Result generation method and device, equipment and storage medium

Abstract

The invention belongs to the technical field of computers, and discloses a result generation method and device, equipment and a storage medium. The method comprises the following steps: acquiring query information; selecting target question and answer data from the related question and answer data corresponding to the question searching information; and generating question-asking result data based on the question-asking information and the target question-answering data. As the question-asking data with the highest correlation degree with the question-asking information is searched in the related question-asking data corresponding to the question-asking information to serve as the target question-asking data, and the new question-asking result data is generated based on the question-asking information and the target question-asking data, it is ensured that a user can try to obtain an actually required result only by asking without modifying questions or keywords by himself / herself, and the user experience is improved. The operation complexity of the user is reduced, and the actual use experience of the user is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a result generation method, device, equipment and storage medium. Background Art

[0002] Nowadays, in the field of human-computer interaction, computers can often only feedback the same results (such as text, pictures or videos, etc.) for the same search term. If the user is not satisfied with the feedback result, he can only modify the question or search term he asked and try to obtain the result again. When adjusting the search term, the user needs to have relatively rich search or query experience to adjust it accurately, so as to ensure that the user can obtain the results he really needs. However, users generally do not have rich experience, which leads to high overall complexity of use and poor user experience.

[0003] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0004] The main purpose of the present invention is to provide a result generation method, device, equipment and storage medium, aiming to solve the technical problem in the prior art that when users are not satisfied with the obtained results, they can only adjust the questions or search terms by themselves.

[0005] To achieve the above object, the present invention provides a result generation method, which comprises the following steps:

[0006] Get follow-up information;

[0007] Selecting target question and answer data from the relevant question and answer data corresponding to the follow-up question information;

[0008] Generate question result data based on the question information and the target question and answer data.

[0009] Optionally, the step of selecting target question and answer data from the relevant question and answer data corresponding to the follow-up question information includes:

[0010] Searching for relevant question and answer data corresponding to the follow-up question information;

[0011] Detecting whether a type keyword exists in the inquiry information;

[0012] If it exists, extracting the type keyword from the query information;

[0013] Target question and answer data are selected from the related question and answer data based on the type keyword.

[0014] Optionally, after the step of detecting whether the query information contains a type keyword, the step further includes:

[0015] If not, obtaining the information modification intention corresponding to the inquiry information;

[0016] Screening the relevant question and answer data according to the information modification intention to obtain question and answer data to be selected;

[0017] Matching the information modification intention with the question and answer data to be selected, and determining the data matching degree corresponding to each question and answer data to be selected;

[0018] Target question and answer data are selected from the candidate question and answer data based on the data matching degree.

[0019] Optionally, the step of obtaining the information modification intention corresponding to the inquiry information includes:

[0020] Extracting keywords from the follow-up question information to obtain a set of follow-up question keywords;

[0021] Performing weight analysis on the query keyword set to obtain a keyword weight sequence;

[0022] Matching is performed in a preset knowledge base according to the keyword weight sequence to determine the information modification intention corresponding to the follow-up information.

[0023] Optionally, the step of generating follow-up question result data based on the follow-up question information and the target question and answer data includes:

[0024] Obtaining the information modification intention corresponding to the inquiry information;

[0025] Obtaining question information and result type corresponding to the target question and answer data, and extracting question information from the target question and answer data;

[0026] Adjust the question information according to the result type and the information modification intention to obtain follow-up question information;

[0027] Generate follow-up question result data based on the target question and answer data and the follow-up question information.

[0028] Optionally, the inquiry result data includes text type data;

[0029] The step of generating follow-up question result data based on the target question and answer data and the follow-up question information includes:

[0030] Searching in a preset corpus according to the follow-up question information to obtain at least one answer corpus information;

[0031] Perform keyword extraction on the answer corpus information to obtain result keywords;

[0032] Construct the follow-up question result data according to the result keywords and the target question and answer data.

[0033] Optionally, the step of extracting keywords from the answer corpus information to obtain result keywords includes:

[0034] Selecting target corpus information from the at least one answer corpus information according to the target result data;

[0035] Obtaining question keywords corresponding to the follow-up question information;

[0036] Find the answer keywords corresponding to the question keywords;

[0037] Keyword extraction is performed on the target corpus information according to the answer keywords to obtain result keywords.

