Question and answer method and device, electronic equipment, storage medium and program product

Abstract

The invention discloses a question and answer method and device, electronic equipment, a storage medium and a program product, and relates to the field of data processing technologies, artificial intelligence technologies, large model technologies and large language models.The method comprises the steps that first information is obtained, the first information comprises a first question and preceding text information of the first question, and processing the first information based on a question-answer model to obtain a first processing result, the question-answer model being configured to call a first information processing tool based on the first information to obtain associated information, and processing the first information based on the associated information to obtain the first processing result. The method can improve the question-answering effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, artificial intelligence, large model and large language model, in particular to a question and answer method and device, an electronic device, a storage medium and a program product. BACKGROUND

[0002] As an important communication bridge between resource providers and resource seekers, intelligent customer service generally relies on preset rules to process questions raised by resource seekers and provide corresponding answers to resource seekers. Although the question and answer method based on preset rules can handle basic questions, it may have poor reply effects such as mechanical rigidity and lack of flexibility when facing complex and variable questions. Therefore, there is an urgent need to provide a new question and answer method. SUMMARY

[0003] Therefore, the present application provides a question and answer method, device, electronic device, storage medium and program product to solve the problem of poor reply effect.

[0004] In a first aspect, the present application provides a question and answer method, comprising:

[0005] obtaining first information, the first information comprising a first question and context information of the first question;

[0006] processing the first information based on a question and answer model to obtain a first processing result, the question and answer model being configured to call a first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result.

[0007] In a second aspect, the present application provides a question and answer device, comprising:

[0008] an information obtaining module configured to obtain first information, the first information comprising a first question and context information of the first question;

[0009] an information processing module configured to process the first information based on a question and answer model to obtain a first processing result, the question and answer model being configured to call a first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result.

[0010] In a third aspect, the present application provides an electronic device, comprising a memory and a processor, which are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the question and answer method of the first aspect or any of the corresponding embodiments thereof.

[0011] In a fourth aspect, the present application provides a computer readable storage medium, having stored thereon computer instructions for causing a computer to execute the question and answer method of the first aspect or any of the corresponding embodiments thereof.

[0012] In a fifth aspect, the present application provides a computer program product, comprising computer instructions for causing a computer to execute the question and answer method of the first aspect or any of the corresponding embodiments thereof.

[0013] The question and answer method provided by the embodiments of the present application comprises obtaining first information, the first information comprising a first question and context information of the first question, processing the first information based on a question and answer model to obtain a first processing result, the question and answer model being configured to call a first information processing tool based on the first information to obtain associated information and process the first information based on the associated information to obtain the first processing result. In the method, the question and answer model can automatically call the first information processing tool in the process of processing the first information, that is, the necessary associated information can be obtained through the first information processing tool, relevant operations are performed, and all data does not need to be input at one time, thereby greatly reducing the waste of computing resources, enabling the question and answer model to more accurately obtain and process information, and thus improving the question and answer effect. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the specific embodiments or the related art, the drawings needed to be used in the specific embodiments or the related art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0015] Figure 1 is a first schematic diagram of a question and answer method in the related art;

[0016] Figure 2 is a second schematic diagram of a question and answer method in the related art;

[0017] Figure 3 is a schematic diagram of an application scenario according to an embodiment of the present application;

[0018] Figure 4 is a first flowchart of a question and answer method according to an embodiment of the present application;

[0019] Figure 5 is a second flowchart of a question and answer method according to an embodiment of the present application;

[0020] Figures 6a-6c is a schematic diagram of processing of continuous questions according to an embodiment of the present application;

[0021] Figures 7a-7b is a processing schematic diagram of tool invocation timeout according to an embodiment of the present application;

[0022] Figure 8 is a processing schematic diagram of feedback information according to an embodiment of the present application;

[0023] Figure 9 is a structural block diagram of a question and answer device according to an embodiment of the present application;

[0024] Figure 10 is a hardware structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0026] It can be understood that, before using the technical solutions disclosed in the embodiments of the present application, the type, use range, use scenario and the like of personal information involved in the present application should be informed to the user and the authorization of the user should be obtained through appropriate means according to relevant laws and regulations.

[0027] For example, in response to receiving the active request of the user, prompt information is sent to the user to explicitly prompt the user that the operation requested to be executed will need to obtain and use the personal information of the user. Thus, the user can voluntarily choose whether to provide the personal information to the electronic device, application program, server or storage medium and the like software or hardware that execute the operation of the technical solutions of the present application according to the prompt information.

