Intelligent agent question and answer method, device and equipment based on large model, medium and product

Abstract

The invention discloses an agent question answering method and device based on a large model, equipment, a medium and a 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 in response to an input instruction of a question answering task, a first question is displayed in a first area of an interaction page; in response to the first question, displaying a first reply plan for the first question in a second area of the interaction page; in response to a modification instruction for the first reply plan, triggering modification of the first reply plan and displaying a second reply plan in the second area; and displaying an answer obtained based on the second reply plan in the first area. The method can ensure the accuracy of the answer obtained based on the second reply plan.

Description

Technical Field

[0001] The present disclosure relates to the fields of data processing technology, artificial intelligence technology, large model technology, and large language model technology, and specifically to intelligent agent question-answering methods, devices, equipment, media, and products based on large models. Background Art

[0002] The interaction process between a user and an agent involves the user asking a question, and the agent then responds after considering the question. The interactive page typically displays the agent's thinking process and final response. The accuracy of the agent's response depends on the agent's thinking process. Therefore, a large-scale model-based agent question-answering method is needed to ensure accurate responses. Summary of the Invention

[0003] In view of this, the present disclosure provides a large-model-based intelligent agent question-answering method, device, equipment, medium and product to solve the problem of intelligent agent question-answering accuracy.

[0004] In a first aspect, the present disclosure provides a large-model-based intelligent agent question-answering method, comprising:

[0005] In response to an input instruction of a question-and-answer task, displaying a first question in a first area of an interactive page;

[0006] In response to the first question, displaying a first response plan for the first question in a second area of the interactive page;

[0007] In response to a modification instruction for the first recovery plan, triggering modification of the first recovery plan and displaying a second recovery plan in the second area;

[0008] An answer based on the second response plan is displayed in the first area.

[0009] In a second aspect, the present disclosure provides an intelligent agent question-answering device based on a large model, comprising:

[0010] A first display module is configured to display a first question in a first area of an interactive page in response to an input instruction of a question-and-answer task;

[0011] a second display module, configured to display a first response plan for the first question in a second area of the interactive page in response to the first question;

[0012] a plan modification module, configured to trigger modification of the first response plan and display a second response plan in the second area in response to a modification instruction for the first response plan;

[0013] The third display module is configured to display an answer obtained based on the second response plan in the first area.

[0014] In a third aspect, the present disclosure provides an electronic device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, computer instructions being stored in the memory, and the processor executing the computer instructions to execute the large-model-based intelligent agent question-answering method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0015] In a fourth aspect, the present disclosure provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the large-model-based intelligent agent question-answering method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0016] In a fifth aspect, the present disclosure provides a computer program product, comprising computer instructions for enabling a computer to execute the large-model-based intelligent agent question-answering method of the above-mentioned first aspect or any corresponding embodiment thereof.

[0017] The large-model-based intelligent agent question-answering method provided by the embodiment of the present disclosure displays a first question in the first area of an interactive page in response to an input instruction of a question-answering task; displays a first reply plan for the first question in the second area of the interactive page in response to the first question; triggers the modification of the first reply plan and displays the second reply plan in the second area in response to a modification instruction for the first reply plan; and displays the answer derived based on the second reply plan in the first area. After obtaining the first question to be answered, the method first generates and displays a first reply plan for the first question. The first reply plan is used to represent the thinking steps for the first question. After displaying the first reply plan, a modification function for the first reply plan is also provided. The first reply plan is modified in an interactive manner to obtain a second reply plan, which can ensure the accuracy of the thinking process for the first question, and thus can ensure the accuracy of the answer derived based on the second reply plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the related technologies, the following briefly introduces the drawings required for use in the specific embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 is a schematic diagram of an application scenario according to an embodiment of the present disclosure;

[0020] Figure 21 is a first flow chart of a large-model-based intelligent agent question-answering method according to an embodiment of the present disclosure;

[0021] Figure 3 This is a first schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0022] Figure 4 2 is a second flow chart of the large-model-based intelligent agent question-answering method according to an embodiment of the present disclosure;

[0023] Figure 5 is a second schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0024] Figure 6 is a third schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0025] Figure 7 is a fourth schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0026] Figure 8 is a fifth schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0027] Figure 9 is a sixth schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0028] Figure 10 is a seventh schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0029] Figure 11 is an eighth schematic diagram of a question-and-answer page according to an embodiment of the present disclosure;

[0030] Figure 12 is a structural block diagram of an intelligent agent question-answering device based on a large model according to an embodiment of the present disclosure;

[0031] Figure 13 Schematic diagram of the hardware structure of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0032] To make the purpose, technical solutions, and advantages of the embodiments of the present disclosure more clear, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present disclosure.

[0033] It is understandable that before using the technical solutions disclosed in the various embodiments of this disclosure, the type, scope of use, usage scenarios, etc. of the personal information involved in this disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0034] For example, in response to a user's active request, a prompt message is sent to the user to clearly inform the user that the operation requested will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the electronic device, application, server, storage medium, or other software or hardware that performs the operations of the disclosed technical solution based on the prompt message.