[0038] Optionally, the inquiry result data includes picture type data;

[0039] The step of generating follow-up question result data based on the target question and answer data and the follow-up question information includes:

[0040] Searching in a preset material library according to the follow-up question information to obtain a drawing template and picture material;

[0041] Generate a picture background based on the drawing template;

[0042] Adjust the picture material based on the drawing template to obtain the material to be inserted;

[0043] Fill the material to be inserted into the picture background to obtain the picture to be displayed;

[0044] Construct question result data according to the to-be-displayed picture and the target question and answer data.

[0045] Optionally, the step of constructing the follow-up question result data according to the to-be-displayed picture and the target question and answer data includes:

[0046] Compare the image to be displayed with the image data in the target question and answer data to obtain image difference information;

[0047] Constructing adjustment description information according to the information modification intention and the picture difference information;

[0048] Construct the inquiry result data according to the picture to be displayed, the target question and answer data and the adjustment description information.

[0049] Optionally, the inquiry result data includes video type data;

[0050] The step of generating follow-up question result data based on the target question and answer data and the follow-up question information includes:

[0051] Searching for corresponding video data in a preset knowledge base according to the follow-up question information to obtain result video data;

[0052] Performing speech recognition on the result video data to obtain a video explanation text;

[0053] Compare the video explanation text with the video explanation text in the target question and answer data to obtain text difference information;

[0054] The question result data is constructed according to the text difference information, the result video data and the target question and answer data.

[0055] Optionally, the step of constructing the question result data according to the text difference information, the result video data and the target question and answer data includes:

[0056] Extract keywords from the video explanation text to obtain explanation keywords;

[0057] Construct a video content summary based on the explanation keywords;

[0058] The question result data is constructed according to the video content summary, the text difference information, the result video data and the target question and answer data.

[0059] Optionally, the step of adjusting the question information according to the result type and the information modification intention to obtain follow-up question information includes:

[0060] Determine the problem modification location and information modification instructions based on the information modification intention;

[0061] Determining the keywords to be modified in the question information according to the question modification position;

[0062] Determine a replacement keyword according to the result type, the keyword to be modified and the information modification instruction;

[0063] The keywords to be modified in the question information are adjusted according to the replacement keywords to obtain follow-up question information.

[0064] Optionally, the step of obtaining follow-up information includes:

[0065] Obtain the follow-up information input by the user through the interactive input box, or obtain the follow-up information input by the user through the follow-up prompt button.

[0066] Optionally, the interactive input box and / or the follow-up prompt button includes at least one of the following:

[0067] The interactive input box is displayed independently;

[0068] The interactive input box and the interactive topic are displayed correspondingly;

[0069] Displayed in association with question and answer pairs;

[0070] Display in association with the question;

[0071] Display in association with the answer;

[0072] After receiving the user's operation of selecting page content, it is displayed in association with the selected content.

[0073] In addition, to achieve the above object, the present invention also proposes a result generating device, which includes the following modules:

[0074] An acquisition module is used to obtain follow-up information;

[0075] A selection module, used to select target question and answer data from the relevant question and answer data corresponding to the follow-up question information;

[0076] A generation module is used to generate question result data based on the question information and the target question and answer data.

[0077] Optionally, the selection module is also used to obtain the information modification intention corresponding to the follow-up information, and search for related question and answer data corresponding to the follow-up information; detect whether there is a type keyword in the follow-up information; if so, extract the type keyword from the follow-up information; and select target question and answer data from the related question and answer data based on the type keyword and the information modification intention.

[0078] Optionally, the selection module is further used to extract keywords from the follow-up information to obtain a follow-up keyword set; perform weight analysis on the follow-up keyword set to obtain a keyword weight sequence; and match the keyword weight sequence in a preset knowledge base to determine the information modification intention corresponding to the follow-up information.

[0079] Optionally, the selection module is also used to determine the modification target type and the information modification indication according to the information modification intention if it does not exist; filter the relevant question and answer data according to the modification target type to obtain the question and answer data to be selected; match the information modification indication with the question and answer data to be selected to determine the data matching degree corresponding to each question and answer data to be selected; and select the target question and answer data from the question and answer data to be selected based on the data matching degree.

[0080] In addition, to achieve the above-mentioned purpose, the present invention also proposes a result generating device, which includes: a processor, a memory, and a result generating program stored in the memory and executable on the processor, and the result generating program implements the steps of the result generating method described above when executed.