[0028] As an optional but non-limiting implementation manner, in response to receiving the active request of the user, the manner of sending the prompt information to the user may, for example, be a pop-up window manner, and the prompt information may, for example, be presented in the form of text in the pop-up window. In addition, the pop-up window may, for example, also carry a selection control for the user to select "agree" or "disagree" to provide the personal information to the electronic device.

[0029] It can be understood that the above notification and obtaining of the authorization of the user are only schematic and do not limit the implementation manner of the present application, and other manners meeting the relevant laws and regulations can also be applied to the implementation manner of the present application.

[0030] It can be understood that the data involved in the technical solution (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of the corresponding laws and regulations and relevant provisions.

[0031] In some related technologies, the question and answer method provided is as shown in Figure 1 The main way is to understand the user's given question and the question's context information through natural language understanding of a non-large language model (Non-LLM), and match a pre-set Frequently Asked Questions (FAQ) library based on the understanding result, or reply according to a configured Standard Operating Procedure (SOP). That is, in the reply module, the understanding result (claim or knowledge point) is matched in the FAQ, or the SOP gives a reply, and the reply is provided to the user.

[0032] However, this way is difficult to handle out-of-process questions, resulting in obvious mechanical feeling of the dialogue. The upper limit of the question and answer system is limited by the artificially pre-set rules and processes, and it is difficult to cope with complex and variable user needs.

[0033] In some related technologies, the question and answer method provided is as shown in Figure 2 It is implemented by a direct service mode of a large model. That is, all relevant information is concatenated into a prompt, including but not limited to the question, the question's context information, after-sales policies, and resource logistics information, etc. That is, the question and context information are understood and recognized by a large language model, and the human agent then queries the relevant information through corresponding tools (including but not limited to a scheme verification tool, a business knowledge query tool, and a FAQ query tool) according to the processing result of the large language model, concatenates the queried information into the prompt as a reply generation model prompt, so that the model generates a reply and pushes it to the user.

[0034] However, in this way, all information is concatenated into the prompt, resulting in information redundancy and a large amount of unnecessary information being input into the model. At the same time, there may be a lack of information, and some key but infrequently used information may be ignored, and all information needs to be processed again each time, lacking coherence.

[0035] Further, in the above scheme, a large amount of information needs to be input into the model at one time, resulting in waste of computing resources and distraction of attention. That is, the single input information overload limits the depth of thinking ability of the model, because the model processes too much information, which consumes a lot of resources, and in the limited processing time, the input information takes too much time, limiting the depth of thinking ability.

[0036] Based on this, the embodiment of the application provides a question and answer method, including obtaining first information, the first information including a first question and context information of the first question, processing the first information based on a question and answer model to obtain a first processing result, the question and answer model being configured to call a first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result. In the method, in the process of processing the first information, the question and answer model can automatically call the first information processing tool, that is, the necessary associated information can be obtained through the first information processing tool, relevant operations are performed, all data does not need to be input at one time, the waste of computing resources is greatly reduced, the question and answer model can more accurately obtain and process information, and therefore the question and answer effect can be improved. Further, the question and answer model in the embodiment of the application can also be understood as providing a question and answer service based on reasoning and acting of a large model.

[0037] As an optional application scenario of the embodiment of the present disclosure, as shown in Figure 3 The terminal device 310 is installed with an application 301, and the user 330 can interact with the application 301 through the terminal device 310 and / or an access device of the terminal device 310.

[0038] Exemplarily, the application 301 can be any application that can provide a question and answer service. For example, the application 301 can be a question and answer interaction application, such as a text-to-text application, a figure-to-text application, and the like. In Figure 1 In the application scenario shown in the figure, if the application 301 is in an active state, the terminal device 310 can present an interface 302 of the application 301. The interface 302 can include various pages that can be provided by the application 301, such as an interaction page, a setting page, a query page, and the like.

[0039] In some embodiments, the terminal device 310 is in communication connection with the server 320 to implement the provision of services of the application 301. The terminal device 310 can be a mobile terminal, a fixed terminal, or a portable terminal, and the like, including but not limited to a mobile phone, a desktop computer, a notebook computer, a multimedia tablet, an electronic book device, a game device, or any combination of the above, including accessories and peripherals of the devices or any combination thereof. In some embodiments, the terminal device 310 can also support any type of interface, and the server 320 can be various types of computing systems, servers that can provide computing capabilities, including but not limited to mainframes, edge computing nodes, computing devices in cloud environments, and the like.