[0035] As an optional but non-limiting implementation, in response to receiving a user's active request, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. Furthermore, the pop-up window may also contain a selection control for the user to select "agree" or "disagree" to provide personal information to the electronic device.

[0036] It is understandable that the above notification and user authorization process are merely illustrative and do not limit the implementation of the present disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0037] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) must comply with the requirements of relevant laws, regulations and relevant provisions.

[0038] It should be noted that the specific implementation methods / technical features of each embodiment of the present application can be arbitrarily combined without violating the principles of the invention.

[0039] In related technologies, question-and-answer tasks typically involve displaying the thought process for a question on an interactive page after the question is given. Once the thought process is complete, the answer is displayed. During this process, the thought process cannot be modified; it simply serves a display purpose. If there's a problem with the thought process, the resulting answer may not meet expectations.

[0040] Based on this, the disclosed embodiments provide a large-scale model-based intelligent question-answering method. After displaying a response plan for a question, a modification function is provided for the response plan. By modifying the response plan, the response plan is continuously updated, ensuring the accuracy of the response plan and, accordingly, improving the accuracy of the obtained answer.

[0041] As used in the embodiments of the present application, the term "model" can learn the association between the corresponding input and output from the training data, so that after the training is completed, the corresponding output can be generated for a given input. The generation of the model can be based on machine learning technology, etc., taking deep learning as an example, deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs by using multiple layers of processing units. In the embodiments of the present application, the model can also be referred to as a machine learning model, a machine learning network or a network, and these terms can be used interchangeably in this article. Among them, a model can also include different types of processing units or networks.

[0042] As an optional application scenario of the embodiment of the present disclosure, Figure 1 As shown, the terminal device 110 has an application 101 installed therein, and the user 130 can interact with the application 101 through the terminal device 110 and / or an access device of the terminal device 110 .

[0043] For example, application 101 can be any application that can provide question-answering related services. For example, application 101 can be a question-answering interactive application, such as a text-to-text application, a picture-to-text application, etc. Figure 1 In the application scenario shown, if the application 101 is active, the terminal device 110 can present the interface 102 of the application 101. The interface 102 can include various pages that the application 101 can provide, such as an interaction page, a setting page, a query page, and the like.

[0044] In some embodiments, the terminal device 110 is in communication with the server 120 to provide services for the application 101. The terminal device 110 can be a mobile terminal, a fixed terminal, or a portable terminal, including but not limited to a mobile phone, a desktop computer, a laptop computer, a multimedia tablet, an e-book device, a gaming device, or any combination thereof, including accessories and peripherals of these devices, or any combination thereof. In some embodiments, the terminal device 110 can also support any type of interface, and the server 120 can be any type of computing system or server that can provide computing capabilities, including but not limited to mainframes, edge computing nodes, computing devices in cloud environments, and the like.

[0045] It should be noted that Figure 1 This is merely an example of an application scenario and does not limit the scope of protection of the present disclosure.

[0046] The embodiments of the present disclosure will be described below with reference to the accompanying drawings. It should be understood that the pages shown in the accompanying drawings are merely examples, and various page designs may actually exist. The various graphic elements in the page may have different arrangements and different visual representations, one or more of which may be omitted or replaced, and one or more other elements may also exist, which are not limited in the embodiments of the present disclosure. In addition, the embodiments are described below mainly with respect to the terminal device 110. It should be understood that the actions described with respect to the terminal device 110 may be performed by the application 101 on the terminal device 110, or may be performed by the application 101 in collaboration with its service end (e.g., server 120).

[0047] According to an embodiment of the present disclosure, an embodiment of an intelligent agent question-answering method based on a large model is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0048] In this embodiment, a large model-based intelligent agent question-answering method is provided, which can be used in the above-mentioned terminal device. Figure 2 is a flow chart of the intelligent agent question answering method based on a large model according to an embodiment of the present disclosure, such as Figure 2 As shown, the process includes the following steps:

[0049] Step S201: In response to an input instruction of a question-and-answer task, a first question is displayed in a first area of an interactive page.

[0050] Question-and-answer tasks are characterized by users asking questions through interaction with an interactive page, and the agent then thinks about the questions and gives answers. Input instructions for question-and-answer tasks include, but are not limited to, text input instructions, voice input instructions, text and image input instructions, text and file input instructions, and so on, and are set according to actual needs. Accordingly, different types of input instructions correspond to different input methods. For example, a voice interaction control is displayed on the interactive page, and voice input instructions are obtained by interacting with the voice interaction control.

[0051] The interactive page represents the page provided by the Q&A application, with the first question displayed in the first area of the interactive page. The first question is obtained in response to the input instruction of the Q&A task. Corresponding to the input instruction described above, the first question includes, but is not limited to, text, text and an image, text and a file, and so on. For example, the first question may include only the question description, or include an image and the question description, or include a file and the question description, and so on.