[0081] In addition, to achieve the above-mentioned purpose, the present invention also proposes a computer-readable storage medium, on which a result generation program is stored, and when the result generation program is executed, the steps of the result generation method described above are implemented.

[0082] The present invention obtains follow-up information; selects target question and answer data from the relevant question and answer data corresponding to the follow-up information; and generates follow-up result data based on the follow-up information and the target question and answer data. Since the most relevant question and answer data to the follow-up information is searched from the relevant question and answer data corresponding to the follow-up information as the target question and answer data, and new follow-up result data is generated based on the follow-up information and the target question and answer data, it is ensured that the user does not need to modify the question or keywords by himself, and only needs to ask questions to try to obtain the actual required results, which reduces the user's operation complexity and improves the user's actual use experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 It is a schematic diagram of the structure of an electronic device in a hardware operating environment involved in an embodiment of the present invention;

[0084] Figure 2 It is a flowchart of the first embodiment of the result generation method of the present invention;

[0085] Figure 3 It is a flow chart of the second embodiment of the result generating method of the present invention;

[0086] Figure 4 It is a flowchart of a third embodiment of the result generation method of the present invention;

[0087] Figure 5 It is a flowchart of a fourth embodiment of the result generating method of the present invention;

[0088] Figure 6 It is a flowchart of a fifth embodiment of the result generating method of the present invention;

[0089] Figure 7 It is a flowchart of a sixth embodiment of the result generation method of the present invention;

[0090] Figure 8 It is a structural block diagram of the first embodiment of the result generating device of the present invention.

[0091] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0092] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0093] Reference Figure 1 , Figure 1 A schematic diagram of a device structure is generated as a result of the hardware operating environment involved in the embodiment of the present invention.

[0094] like Figure 1 As shown, the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0095] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the electronic device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0096] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a result generation program.

[0097] exist Figure 1 In the electronic device shown, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the electronic device of the present invention can be set in a result generating device, and the electronic device calls the result generating program stored in the memory 1005 through the processor 1001, and executes the result generating method provided by an embodiment of the present invention.

[0098] The embodiment of the present invention provides a method for generating a result, referring to Figure 2 , Figure 2 The figure is a flow chart of a first embodiment of a result generating method of the present invention.

[0099] In this embodiment, the result generation method includes the following steps:

[0100] Step S10: Obtain follow-up information.

[0101] It should be noted that the executor of this embodiment may be the result generating device, in which a result generating program is provided (it may also be referred to as an intelligent assistant or an intelligent big model, and the result generating program may be a pre-trained human-computer interaction model constructed based on an intelligent language model and / or an intelligent drawing model, wherein, when conducting model training, sample data may be collected through big data or other means, and the model may be trained through the collected sample data), and the result generating device may be an electronic device such as a personal computer or a server, or other devices that may realize the same or similar functions, which is not limited in this embodiment. In this embodiment and the following embodiments, the result generating method of this embodiment and the following embodiments is described by taking the result generating device as an example.

[0102] It should be noted that the intelligent assistant will divide the process of interacting with the user into sessions. The cycle of a session is the entire process from the user opening the interaction window to closing the interaction window. After the user obtains the result data once, if the user enters the question again, the intelligent assistant will determine that it has received follow-up information.

[0103] In order to avoid misjudgment of follow-up questions, after receiving the question or search term input by the user, it can also be matched with other questions in the current session to determine the relevance. If the relevance reaches a preset threshold, it is determined that the follow-up question information has been received.

[0104] In a specific implementation, the user can input the follow-up information through the interactive input box set in the interactive window, or can input the follow-up information by clicking the follow-up prompt button displayed in the interactive window. The display of the interactive input box and / or the follow-up prompt button may include at least one of the following:

[0105] The interactive input box is displayed independently (i.e., an independent area is divided in the interactive window to display the interactive input box);

[0106] The interactive input box is displayed corresponding to the interactive topic (i.e., an interactive input box is set for each interactive topic in the interactive window);

[0107] Displaying in association with question-answer pairs (i.e., displaying an interactive input box and / or at least one follow-up prompt button in the display area of ​​each question-answer pair);

[0108] Display in association with the question (i.e., displaying an interactive input box and / or at least one follow-up question prompt button in the display area for each question input by the user);

[0109] Display in association with the answer (i.e., display an interactive input box and / or at least one follow-up prompt button in the display area of ​​each answer);

[0110] After receiving the user's operation of selecting page content, it is displayed in association with the selected content (that is, after the user selects page content in the interactive interface, an interactive input box and / or at least one follow-up prompt button is displayed near the selected content).