[0040] It should be noted that Figure 3 This is only an example of an application scenario, and does not limit the protection scope of the present disclosure.

[0041] The embodiments of the present disclosure will be described below with reference to the drawings. It should be understood that the pages shown in the drawings are merely examples, and various page designs can actually exist. Various graphical elements in the page can have different arrangements and different visual representations, one or more elements of which can be omitted or replaced, and one or more other elements can also exist, which are not limited in the embodiments of the present disclosure. In addition, the embodiments are mainly described below with respect to the server 320.

[0042] According to the embodiments of the present application, a question and answer method embodiment is provided. It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from here.

[0043] In the present embodiment, a question and answer method is provided, which can be used in the server described above, Figure 4 is a flowchart of the question and answer method according to the embodiments of the present application, as Figure 4 shown, the flow includes the following steps:

[0044] Step S401, obtaining first information.

[0045] The first information includes a first question and context information of the first question.

[0046] The first question is used to represent information input by the user through the question and answer application, including but not limited to voice input, text input, etc. Further, the information input through the question and answer application can be input by the user through the interactive mode, or the question and answer application can provide recommended questions, and the user can interact with these recommended questions as the first question.

[0047] It should be noted that the first question is not limited to information expressed by interrogative sentences, and in the present embodiment, the description information provided by the user side is referred to as the first question.

[0048] In combination with the application scenario shown in Figure 3 , the user 330 provides the first question by interacting with the interface 302 provided by the application 301, and accordingly, the terminal device 310 can obtain the first question, and the terminal device 310 sends the obtained first question and the context information of the first question to the server 320. Thus, the server 320 can obtain the first question and its context information.

[0049] The context information of the first question is used to represent all context information of the first question in the current session, or N context contents close to the first question, etc. The range of the context information is not limited here, and can be set according to actual needs.

[0050] At step S402, the first information is processed based on the question and answer model to obtain a first processing result.

[0051] The question and answer model is configured to call a first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result.

[0052] After obtaining the first information, the question and answer model can first analyze and understand the first information, and actively call the corresponding first information processing tool to obtain the associated information required to solve the first problem. Then, the first information is processed based on the associated information to obtain the first processing result.

[0053] Specifically, in the process of processing the first information, the first information processing tool called by the question and answer model can be one or more, and the specific data is related to the processing of the first information, which is not limited herein. The first information processing tool includes but is not limited to a business knowledge query tool, an FAQ query tool, and a scheme verification tool, etc. The question and answer model calls the corresponding first information processing tool as needed according to the specific situation to obtain the associated information.

[0054] By encapsulating various information processing tools (such as after-sales operation tools and query tools, etc.), the question and answer model is called. Accordingly, the question and answer model determines the first information processing tool to be called by thinking.

[0055] In order to ensure that the question and answer model can independently call the first information processing tool as needed, in the process of training the question and answer model, the target thinking chain can be provided in the sample data. Accordingly, the result loss and the thinking chain loss can be included when calculating the loss, and the fusion result of the two is taken as the total loss of the question and answer model training. The thinking chain is used to represent the thinking process of the question and answer model for the sample problem, including which information processing tools are called and the calling order between these called information processing tools. Specifically, each sample data includes a sample question and its context information, a sample thinking chain, and a target reply. The sample question and its context information are input into the initial question and answer model to obtain the executed thinking chain and the predicted reply. The thinking chain loss is obtained by the difference between the executed thinking chain and the sample thinking chain, and the result loss is obtained by the difference between the target reply and the preset reply. The total loss is obtained by fusing the two. The fusion method includes but is not limited to weighting and taking the average, etc.

[0056] Of course, the training of the question and answer model can also be implemented in other ways, which is not limited herein. The specific setting is required according to the actual demand, and it is required to ensure that the question and answer model obtained by training can independently call the first information processing tool as needed to obtain the associated information.

[0057] The question and answer method provided in the embodiment, in the process of processing the first information, the question and answer model can automatically call the first information processing tool, that is, through the first information processing tool, the necessary associated information can be obtained, and relevant operations are performed, without inputting all data at one time, greatly reducing the waste of computing resources, so that the question and answer model can more accurately obtain and process information, thereby improving the question and answer effect.

[0058] A question and answer method is provided in the embodiment, which can be used for the server side described above, Figure 5 is a flowchart of the question and answer method according to the embodiment of the application, as Figure 5 shown, the flowchart includes the following steps:

[0059] In step S501, the first information is obtained.