[0052] The interactive page can be divided into regions according to the functions of the page, and the number of regions divided is not limited. Accordingly, the first region can be located in any region of the interactive page, and can be set according to actual needs.

[0053] For example, Figure 3 FIG. 3 shows a first schematic diagram of an interactive page 310. In the first area 311 of the interactive page 310, a conversation between a user 313 and an agent 314 is displayed. Figure 3 The figure also shows the identification of the user 313 and the agent 314. The identification can be represented by images or text, etc., and there is no limitation on this.

[0054] Step S202: In response to the first question, a first response plan for the first question is displayed in a second area of the interactive page.

[0055] After the first question is given, the terminal device displays the thinking process for the first question, i.e., the first response plan, in the second area of the interactive page in response to the first question. The location of the second area in the interactive page is set according to actual needs and is not limited here.

[0056] The first reply plan includes a plan outline and the specific implementation method of each part of the plan outline. For example, the plan outline includes 5 steps, and each step includes the specific implementation method of the step.

[0057] For example, Figure 3 As shown, a first response plan 315 is displayed in the second area 312 of the interactive page 310. The first response plan 315 includes four steps. The specific implementation of each step can be displayed by interacting with the corresponding step; alternatively, the specific implementation can be displayed during the execution of each step. There is no limitation on the display method for the specific implementation of each step, and it can be set according to actual needs.

[0058] Step S203 : In response to the modification instruction for the first recovery plan, triggering the modification of the first recovery plan and displaying the second recovery plan in the second area.

[0059] The first response plan in the interactive page also provides a modification function, which can modify the plan outline of the first response plan and the specific implementation of each step in the plan outline.

[0060] The first reply plan can be modified by editing the text, commenting on the corresponding content, or modifying the logic diagram of the first reply plan through the canvas, etc. Of course, other methods can also be used to modify the first reply plan. There is no limitation on the modification method of the first reply plan here, and it can be set according to actual needs.

[0061] After the first response plan is modified, a second response plan is obtained. It should be noted that the second response plan does not specifically refer to the response plan obtained after the first response plan is modified once, but generally refers to the second response plan that is finally determined after the first response plan is modified. Among them, the modification of the first response plan here can be 1 time, 2 times or multiple times, etc. It should be understood that the described two modifications to the first response plan can be modifications made on the basis of the response plan after the first modification to the first response plan, that is, each modification can be made on the basis of the previous modification, or it can be made on the basis of a previous modification. There is no limitation on the modification basis of each modification here.

[0062] The second reply plan obtained after modifying the first reply plan is displayed in the second area 312 of the interactive page 310. The second area 312 can be displayed on the interactive page 310 all the time, or can be hidden by interacting with the interactive control 316. The display style of the interactive control 316 is not limited to Figure 3 As shown, other display styles are possible and are not limited to them. The hidden display of the second area 312 can be a floating control folded into a preset size on the interactive page 310, and the display of the second area 312 can be expanded by interacting with the floating control; or the hidden display of the second area 312 can be a representation that it is not displayed on the interactive page 310, and the display of the second area 312 can be expanded by interacting with the corresponding functional control on the interactive page 310. There is no limitation on the hiding and displaying methods of the second area 312, and it can be set according to actual needs.

[0063] Step S204: Displaying an answer based on the second response plan in the first area.

[0064] After the second response plan is determined, you can think according to the thinking process given in the second response plan to obtain the corresponding answer, and display the obtained answer in the first area.

[0065] The answer display style can be text, table, slide or other formats of documents, etc., which can be displayed according to actual needs and is not limited here.

[0066] The large-model-based intelligent agent question-answering method provided in this embodiment, after obtaining the first question that needs to be answered, first generates and displays a first response plan for the first question. The first response plan is used to represent the thinking steps for the first question. After displaying the first response plan, it also provides a modification function for the first response plan. The first response plan is modified in an interactive manner to obtain a second response plan, which can ensure the accuracy of the thinking process for the first question, and thus ensure the accuracy of the answer obtained based on the second response plan.

[0067] In this embodiment, a large model-based intelligent agent question-answering method is provided, which can be used in the above-mentioned terminal device. Figure 4 is a flow chart of the intelligent agent question answering method based on a large model according to an embodiment of the present disclosure, such as Figure 4 As shown, the process includes the following steps:

[0068] Step S401: In response to the input instruction of the question-answering task, the first question is displayed in the first area of the interactive page. Figure 2 The step S201 of the illustrated embodiment is not limited in any way herein.

[0069] Step S402: In response to the first question, a first response plan for the first question is displayed in the second area of the interactive page. Figure 2 The step S202 of the illustrated embodiment is not limited herein.

[0070] Step S403 : In response to the modification instruction for the first recovery plan, triggering the modification of the first recovery plan and displaying the second recovery plan in the second area.

[0071] Exemplarily, the above step S403 includes:

[0072] Step S4031: determining a modification method in response to an interactive instruction for modifying a target control in the second area.

