Q&A function. In this way, it can quickly respond to user needs and improve the interaction efficiency. In addition, this application also sets up a question feedback button in the social tool, enabling users to easily submit dissatisfaction or suggestions about the answers, and automatically records these feedbacks for subsequent analysis and optimization of the performance of the target model. It can be seen that this application can instantly respond to user needs, continuously improve the automatic answer service for questions through the question feedback mechanism, and significantly enhance the overall user experience.

[0081] (6) Effect evaluation and optimization, specifically: This application establishes a monitoring mechanism to evaluate the accuracy rate and user satisfaction of the automatic answer results of questions in real time for timely adjustment of the target model. In addition, this application can continuously optimize the model according to the monitoring results to ensure that it can always better meet user needs.

[0082] Based on the answer method provided by the embodiments of this application, the embodiments of this application also provide an answer device, which will be explained and described below in combination with Figure 3 for explanation and illustration. Among them, Figure 3 is a schematic structural diagram of an answer device provided by the embodiments of this application. It should be noted that for the technical details of the answer device provided by the embodiments of this application, please refer to the relevant content of the above answer method.

[0083] As Figure 3 shown, the answer device 300 provided by the embodiments of this application includes:

[0084] A data processing unit 301, configured to respond to a question raised for a target object, and use the target model to process the question to obtain a first answer;

[0085] A first display unit 302, configured to display the first answer;

[0086] A second display unit 303, configured to, in response to a feedback operation triggered for the first answer indicating that the first answer does not match the question, obtain and display a second answer obtained by manually answering the question.

[0087] In a possible implementation manner, the question is raised by using a target tool, and the target tool is used to input the question into the target model;

[0088] The first display unit 302 is specifically configured to: in response to the target tool obtaining the first answer from the target model, display the first answer and a feedback component;

[0089] The second display unit 303 is specifically configured to: in response to a feedback operation triggered by the feedback component indicating that the first answer does not match the question, obtain and display a second answer obtained by manually replying to the question.

[0090] In a possible implementation manner, the question is proposed through a session group of the target tool, and the session group is used to display at least one round of sessions, and the at least one round of sessions includes the question.

[0091] The second display unit 303 is specifically configured to: in response to a feedback operation triggered by the feedback component indicating that the first answer does not match the question, add the target login account of the target tool to the session group, where the target login account is used to manually reply to the question, and display the second answer provided by the target login account for the question in the session group.

[0092] In a possible implementation manner, the target tool is used to provide an instant messaging service.

[0093] In a possible implementation manner, the answering device 300 further includes:

[0094] A model update unit, configured to update the target model according to the second answer and the question in response to the second answer matching the question.

[0095] In a possible implementation manner, the first answer is obtained by the target model performing answer prediction processing using a knowledge base, and the knowledge base includes the first answer.

[0096] The model update unit is specifically configured to: in response to the second answer matching the question, generate a question-answer pair according to the second answer and the question; after adding the question-answer pair to the knowledge base, update the target model according to the knowledge base.

[0097] In a possible implementation manner, the question includes some or all of text, pictures, audio, and video.

[0098] Based on the relevant content of the above-mentioned response device 300, the working principle of the device 300 includes: after obtaining the question raised by the user for the target object (such as a certain product), first use the target model to process the question to obtain a first answer, so that the first answer is used to represent the result obtained by automatically answering the question; then display the first answer to the user. After detecting the feedback operation triggered by the user for the first answer, if the feedback operation is used to indicate that the first answer does not match the question, it is determined that the answer determined by the automatic answering method for the question is inaccurate. Therefore, obtain and display the second answer obtained by manually answering the question. In this way, it is possible to reduce the response cost on the premise of ensuring the response accuracy as much as possible through the automatic answering + manual correction method.

[0099] In addition, an embodiment of the present application also provides an electronic device, the device includes a processor and a memory: the memory is used to store instructions or computer programs; the processor is used to execute the instructions or computer programs in the memory, so that the electronic device executes any implementation manner of the response method provided by the embodiment of the present application.

[0100] See Figure 4 , which shows a schematic structural diagram of an electronic device 400 suitable for implementing the embodiments of the present disclosure. The terminal device in the embodiments of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The electronic device shown is only an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.

[0101] As Figure 4 shown, the electronic device 400 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 401, which may perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage device 408 into the random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the electronic device 400 are also stored. The processing device 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. The input / output (I / O) interface 405 is also connected to the bus 404.

[0102] Typically, the following devices can be connected to the I / O interface 405: an input device 406 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 407 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 408 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 409. The communication device 409 can allow the electronic device 400 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 4 the electronic device 400 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. Instead, more or fewer devices can be implemented or had.

[0103] Specifically, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program codes for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication device 409, or installed from the storage device 408, or installed from the ROM 402. When the computer program is executed by the processing device 401, the above functions defined in the method of the embodiment of the present disclosure are executed.