[0060] The first information includes the first question and the context information of the first question. For details, refer to step S401 of the embodiment shown in Figure 4 , which will not be described here again.

[0061] In step S502, the first information is processed based on the question and answer model to obtain a first processing result.

[0062] The question and answer model is configured to call the first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result.

[0063] Specifically, the above step S502 includes:

[0064] In step S5021, the question and answer model performs semantic analysis on the first information to determine the first information processing tool to be called.

[0065] The first information is input into the question and answer model, and semantic understanding is performed by the question and answer model to analyze the appeal or knowledge point, etc., and then the first information processing tool to be called is determined based on this. For example, a resource query tool is called to obtain detailed information of resources involved in the first question, etc.

[0066] In some optional embodiments, if the first information processing tool to be called is multiple, the question and answer model is further used to determine the calling sequence between the first information processing tools.

[0067] The question and answer system autonomously decides the first information processing tool to be called, and if the first information processing tool is multiple, the question and answer model obtains the calling sequence between the first information processing tools according to the thought chain in the thinking process. That is, the sequence of calling the multiple first information processing tools. The question and answer model can combine the first information processing tools as needed according to the result of semantic understanding to obtain the required associated information.

[0068] The question and answer model can flexibly combine the first information processing tool according to the semantic analysis result of the first information, so that it can process complex scenarios, thereby ensuring that the question and answer model does not need to distinguish scenarios, that is, one question and answer model can meet the needs of multiple business scenarios.

[0069] In some optional embodiments, the question and answer model is configured to call one or more of the following tool types:

[0070] The first tool type is used to verify whether the corresponding interface can be successfully executed; the second tool type is used to query information; and the third tool type is used for session summary and session planning.

[0071] The tool types that can be called by the question and answer model are configured to include one or more information processing tools.

[0072] As described above, the information processing tools are encapsulated for the question and answer model to call. Specifically, each information processing tool is encapsulated to leave a corresponding interface for the question and answer model to call. Depending on the different tool types to which each information processing tool belongs, it can be divided into a first tool type, a second tool type, and a third tool type, and different tool types are used to be responsible for corresponding processing matters. That is, the first tool type is used to verify whether the corresponding interface can be successfully executed. For example, the question and answer model can verify whether a processing strategy can be successfully executed by calling the information processing tool of the first tool type. The second tool type is used to query information, such as logistics track query, after-sales track query, etc. The underlying can reuse the interface of the agent query tool, which is consistent with the information displayed on the agent workstation. The third tool type is used for session summary and service planning. This type of information processing tool is implemented in the form of a workflow. It should be noted that the information processing tool can also call a corresponding large model to implement the corresponding function.

[0073] Further, the information processing tools included under each tool type can be one or more, which are specifically set according to actual needs, and are not limited herein. For example, for the second tool type, some information processing tools are used to query the logistics track, and some information processing tools are used to query the after-sales strategy, etc.

[0074] Through the setting of one or more tool types, multiple application scenarios are covered for the session model to call, so that the answer model can autonomously decide which information processing tools to use, etc., so that the question and answer method can process more complex question and answer scenarios.

[0075] In step S5022, the question and answer model calls the first information processing tool to obtain the associated information.

[0076] The question and answer model sequentially calls the corresponding first information processing tool in order to obtain the associated information obtained by the first information processing tool. The associated information includes but is not limited to query results, session summaries, execution verification results, and the like, and the specific content is related to the called first information processing tool.

[0077] In step S5023, the question and answer model processes the first information based on the associated information to obtain a first processing result.

[0078] In the process of processing the first information by using the associated information, the question and answer model is not limited to processing after all the associated information is complete, but sequentially processes according to the logical order according to the thinking chain of the question and answer model. The specific processing process is related to the thinking process of the question and answer model, and is not limited herein.

[0079] After processing the first information, the question and answer model obtains a first processing result. The first processing result output by the question and answer model can be directly pushed to the corresponding session page of the user for display, or can be pushed to the information confirmation page of the agent side for confirmation before being sent to the session page for display. The display form of the first processing result is not limited herein, and can be processed according to actual needs.

[0080] The question and answer method provided in this embodiment can automatically determine which first information processing tools need to be called by performing semantic analysis on the first information, that is, call the first information processing tools as needed to obtain necessary associated information, ensure that necessary instructions can be obtained to solve problems, and improve the quality of question and answer services.