[0073] The second area displays a target modification control, and different target modification controls are used to represent different modification methods. Different target modification controls are represented by corresponding logos, and the specific styles correspond to the modification methods.

[0074] The target modification control can be a sub-control under the modification control. For example, a modification control is displayed in the second area, and sub-controls are displayed through interaction with the modification control, with the sub-controls corresponding to the modification methods. Of course, other methods can also be used to display the target modification control in the second area. There is no limitation on this method, and the specific method can be set according to actual needs.

[0075] After interacting with the target modification control, a modification method is determined accordingly. After the modification method is determined, the second area can display a style corresponding to the modification method to ensure that the modification operation can be performed.

[0076] For example, Figure 5 As shown, a modification control 501 is displayed in the second area of the interactive page, and the modification control 501 includes sub-controls corresponding to the modification methods, that is, the target modification control described above.

[0077] Step S4032: In response to the modification instruction corresponding to the modification method, triggering modification of the first reply plan.

[0078] After determining the modification method, the first response plan is interacted with based on the modification method to generate modification instructions, which are used to trigger modifications to the first response plan, such as comments on the first response plan or logical adjustments to the first response plan.

[0079] In some optional implementations, the above step S4032 includes:

[0080] Step a1: If the modification mode is the comment mode, in response to the comment instruction for the first plan content in the first reply plan, the comment content is displayed at a corresponding position of the first plan content.

[0081] Step a2: In response to the instruction to regenerate the plan, triggering modification of the first recovery plan.

[0082] If the determined modification method is a comment method, the first plan content is commented on by interacting with the first plan content in the first reply plan, generating a comment instruction for the first plan content, and correspondingly displaying the comment content at the corresponding position of the first plan content. The comment content is used to indicate the modification method of the first plan content.

[0083] For example, Figure 6 As shown, if the selected target modification control is the comment control 601, the modification of the first reply plan is given in the form of comment. Figure 6 In the example, a comment 602 for the first plan content is given. Based on the location of the comment 602, the first plan content targeted by the comment can be determined. Specifically, Figure 6 The comment content 602 is located at the second step in the first reply plan, which indicates that the comment content 602 is a modification description given for the second step.

[0084] After the first response plan is generated, the execution status of the response plan is displayed in the first area. For example, Figure 7As shown, "Reply plan confirmation" is displayed in the reply information given by the intelligent agent, which is used to indicate that the first reply plan is waiting for confirmation. The confirmation of the first reply plan can be achieved through interaction with the start execution control 701 displayed in the second area, or the timing can be started after the first reply plan is generated to count the continuous unmodified time of the first reply plan. If the time reaches the preset time, it indicates that the first reply plan is confirmed and the processing of the first reply plan is started. For example, the prompt information 702 given by the intelligent agent includes a text description of "Reply plan confirmation" and a timer 7021. The timer 7021 is used to trigger the timing after the first reply plan is generated. If there is any modification to the first reply plan, the timing is stopped and the timing result is cleared. The timing can be started again after the modification. If the timing reaches the preset time, the execution of the modified reply plan is actively triggered.

[0085] After the first response plan is modified, Figure 8 As shown, the modification of the first reply plan can be triggered by interacting with the regeneration plan control 801 in the second area, and accordingly, the modification result is displayed in the second area. It should be noted that the triggering method for regenerating the plan can be through Figure 8 The regeneration plan control 801 shown can also be triggered by a shortcut key, or by other methods. There is no limitation on this, and it can be set according to actual needs.

[0086] If the modification method is the comment method, you can comment on the first plan content in the first reply plan and display the comment content in the corresponding position. The comment content is used to indicate the modification of the first plan content. Accordingly, the modification of the first reply plan is triggered by giving an instruction to regenerate the plan.

[0087] In some optional implementations, the above step S4032 includes:

[0088] Step b1: If the modification mode is canvas mode, the first response plan is displayed in canvas form.

[0089] Step b2: Displaying the modified first response plan on the canvas in response to the modification instruction for the first response plan.

[0090] Step b3: In response to the instruction to regenerate the plan, triggering the effectiveness of the modified first response plan.

[0091] If the selected modification method is canvas, the first response plan is displayed in canvas format. By interacting with the first response plan on the canvas, modifications are made to the first response plan, modification instructions are generated, and the modified first response plan is displayed accordingly. Regeneration of the response plan is triggered by interacting with the Regenerate Plan control, using shortcuts, or other methods. Regeneration instructions are generated, which in turn triggers the implementation of the modified first response plan.

[0092] By modifying the first response plan through the canvas modification method, the processing logic of the first response plan can be flexibly adjusted, and the operation is flexible and simple.

[0093] Step S4033: Display the second response plan in the second area.

[0094] After one or more modifications to the first response plan, a final response plan is obtained, which is referred to as the second response plan. Accordingly, the second response plan is displayed in the second area. For a description of the second response plan, see above. Figure 2 The description in the illustrated embodiment will not be repeated here.

[0095] Step S404: Display the answer based on the second response plan in the first area. Figure 2 The description of step S204 of the illustrated embodiment will not be repeated here.