[0111] The question may be a question input by the user, and the answer may be the displayed result data.

[0112] In a specific implementation, the result generation method provided by the present invention can be applied in the interactive scenario of intelligent question and answer of the model. For example: after the user inputs the question for the first time, the intelligent assistant (also referred to as the intelligent big model) will display at least one result data in the interactive window according to the question input by the user (at this time, the interactive input box and / or the follow-up prompt button can be displayed in the interactive window in at least one of the above methods). If the user is not satisfied with the displayed result data (such as incomplete results or hope to go deeper), and needs to ask questions, then at this time, the follow-up information can be input by clicking the follow-up prompt button and / or entering data in the interactive input box. When the intelligent assistant receives the follow-up information, it will execute the result generation method provided by any embodiment of the present invention, further generate result data, and display it in the interactive window. Similarly, the interactive window further displayed at this time still has the interactive input box and / or the follow-up prompt button. If the user is still not satisfied with the result data, the follow-up information can be input again for questioning until the result data required by the user is obtained.

[0113] Step S20: Select target question and answer data from the relevant question and answer data corresponding to the follow-up question information.

[0114] It should be noted that when generating result data, the intelligent assistant will determine the question information based on the question or keyword input by the user, combined with the current context and the user's historical interactive feedback data, and generate result data based on the question information. At this time, the relevant question and answer data corresponding to the follow-up information may include the question input by the user in the same session as the follow-up information, the question information used by the intelligent assistant when generating the result data, and the generated result data. Selecting the target question and answer data from the relevant question and answer data may be selecting the question and answer data with the highest degree of relevance to the follow-up information from the relevant question and answer data as the target question and answer data.

[0115] In the specific implementation, in order to ensure that historical result data can be accurately obtained, the intelligent assistant can set a corresponding unique identifier for each conversation. When it is necessary to obtain relevant question and answer data corresponding to the follow-up information, the session unique identifier of the conversation to which the follow-up information belongs can be obtained, and the relevant question and answer data can be searched according to the session unique identifier.

[0116] Step S30: Generate inquiry result data based on the inquiry information and the target question and answer data.

[0117] It should be noted that generating follow-up question result data based on follow-up question information and target question and answer data can be to modify the question information in the target question and answer data according to the follow-up question information to obtain follow-up question information, and then search for relevant corpus data based on the follow-up question information to generate follow-up question result data according to the corpus data.

[0118] This embodiment obtains follow-up information; selects target question and answer data from the relevant question and answer data corresponding to the follow-up information; and generates follow-up question result data based on the follow-up information and the target question and answer data. Since the target question and answer data with the highest degree of relevance to the follow-up information will be searched from the relevant question and answer data corresponding to the follow-up information, and new follow-up question result data will be generated based on the follow-up information and the target question and answer data, it is ensured that the user does not need to modify the question or keywords by himself, and only needs to ask questions to try to obtain the actual required results, which reduces the user's operation complexity and improves the user's actual use experience.

[0119] refer to Figure 3 , Figure 3 It is a flowchart of a second embodiment of a result generating method of the present invention.

[0120] Based on the first embodiment, step S20 of the result generation method of this embodiment includes:

[0121] Step S201: searching for relevant question and answer data corresponding to the follow-up question information.

[0122] It should be noted that searching for relevant question and answer data corresponding to follow-up question information can be to search for questions input by the user in the same session as the follow-up question information, the question information used when the intelligent assistant generates result data corresponding to the question input by the user, and the generated result data, thereby obtaining the relevant question and answer data corresponding to the follow-up question information.

[0123] Step S202: Detect whether there is a type keyword in the inquiry information.

[0124] It should be noted that the type keyword may be a keyword used to indicate the result type, such as keywords such as picture, video, etc., or other keywords with the same meaning.

[0125] Step S203: If it exists, extract the type keyword from the query information.

[0126] It is understandable that if there is a type keyword, then an attempt can be made to determine the question-answer data with the highest degree of relevance to the follow-up information through the type keyword, and therefore, the type keyword can be extracted from the follow-up information.