[0104] The electronic device provided by the embodiment of the present disclosure and the method provided by the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0105] An embodiment of the present application also provides a computer-readable medium, in which instructions or a computer program are stored. When the instructions or the computer program run on a device, the device is enabled to execute any implementation manner of the reply method provided by the embodiment of the present application.

[0106] It should be noted that the computer-readable medium described above in the present disclosure may be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present disclosure, the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium, and the computer-readable signal medium can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0107] In some embodiments, the client and the server can communicate using any currently known or future-developed network protocol such as HTTP (Hyper Text Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks ("LANs"), wide area networks ("WANs"), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0108] The above computer-readable medium may be included in the above electronic device; or it may exist separately and not be assembled into the electronic device.

[0109] The above computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device can execute the above method.

[0110] Computer program code for performing the operations of this disclosure may be written in one or more programming languages or combinations thereof. The foregoing programming languages include, but are not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0111] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that, in some alternative implementations, the functions noted in the blocks may occur in an order different from that noted in the drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0112] The units involved in the embodiments described in this disclosure may be implemented in software or in hardware. Among them, the name of the unit / module does not constitute a limitation to the unit itself in some cases.

[0113] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGA), application specific integrated circuits (ASIC), application specific standard products (ASSP), system on a chip (SOC), complex programmable logic devices (CPLD), and so on.

[0114] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0115] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the systems or devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description in the method section.

[0116] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships can exist. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist simultaneously. Here, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or a similar expression means any combination of these items, including any combination of single item (one) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0117] It should also be noted that, in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0118] The steps of the methods or algorithms described in connection with the embodiments disclosed herein can be implemented directly in hardware, in software modules executed by a processor, or in a combination thereof. The software modules can be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium well-known in the art.

[0119] The foregoing description of the disclosed embodiments enables those skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A reply method, characterized in that: The method comprises: In response to a question raised to a target object, the question is processed using the target model to obtain a first answer; Display the first answer; In response to a feedback operation triggered for the first answer indicating that the first answer does not match the question, a second answer obtained by manually replying to the question is obtained and displayed.

2. The method according to claim 1, characterized in that The problem is posed using a target tool, and the target tool is used to input the problem into the target model; The displaying of the first answer comprises: In response to the target tool obtaining the first answer from the target model, displaying the first answer and a feedback component; The feedback operation triggered in response to the first answer is used to indicate that the first answer does not match the question, and obtain and display a second answer obtained by manually answering the question, including: In response to a feedback operation triggered by the feedback component indicating that the first answer does not match the question, a second answer obtained by manually replying to the question is obtained and displayed.

3. The method according to claim 2, characterized in that The question is asked through a conversation group of the target tool, and the conversation group is used to present at least one round of conversation, and the at least one round of conversation includes the question; The step of obtaining and displaying a second answer obtained by manually answering the question in response to a feedback operation triggered by the feedback component indicating that the first answer does not match the question comprises: In response to a feedback operation triggered by the feedback component indicating that the first answer does not match the question, a target login account of the target tool is added to the conversation group, the target login account is used to manually reply to the question, and a second answer to the question provided by the target login account is displayed in the conversation group.

4. The method according to claim 2, characterized in that: The target tool is used to provide instant messaging services.

5. The method according to claim 1, characterized in that The method further comprises: In response to the second answer matching the question, the target model is updated according to the second answer and the question.

6. The method according to claim 5, characterized in that The first answer is obtained by the target model performing answer prediction processing using a knowledge base, and the knowledge base includes the first answer; In response to the second answer matching the question, updating the target model according to the second answer and the question, comprises: In response to the second answer matching the question, generating a question-answer pair based on the second answer and the question; After the question-answer pair is added to the knowledge base, the target model is updated according to the knowledge base.

7. The method according to any one of claims 1 to 6, characterized in that: The question may include part or all of the text, picture, audio and video.

8. A reply device, characterized in that: include: A data processing unit, configured to respond to a question raised to a target object and process the question using a target model to obtain a first answer; A first display unit, used to display the first answer; The second display unit is used to obtain and display a second answer obtained by manually replying to the question in response to a feedback operation triggered for the first answer indicating that the first answer does not match the question.

9. An electronic device, characterized in that: The device comprises: a processor and a memory; The memory is used to store instructions or computer programs; The processor is used to execute the instructions or computer programs in the memory so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable medium, characterized in that The computer-readable medium stores instructions or computer programs, and when the instructions or computer programs are executed on a device, the device is enabled to execute the method according to any one of claims 1 to 7.

11. A computer program product, characterized in that It comprises a computer program carried on a non-transitory computer-readable medium, the computer program comprising a program code for executing the method according to any one of claims 1 to 7.

Images