[0096] The intelligent agent question-answering method based on a large model provided in this embodiment displays a target modification control for the response plan in the second area, determines the modification method by interacting with the target modification control, and triggers the modification of the first response plan according to the corresponding modification method, that is, it can provide multiple modification methods to meet the modification needs in different scenarios.

[0097] In some optional implementations, the above-mentioned large-model-based agent question-answering method further includes:

[0098] Step c1, obtaining the complexity of the first problem.

[0099] Step c2: If the complexity characterization first question requires a response based on the generated plan, then a step of displaying a first response plan for the first question in the second area of the interactive page in response to the first question is executed.

[0100] Step c3: If the complexity indicates that the first question can be answered directly, the answer to the first question is displayed in the first area.

[0101] After a first question is given, a semantic analysis of the first question is performed to determine its complexity. The complexity can be expressed as a probability value or as a corresponding text representation. The display format of the complexity is not limited and can be set based on actual needs. For example, after the first question is obtained, a semantic analysis of the first question is performed using a language model to output the complexity of the first question.

[0102] Complexity represents the difficulty of answering the first question. If the complexity value is large, it means that the first question needs to be answered based on the generated plan. Then execute the above Figure 2 Step S202 of the embodiment shown, or Figure 4 Step S402 of the illustrated embodiment. If the complexity value is small, indicating that the first question can be answered directly without generating a response plan, the answer to the first question is displayed in the first area. Determining whether the complexity value is large or small can be done by comparing the complexity value with a preset threshold. If the complexity value is greater than the preset threshold, the complexity value is large; otherwise, the complexity value is small.

[0103] Whether a response plan needs to be generated is determined based on the complexity of the first question. A first response plan is generated for complex first questions; and the answer to the first question is directly displayed for non-complex questions, thereby improving question-answering efficiency.

[0104] In some optional implementations, the above-mentioned large-model-based agent question-answering method further includes:

[0105] Step d1: Record the first plan version number of the first reply plan.

[0106] Step d2: After the first response plan is modified, the second plan version number of the second response plan is recorded to obtain the first plan version number and the second plan version number corresponding to the first problem.

[0107] The version number of the first response plan generated initially is V1. Each subsequent modification to the response plan will record a version number, for example, incrementing it in the order of V2, V3, ... Vn. Specifically, the first response plan corresponds to the first plan version number, and the second response plan corresponds to the second plan version number. Each time a response plan is modified, the corresponding plan version number can be obtained, and the plan version number serves as the unique identifier of the response plan.

[0108] For the first question, the corresponding version of the response plan and the plan version number are obtained based on the number of times the response plan has been modified. It should be noted that the phrase "after modification to the response plan" indicates that the modification to the response plan has taken effect. If the modification to the response plan is invalid, there is no need to record the version number. For example, the modification to the response plan described may take effect by triggering the regeneration of the response plan after the modification.

[0109] For the same issue, each modification to the response plan generates a corresponding plan version number. That is, different response plans are distinguished by the plan version number, making it easier to view the corresponding response plan later.

[0110] In some optional implementations, the above-mentioned large-model-based agent question-answering method further includes:

[0111] Step e1: In response to a selection instruction for the plan description in the first area, display a plan version number selection control in the second area.

[0112] Step e2: determining the target plan version number in response to an interactive instruction on the plan version number selection control in the second area.

[0113] Step e3: Displaying the answer obtained based on the response plan corresponding to the target plan version number in the first area.

[0114] The first area shows the question-answer pair, that is, the response given by the agent to the question given by the user. As described above, the agent also gives a description of the execution status of the response plan. For example, Figure 9 As shown, the first area displays the agent's reply 901 and the execution status of the reply plan 904. After the reply plan is executed, the answer 905 to the first question is displayed in the first area.

[0115] Therefore, the first area displays the plan description and the answer for the reply plan, and then, by interacting with the plan description, the plan version number selection control is displayed in the second area. For example, Figure 9 The plan version number selection control 902 is shown in FIG. By interacting with the plan version number selection control 902, the display of the plan version number for the first question is triggered, and the target plan version number to be selected is determined by the interactive instruction with the plan version number selection control.

[0116] After the target plan version number is determined, it means that you want to get the answer of the response plan corresponding to the target plan version number. Therefore, after selecting the target plan version number, you can directly trigger the execution of the response plan corresponding to the target plan version number to get the answer; or after selecting the target plan version number, you can trigger the execution of the response plan corresponding to the target plan version number to get the answer. Figure 9The interaction of the start execution control 903 shown triggers the execution of the corresponding reply plan to obtain the answer, etc. There is no limitation on this, and it can be set according to actual needs.

[0117] For example, Figure 9 As shown, if the selected target plan version number is V2, the response plan of version V2 can be directly executed, and the answer obtained by executing the response plan will be displayed in the first area.

[0118] By selecting the plan version number, you can determine the answer to the response plan of the target plan version number you currently want to view, making it easier to compare the answers to different response plans.