[0081] In some optional embodiments, the question and answer method further includes: if the associated information corresponding to the first information processing tool is obtained, the call of the first information processing tool and the associated information are associatedly stored, and the content of the associated storage is used for inheritance by the question and answer model when processing a second question.

[0082] In order to ensure that the associated information of the first information processing tool that has been called can be directly inherited and used in subsequent processing, the associated information obtained by the first information processing tool can be associatedly stored with the call of the first information processing tool, that is, the content of the associated storage includes the call action of the first information processing tool and the associated information obtained by the call.

[0083] Due to the existence of the content of the associated storage, the relevant information can be directly inherited in the process of processing a second question, and the same first information processing tool does not need to be called again, thereby reducing the processing time and improving the response efficiency of the question and answer service.

[0084] Exemplarily, the call of the first information processing tool and the result obtained by the call are added into the message queue in pairs, and subsequent processing of the second question can directly access the message queue to obtain the corresponding associated information.

[0085] By associatively storing the call of the first information processing tool and the associated information obtained by the call, the question and answer model can inherit the existing result when processing continuous questions, without repeatedly processing all information each time, thereby improving processing efficiency.

[0086] In some optional embodiments, the above question and answer method comprises: if the second question is obtained in the process of processing the first information, the question and answer model processes the second question by inheriting the content of the associated storage obtained by processing the first information, and the content of the associated storage comprises the call of the first information processing tool and the associated information corresponding to the first information processing tool.

[0087] In actual application scenarios, users often continuously send questions due to the coherence of needs, which puts higher requirements on the processing capability of the question and answer model. For example, as shown in Figure 6a In this embodiment, if the second question is obtained in the process of processing the first question, the question and answer model processes the second question by inheriting the content of the associated storage obtained by processing the first information. That is, the question and answer model inherits the previous tool call thinking and result, which can avoid a lot of repetitive work and greatly improve the response efficiency of the question and answer model, thereby ensuring that the question and answer model can give a reply in a shorter time.

[0088] Exemplarily, as shown in Figure 6b In the related art, for the scenario as shown in Figure 6a Generally, problem A and problem B are treated as two independent problems and have no influence on each other, and the question and answer model processes problem A and problem B, and the reply corresponding to which problem is pushed first.

[0089] However, in this embodiment, as shown in Figure 6c If the question and answer model has called the first information processing tool and obtained the corresponding associated information in the process of processing problem A, the associated storage content of problem A can be inherited when processing problem B, and subsequent processing can be continued. This way of inheritance can efficiently process multiple questions sent by the user in sequence.

[0090] The question and answer model can inherit the previous thinking and result when processing continuous questions, without repeatedly processing all information each time, and thus can provide more coherent and natural question and answer services.

[0091] In some optional embodiments, the above-mentioned question and answer method further comprises: if the calling duration of the first information processing tool exceeds a preset threshold, triggering the question and answer model to generate first prompt information based on the first information.

[0092] In the process of calling the first information processing tool by the question and answer model, as shown in Figure 7a , due to the complexity of the business, the calling time process may be encountered. In this case, the present embodiment provides a preset threshold as shown in Figure 7b , if the calling duration of the first information processing tool exceeds the preset threshold, the question and answer model is automatically triggered to generate first prompt information based on the first information.

[0093] The first prompt information is generated according to the first information, for example, it can be "please wait while helping you query", or "please wait while querying order details", etc.

[0094] The first prompt information will be displayed on the conversation page, so that the user understands that the question and answer model is processing their request and is not in a stagnant state. For the question and answer model, after sending the first prompt information, it is still processing until the first processing result is obtained.

[0095] In the case where the calling duration of the first information processing tool exceeds the preset threshold, the question and answer model is triggered to generate the first prompt information for the current task scenario, which improves the user experience.

[0096] In some optional embodiments, the above-mentioned question and answer method further comprises:

[0097] Step a1, sending the first processing result to the information confirmation page for display.

[0098] Step a2, if feedback information is obtained for the first processing result, triggering the question and answer model to process the first information based on the feedback information to obtain an updated first processing result.

[0099] The information confirmation page is displayed on the agent side, and the processing results of multiple user-initiated question and answers can be displayed on the information confirmation page at the same time. The agent gives feedback information for the first processing result through interaction with the information confirmation page, and accordingly, the question and answer model can be triggered to process the first information based on the feedback information to obtain an updated first processing result.

[0100] In the present embodiment, as shown in Figure 8 , the first processing result generated by the question and answer model is not directly pushed to the question and answer page for display, but is first pushed to the information confirmation page for display, and then is pushed to the question and answer page for display after the accuracy of the first processing result is further confirmed by the agent personnel through manual intervention.

[0101] Specifically, pushing the first processing result to the information confirmation page display can ensure that the deviation of the communication direction or the specific type of problem is found in the first time. In this case, by obtaining feedback information and processing the first information again, a suitable processing scheme can be quickly selected, so that the processing result given is more accurate to meet the needs of the problem given by the user, and the communication efficiency and quality are improved.

[0102] The first processing result obtained by the question and answer model is sent to the information confirmation page for agent intervention processing to ensure the quality of the question and answer and provide protection in the case of abnormal processing. This human-machine cooperation mode not only plays the automation advantage of the question and answer model, but also retains the flexibility of manual intervention.

[0103] In some optional embodiments, the feedback information includes one or more of preset information and input information, the information confirmation page is configured to include one or more of the preset information, and the information confirmation page is further configured to have an information input control for obtaining the input information.

[0104] The information confirmation page can be pre-configured with one or more preset information, and the agent personnel can determine the corresponding feedback information by interacting with the preset information. The preset information can be a processing strategy, a processing method, and the like for some problems.

[0105] Further, the information confirmation page is also provided with an information input control, and the agent personnel directly gives input information by interacting with the information input control, and accordingly obtains the feedback information described above. The input information can be obtained by modifying the first processing result, or a processing strategy is given again, and the like, which is not limited herein.

[0106] The feedback information is obtained by one or more of the preset information and the input information, that is, a flexible feedback mechanism is provided, and these feedback mechanisms work together to further improve the question and answer effect.

[0107] In some optional embodiments, the question and answer method described above further includes: if feedback information for the first processing result is not obtained within a preset time length, the first processing result is pushed to the conversation page for display.

[0108] If feedback information is not obtained within a preset time length after the first processing result is pushed to the information processing page, the first processing result is considered to be correct by default, and can be directly pushed to the conversation page. That is, the first processing result is pushed to the conversation page for display. The specific value of the preset time length is set according to actual needs, which is not limited herein.

[0109] If no feedback is received regarding the first processing result within the preset time period, the first processing result will be proactively pushed to the chat page to avoid poor user experience caused by long waiting times.

[0110] As a specific application embodiment of this application, combined with Figure 3 In the application scenario shown, terminal device 310 is a mobile phone, application 301 is a question-and-answer application installed on the mobile phone, and interface 302 is the interface provided by the question-and-answer application. User 330 interacts with interface 302 to give a first question. Correspondingly, the mobile phone obtains the first question, extracts the context information of the first question, and sends it to server 320 for processing. Server 320 is equipped with the question-and-answer model described in this embodiment. The question-and-answer model performs semantic analysis on the first information (the first question and its context information) to determine the first information processing tool to be called, and calls the first information processing tool in sequence to obtain the corresponding associated information. If, during this process, user 330 again interacts with interface 302 to give a second question, similarly, mobile phone 310 sends the second question and its context information to server 320. In processing the second question, the question-and-answer model inherits the associated storage content stored in the first question and uses the inherited content to process the second question. During the processing, the question-and-answer model sends the processing result to the information confirmation page on the agent side. Figure 3 (Not shown in the image) If feedback information regarding the processing result is received, the question-answering model is triggered to reprocess the corresponding question based on the feedback information; if no feedback information is received within the preset time, the processing result is directly pushed to the user's interface 302 for display.

[0111] The question-answering method provided in this embodiment effectively solves the problems of information redundancy and insufficiency through the encapsulation and invocation mechanism of information processing tools. This allows the question-answering model to obtain information on demand, eliminating the need to input all data at once and significantly reducing the waste of computing resources. Furthermore, when specific information is required, the question-answering model can proactively invoke the corresponding information processing tools to ensure that the necessary knowledge is acquired to solve the problem.

[0112] Furthermore, because question-answering models can inherit previous thoughts and results, they do not need to reprocess all information each time, thus providing a more coherent and natural service experience. Since question-answering models can autonomously decide which information processing tools to use based on the content of the conversation, this question-answering method not only does not need to distinguish between scenarios but can also handle more complex scenarios.

[0113] In the present embodiment, a question and answer device is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and will not be described again. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, implementation of hardware, or a combination of software and hardware, is also possible and contemplated.