[0119] In some optional implementations, the above-mentioned large-model-based agent question-answering method further includes:

[0120] Step f1: In response to a selection instruction for an answer obtained from a second response plan, detailed information of the answer is displayed in a second area.

[0121] Step f2: In response to the selection instruction of the first answer version number in the second area, detailed information of the answer corresponding to the first answer version number is displayed in the second area, where the answer version number corresponds to the reply plan.

[0122] As described above, for the same question, different response plan versions can generate corresponding answers. Therefore, to provide answers for different response plan versions, the answers generated by each response plan are distinguished by version number. That is, the same question is associated with one or more response version numbers, and each response version number corresponds to a single answer.

[0123] When displaying the answer in the first area, you can display only the brief content of the answer or the detailed content of the answer. By selecting the corresponding answer in the first area, the detailed information of the answer is displayed in the second area.

[0124] Furthermore, the answer version number is also displayed in the second area. Specifically, the first answer version number to be displayed is determined by the selection instruction of the first answer version number in the second area, and accordingly, the detailed information of the answer corresponding to the first answer version number is displayed in the second area. For example, Figure 10 As shown, the answer 1001 to the first question is displayed in the first area. By interacting with the answer 1001, an answer version number selection control 1002 is displayed in the second area. By interacting with the answer version number selection control 1002, the first answer version number is selected. Accordingly, the answer corresponding to the first answer version number is displayed in the second area. In other words, by setting the answer version number, it is possible to review historical answers to the same question.

[0125] The answers given to the same question under different response plans are recorded with corresponding answer version numbers. By supporting the switching of answer version numbers, different answers can be viewed.

[0126] In some optional implementations, the above-mentioned large-model-based agent question-answering method further includes:

[0127] Step g1: Before receiving an instruction to start executing the plan, a prompt message for plan confirmation is displayed in the first area.

[0128] Step g2: If an instruction to start executing the plan is received, a prompt message indicating that the plan is being executed is displayed in the first area.

[0129] As described above, the first area of the interactive page displays a description of the execution status of the reply plan. Specifically, before receiving an instruction to start execution of the plan, the execution status of the reply plan is "Plan Confirmation", and accordingly, a prompt message indicating that the plan has been confirmed is displayed in the first area. If an instruction to start execution of the plan is received, the execution status of the reply plan is "Plan Execution", and accordingly, a prompt message indicating that the plan has been executed is displayed in the first area.

[0130] The linked display of the first area and the second area makes it easier to intuitively perceive the current execution status.

[0131] In some optional implementations, the above-mentioned large-model-based agent question-answering method further includes:

[0132] Step h1: Display the resource library in the third area of the interactive page, where the resource library includes file resources.

[0133] Step h2: In response to the selection instruction for the file resource, an identifier of the target file resource is displayed in the first area, and the target file resource is used to assist in answering the first question.

[0134] The third area of the interactive page also displays a resource library, which includes file resources. This resource library can also be a personal knowledge space, belonging to a private domain. Before instructing the agent to respond to the first question, it can first interact with the resource library to select a file resource and obtain a target file identifier. The target file represents the knowledge needed to respond to the first question.

[0135] For example, Figure 11 As shown, a resource library 1101 is displayed in the third area, and a target file resource 1103 is selected by interacting with the resource library. Accordingly, an identifier 1104 of the target file resource is displayed in the first area.

[0136] The third area also displays an add resource control 1102. If the file resources in the resource library do not meet the requirements for answering the question, interacting with add resource control 1102 triggers the addition of file resources. The source of the file resources can be local file resources, favorite file resources, etc., and there is no limitation on the source of the file resources.

[0137] After being given the target file resources, the agent combines the target file resources and public domain resources to obtain the answer to the first question.

[0138] When performing question-answering tasks, file resources in the resource library are provided to assist in answering the first question, ensuring that the answer obtained meets the needs of the current scenario.

[0139] As a specific application example of the present disclosure, Figures 1 to 11 The content shown triggers the operation of the interactive application and displays the interactive page. In the resource library 1101 in the third area of the interactive page, the target file resource is determined by interacting with the file resource 1103 in the resource library 1101, and accordingly, the identifier 1104 of the target file resource is displayed in the first area. The user gives a first question by interacting with the first area of the interactive page, and accordingly, the first reply plan 315 is displayed in the second area 312. By commenting on the first reply plan 315 and triggering the regeneration of the reply plan, the second reply plan is determined and displayed. In this process, the intelligent agent gives the execution status of the reply plan, for example, the reply status is being confirmed. By triggering the execution of the second reply plan, the intelligent agent gives the execution status of the reply plan, for example, the status is being executed. After the execution of the reply plan is completed, the corresponding answer is displayed in the first area of the interactive page.

[0140] Response plans for the same question have corresponding response plan version numbers, and each response plan has a corresponding answer version number. By interactively selecting a response plan version number, you can view the corresponding response plan and display the corresponding answer. By interactively selecting an answer version number, you can view the corresponding answer.

[0141] In this embodiment, a large-scale model-based intelligent agent question-answering device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation methods. The details that have been described will not be repeated here. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation by hardware, or a combination of software and hardware, is also possible and conceivable.

[0142] This embodiment provides an intelligent agent question-answering device based on a large model, such as Figure 12 Shown, including:

[0143] The first display module 1201 is configured to display a first question in a first area of an interactive page in response to an input instruction of a question-and-answer task.

[0144] The second display module 1202 is configured to display a first response plan for the first question in a second area of the interactive page in response to the first question.

[0145] The plan modification module 1203 is configured to trigger modification of the first response plan and display the second response plan in the second area in response to a modification instruction for the first response plan.

[0146] The third display module 1204 is configured to display an answer obtained based on the second response plan in the first area.

[0147] In some optional implementations, the plan modification module 1203 includes:

[0148] The first response unit is configured to determine a modification method in response to an interactive instruction for modifying a target control in the second area.

[0149] The second response unit is configured to trigger modification of the first response plan in response to a modification instruction corresponding to the modification method.

[0150] The first display unit is configured to display the second response plan in the second area.

[0151] In some optional embodiments, the second response unit includes:

[0152] The first responding subunit is configured to display the comment content at a corresponding position of the first plan content in response to a comment instruction on the first plan content in the first reply plan if the modification mode is the comment mode.

[0153] The second response subunit is configured to trigger modification of the first response plan in response to an instruction to regenerate the plan.

[0154] In some optional embodiments, the second response unit includes:

[0155] The first display subunit is configured to display the first reply plan in a canvas format if the modification mode is a canvas format.

[0156] The second display subunit is configured to display the modified first response plan in the canvas in response to the modification instruction for the first response plan.

[0157] The third response subunit is configured to trigger the effectiveness of the modified first response plan in response to the instruction to regenerate the plan.

[0158] In some optional embodiments, the method further includes:

[0159] The complexity acquisition module is used to obtain the complexity of the first problem.

[0160] The first execution module is configured to execute a step of displaying a first response plan for the first question in a second area of the interactive page in response to the first question if the complexity characterization first question requires a response based on the generated plan.

[0161] The second execution module is configured to display an answer to the first question in the first area if the complexity indicates that the first question can be answered directly.

[0162] In some optional embodiments, the method further includes:

[0163] The first recording module is used to record a first plan version number of the first reply plan.

[0164] The second recording module is used to record the second plan version number of the second response plan after the first response plan is modified, and obtain the first plan version number and the second plan version number corresponding to the first problem.

[0165] In some optional embodiments, the method further includes:

[0166] The first response module is configured to display a plan version number selection control in the second area in response to a selection instruction for the plan description in the first area.

[0167] The second response module is used to determine the target plan version number in response to the interactive instruction for the plan version number selection control in the second area.

[0168] The fourth display module is used to display an answer obtained based on the response plan corresponding to the target plan version number in the first area.

[0169] In some optional embodiments, the method further includes:

[0170] The third response module is configured to display detailed information of the answer in the second area in response to a selection instruction for the answer obtained from the second response plan.

[0171] The fourth response module is used to respond to the selection instruction of the first answer version number in the second area and display detailed information of the answer corresponding to the first answer version number in the second area, where the answer version number corresponds to the reply plan.

[0172] In some optional embodiments, the method further includes:

[0173] The first prompt module is used to display a prompt message of plan confirmation in the first area before receiving an instruction to start executing the plan.

[0174] The second prompt module is used to display prompt information in the first area according to the plan execution if an instruction to start the plan execution is received.

[0175] In some optional embodiments, the method further includes:

[0176] The fifth display module is used to display the resource library in the third area of the interactive page, where the resource library includes file resources.

[0177] The fifth response module is configured to display an identifier of a target file resource in the first area in response to a selection instruction for the file resource, where the target file resource is used to assist in answering the first question.

[0178] The intelligent agent question-answering device based on a large model provided by the embodiments of the present disclosure can execute the intelligent agent question-answering method based on a large model provided by any embodiment of the present disclosure, and has the functional modules and beneficial effects corresponding to the execution method. After obtaining the first question that needs to be answered, the device first generates and displays a first response plan for the first question. The first response plan is used to characterize the thinking steps for the first question. After displaying the first response plan, it also provides a modification function for the first response plan. The first response plan is modified in an interactive manner to obtain a second response plan, which can ensure the accuracy of the thinking process for the first question, and thus can ensure the accuracy of the answer obtained based on the second response plan. The further functional description of each of the above modules and units is the same as that of the corresponding embodiments above, and will not be repeated here.

[0179] Figure 13 A schematic structural diagram of an electronic device provided in an embodiment of the present disclosure.

[0180] The following specific reference Figure 13 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present disclosure. The electronic device may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 1301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1302 or a program loaded from a memory 1308 into a random access memory (RAM) 1303. Various programs and data required for the operation of the electronic device are also stored in the RAM 1303. The processor 1301, ROM 1302, and RAM 1303 are connected to each other via a bus 1304. An input / output (I / O) interface 1305 is also connected to the bus 1304.