[0114] The present embodiment provides a question and answer device, as shown in Figure 9 comprises:

[0115] An information obtaining module 901 is configured to obtain first information, the first information comprising a first question and context information of the first question.

[0116] An information processing module 902 is configured to process the first information based on a question and answer model to obtain a first processing result, the question and answer model being configured to call a first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result.

[0117] In some optional embodiments, the information processing module 902 comprises:

[0118] A semantic analysis unit is configured to perform semantic analysis on the first information by the question and answer model to determine the first information processing tool to be called.

[0119] A tool calling unit is configured to call the first information processing tool by the question and answer model to obtain the associated information.

[0120] An information processing unit is configured to process the first information based on the associated information by the question and answer model to obtain the first processing result.

[0121] In some optional embodiments, if the first information processing tool to be called is multiple, the question and answer model is further configured to determine a calling sequence between the first information processing tools.

[0122] In some optional embodiments, further comprising:

[0123] An association storage module is configured to, if the associated information corresponding to the first information processing tool is obtained, store the calling of the first information processing tool and the associated information in association, and the content of the associated storage is used for inheritance by the question and answer model when processing a second question.

[0124] In some optional embodiments, the question and answer model is configured to be capable of calling one or more of the following tool types:

[0125] A first tool type is configured to verify whether a corresponding interface can be executed successfully.

[0126] A second tool type is configured to query information.

[0127] a third tool type for session summarization and session planning.

[0128] The tool types that the question-answering model can invoke are configured to include one or more information processing tools.

[0129] In some optional embodiments, the method comprises:

[0130] The information inheritance module is configured to, if a second question is obtained in the process of processing the first information, process the second question by inheriting the associated stored content obtained by processing the first information, the associated stored content comprising the invocation of the first information processing tool and the associated information corresponding to the first information processing tool.

[0131] In some optional embodiments, the method comprises:

[0132] The prompt information generation module is configured to, if the invocation duration of the first information processing tool exceeds a preset threshold, trigger the question-answering model to generate first prompt information based on the first information.

[0133] In some optional embodiments, the method comprises:

[0134] The information sending module is configured to send the first processing result to the information confirmation page for display.

[0135] The feedback processing module is configured to, if feedback information is obtained for the first processing result, trigger the question-answering model to process the first information based on the feedback information to obtain an updated first processing result.

[0136] In some optional embodiments, the feedback information comprises one or more of preset information and input information, the information confirmation page is configured to include one or more of the preset information, and the information confirmation page is further configured to have an information input control, which is configured to obtain the input information.

[0137] In some optional embodiments, the method further comprises:

[0138] The information pushing module is configured to, if no feedback information is obtained for the first processing result within a preset time duration, push the first processing result to the session page for display.

[0139] The question-answering device provided by the embodiments of the present disclosure can execute the question-answering method provided by any of the embodiments of the present disclosure, and has the corresponding function modules and beneficial effects of the execution method. The further function description of each of the above modules and units is the same as that of the corresponding embodiments, which will not be repeated here.

[0140] Figure 10 A structural schematic diagram of an electronic device provided by the embodiments of the present disclosure.

[0141] Reference will now be made to the following description Figure 10 which shows a structural schematic diagram suitable for being used to implement an electronic device in the embodiments of the present disclosure. The electronic device can include a processor (such as a central processor, a graphics processor, etc.) 1001, which can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for operation of the electronic device are also stored. The processor 1001, the ROM 1002, and the RAM 1003 are connected to each other through a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0142] Generally, the following devices can be connected to the I / O interface 1005: an input device 1006 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 1007 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage 1008 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 10 The electronic device with various devices is shown, but it should be understood that all the shown devices are not required to be implemented or possessed, and more or less devices can be alternatively implemented or possessed.

[0143] In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, the embodiments of the present disclosure include a computer program product including a computer program carried on a non-transitory computer readable medium, the computer program containing program codes for executing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network through the communication device 1009, or installed from the storage 1008, or installed from the ROM 1002. When the computer program is executed by the processor 1001, the above-mentioned functions defined in the question and answer method of the embodiments of the present disclosure are performed.

[0144] Figure 10 The electronic device shown is only an example and should not bring any limitation to the functions and use range of the embodiments of the present disclosure.