[0181] Typically, the following devices may be connected to the I / O interface 1305: an input device 1306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 1307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1309. The communication device 1309 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 13 An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all of the devices shown, and more or fewer devices may be implemented or possessed instead.

[0182] In particular, 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 includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 1309, or installed from the memory 1308, or installed from the ROM 1302. When the computer program is executed by the processor 1301, the above-mentioned functions defined in the large model-based intelligent agent question-answering method of the embodiment of the present disclosure are performed.

[0183] Figure 13 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0184] The embodiments of the present disclosure also provide a computer-readable storage medium. The above-mentioned method according to the embodiments of the present disclosure can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or 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 a computer, a processor or hardware, the large model-based intelligent agent question-answering method shown in the above embodiment is implemented.

[0185] A portion of the present disclosure may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present disclosure through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes but is not limited to a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0186] Although the embodiments of the present disclosure have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A large-model-based intelligent agent question answering method, characterized in that: include: In response to an input instruction of a question-and-answer task, displaying a first question in a first area of an interactive page; In response to the first question, displaying a first response plan for the first question in a second area of the interactive page; In response to a modification instruction for the first recovery plan, triggering modification of the first recovery plan and displaying a second recovery plan in the second area; An answer based on the second response plan is displayed in the first area.

2. The method according to claim 1, characterized in that The triggering of modifying the first restoration plan and displaying the second restoration plan in the second area in response to the modification instruction for the first restoration plan includes: In response to an interactive instruction for modifying a target control within the second area, determining a modification method; In response to a modification instruction corresponding to the modification method, triggering modification of the first response plan; The second response plan is displayed in the second area.

3. The method according to claim 2, characterized in that The triggering of modification of the first response plan in response to the modification instruction corresponding to the modification method includes: If the modification mode is a comment mode, in response to a comment instruction for the first plan content in the first reply plan, displaying the comment content at a corresponding position of the first plan content; In response to the instruction to regenerate the plan, a modification of the first recovery plan is triggered.

4. The method according to claim 2, characterized in that The triggering of modification of the first response plan in response to the modification instruction corresponding to the modification method includes: If the modification mode is canvas mode, the first response plan is displayed in canvas form; displaying a modified first response plan in the canvas in response to a modification instruction for the first response plan; In response to the instruction to regenerate the plan, the modified first recovery plan is triggered to take effect.

5. The method according to claim 1, wherein Also includes: Obtaining the complexity of the first problem; If the complexity indicates that the first question requires a response based on a generated plan, performing the step of displaying a first response plan for the first question in a second area of the interactive page in response to the first question; If the complexity indicates that the first question can be answered directly, the answer to the first question is displayed in the first area.

6. The method according to claim 1, characterized in that Also includes: Recording the first plan version number of the first response plan; After the first response plan is modified, the second plan version number of the second response plan is recorded to obtain the first plan version number and the second plan version number corresponding to the first question.

7. The method according to claim 1, characterized in that Also includes: In response to a selection instruction for the plan description in the first area, displaying a plan version number selection control in the second area; In response to an interactive instruction on a plan version number selection control in the second area, determining a target plan version number; An answer obtained based on the response plan corresponding to the target plan version number is displayed in the first area.

8. The method according to claim 1, characterized in that Also includes: In response to a selection instruction for an answer derived from the second response plan, displaying detailed information of the answer in the second area; In response to a selection instruction of a first answer version number in the second area, detailed information of an answer corresponding to the first answer version number is displayed in the second area, where the answer version number corresponds to a reply plan.

9. The method according to claim 1, characterized in that Also includes: Before receiving an instruction to start executing the plan, displaying a prompt message for plan confirmation in the first area; If an instruction to start executing the plan is received, a prompt message indicating that the plan is being executed is displayed in the first area.

10. The method according to any one of claims 1 to 9, characterized in that Also includes: Displaying a resource library in a third area of the interactive page, wherein the resource library includes file resources; In response to a selection instruction for the file resource, an identifier of a target file resource is displayed in the first area, and the target file resource is used to assist in answering the first question.

11. An intelligent agent question-answering device based on a large model, characterized in that: include: A first display module is configured to display a first question in a first area of an interactive page in response to an input instruction of a question-and-answer task; a second display module, configured to display a first response plan for the first question in a second area of the interactive page in response to the first question; a plan modification module, configured to trigger modification of the first response plan and display a second response plan in the second area in response to a modification instruction for the first response plan; The third display module is configured to display an answer obtained based on the second response plan in the first area.

12. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the large model-based intelligent agent question-answering method described in one of claims 1 to 10 by executing the computer instructions.

13. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, which are used to enable a computer to execute the large model-based intelligent agent question-answering method described in one of claims 1 to 10.

14. A computer program product, characterized in that It includes computer instructions, which are used to cause a computer to execute the large model-based intelligent agent question answering method as described in one of claims 1 to 10.

Images