[0145] The embodiments of the present application further provide a computer readable storage medium, and the method according to the embodiments of the present application can be implemented in hardware, firmware, or recorded in a storage medium, or be implemented as computer code originally stored in a remote storage medium or a non-transitory machine readable storage medium and downloaded through a network and stored in a local storage medium, so that the method described herein can be processed by such software on a storage medium using a general purpose computer, a special purpose processor, or programmable or special hardware. The storage medium can be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid state disk, etc. Further, the storage medium can also include a combination of the above-mentioned types of memories. It can be understood that the computer, the processor, the microprocessor controller, or the programmable hardware includes a storage component that can store or receive software or computer code, when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the question and answer method shown in the above embodiments is implemented.

[0146] Part of the present application can be applied as a computer program product, for example, computer program instructions, when executed by a computer, through the operation of the computer, the method and / or technical solutions according to the present application can be invoked or provided. Those skilled in the art should understand that the form of computer program instructions in a computer readable medium includes but is not limited to source files, executable files, installation package files, etc. Correspondingly, the way of executing computer program instructions by computer includes but is not limited to: the computer directly executes the instructions, or the computer compiles the instructions and then executes the corresponding compiled program, or the computer reads and executes the instructions, or the computer reads and installs the instructions and then executes the corresponding installed program. Here, the computer readable medium can be any available computer readable storage medium or communication medium accessible to the computer.

[0147] Although the embodiments of the present application are described in conjunction with the accompanying drawings, various modifications and changes can be made by those skilled in the art without departing from the spirit and scope of the present application, and such modifications and changes fall within the scope defined by the appended claims.

Claims

1. A question and answer method, characterized by, The method comprises: obtaining first information, the first information comprising a first question and context information of the first question; processing the first information based on a question and answer model to obtain a first processing result, the question and answer model being configured to call a first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result.

2. The method of claim 1, wherein, The processing of the first information based on the question and answer model to obtain the first processing result comprises: the question and answer model performing semantic analysis on the first information to determine the first information processing tool to be called; the question and answer model calling the first information processing tool to obtain the associated information; the question and answer model processing the first information based on the associated information to obtain the first processing result.

3. The method of claim 2, wherein, If there are multiple first information processing tools to be called, the question and answer model is further configured to determine the calling sequence of the first information processing tools.

4. The method of claim 2, wherein, The method further comprises: if the associated information corresponding to the first information processing tool is obtained, the calling of the first information processing tool and the associated information are stored in association, and the content of the associated storage is used for inheritance by the question and answer model when processing a second question.

5. The method of claim 1, wherein, The question and answer model is configured to call one or more of the following tool types: a first tool type for verifying whether a corresponding interface can be successfully executed; a second tool type for querying information; a third tool type for session summarization and session planning; wherein the tool types that can be called by the question and answer model are configured to comprise one or more information processing tools.

6. The method of claim 1, wherein, The method comprises: if a second question is obtained during the processing of the first information, the question and answer model processes the second question by inheriting the content of the associated storage obtained by processing the first information, the content of the associated storage comprising the calling of the first information processing tool and the associated information corresponding to the first information processing tool.

7. The method of claim 1, wherein, The method comprises: if the calling duration of the first information processing tool exceeds a preset threshold, the question and answer model is triggered to generate first prompt information based on the first information.

8. The method of claim 1, wherein, The method comprises: sending the first processing result to an information confirmation page for display; if feedback information is obtained for the first processing result, the question and answer model is triggered to process the first information based on the feedback information to obtain an updated first processing result.

9. The method of claim 8, wherein, The feedback information comprises one or more of preset information and input information, the information confirmation page is configured to comprise one or more of the preset information, and the information confirmation page is further configured to have an information input control, which is used to obtain the input information.

10. The method of claim 8, wherein, The method further comprises: if no feedback information is obtained for the first processing result within a preset time duration, the first processing result is pushed to a session page for display.

11. A question answering apparatus characterized by comprising: The method comprises: an information obtaining module configured to obtain first information, the first information comprising a first question and context information of the first question; An information processing module is configured to process the first information based on a question and answer model to obtain a first processing result. The question and answer model is configured to call a first information processing tool based on the first information to obtain associated information, and process the first information based on the associated information to obtain the first processing result.

12. An electronic device, comprising: The application further provides a computer readable storage medium having stored computer instructions, and a computer program product. The application further provides a computer readable storage medium having stored computer instructions, and a computer program product.

13. A computer-readable storage medium, characterized in that, The application further provides a computer readable storage medium having stored computer instructions, and a computer program product.

14. A computer program product, characterised in that, The application further provides a computer readable storage medium having stored computer instructions, and a computer program product.

Images