[0127] Step S204: Select target question and answer data from the related question and answer data based on the type keyword.

[0128] It should be noted that selecting target question and answer data from related question and answer data based on type keywords can be to obtain the data type of the result data in each related question and answer data, match the type keyword with the data type, and select the related question and answer data whose data type of the corresponding result data matches the type keyword as the target question and answer data.

[0129] Furthermore, if the type keyword does not exist in the follow-up question information, then at this time, the data matching degree between each related question and answer data and the follow-up question information can be calculated, and the target question and answer data can be selected according to the data matching degree. At this time, after step S202 described in this embodiment, the following may also be included:

[0130] If not, obtaining the information modification intention corresponding to the inquiry information;

[0131] Screening the relevant question and answer data according to the information modification intention to obtain question and answer data to be selected;

[0132] Matching the information modification intention with the question and answer data to be selected, and determining the data matching degree corresponding to each question and answer data to be selected;

[0133] Target question and answer data are selected from the candidate question and answer data based on the data matching degree.

[0134] It should be noted that when users ask follow-up questions about other previous questions, they will generally be more specific. At this time, the information modification intention determined based on the follow-up information will include information related to the previously asked questions, such as the field to which the question belongs (such as the mechanical field, software development field, etc.) and more specific question modification instructions. At this time, the relevant question and answer data are screened according to the information modification intention. The question and answer data to be selected can be obtained by determining the field to which the question belongs based on the information modification intention, and then screening the relevant question and answer data based on the field to which the question belongs, filtering out relevant question and answer data with different fields, and using the remaining part as the question and answer data to be selected, and then determining the question modification instructions based on the information modification intention, matching the question modification instructions with the question and answer data to be selected, and determining the data matching degree corresponding to each question and answer data to be selected.

[0135] In actual use, selecting the target question and answer data from the candidate question and answer data based on the data matching degree may be selecting the candidate question and answer data with the largest corresponding data matching degree as the target question and answer data.

[0136] It should be noted that if there is a type keyword in the follow-up information, but there are multiple data in the relevant question and answer data corresponding to the follow-up information that match the data type of the corresponding result data and the type keyword, then at this time, the relevant question and answer data corresponding to the follow-up information can be firstly screened based on the type keyword, and then the data matching degree of the remaining relevant question and answer data can be calculated according to the information modification intention corresponding to the follow-up information, and the target question and answer data can be selected according to the data matching degree.

[0137] In a specific implementation, in order to ensure that the information modification intention corresponding to the follow-up information can be accurately obtained, the step of obtaining the information modification intention corresponding to the follow-up information in this embodiment may include:

[0138] Extracting keywords from the follow-up question information to obtain a set of follow-up question keywords;

[0139] Performing weight analysis on the query keyword set to obtain a keyword weight sequence;

[0140] Matching is performed in a preset knowledge base according to the keyword weight sequence to determine the information modification intention corresponding to the follow-up information.

[0141] It should be noted that extracting keywords from the follow-up information to obtain the follow-up keyword set may be to segment the follow-up information, and after segmentation, obtain a segmentation set, and delete modal particles, conjunctions, etc. in the segmentation set to obtain the follow-up keyword set.

[0142] In actual use, weight analysis of the follow-up keyword set can be performed in combination with context analysis to determine the weight of each keyword in the follow-up keyword set, and then sort the keywords in the follow-up keyword set from high to low according to the weight, and construct a keyword weight sequence based on the sorting results.

[0143] It should be noted that when users input follow-up information, the question structure (the grammatical structure used to ask questions, including the connection relationship, connection order, conjunctions, etc. between key words) and the nature of the information (the type of question asked, such as rhetorical questions, interrogative sentences, etc.) used may affect the weight of the keywords. In order to ensure the normal construction of the keyword weight sequence, the question structure and information nature corresponding to the follow-up information can be obtained first, and then the follow-up information can be semantically analyzed according to the question structure and information nature to determine the weight ratio corresponding to each follow-up keyword in the follow-up keyword set, and then the follow-up keywords in the follow-up keyword set can be sorted based on the weight ratio to obtain the keyword weight sequence.

[0144] Among them, since the keywords at the front of the sequence play a greater role in determining the intent later, when sorting the follow-up keywords in the follow-up keyword set based on the weight ratio, the follow-up keywords in the follow-up keyword set can be sorted from large to small according to the corresponding weight ratio to ensure that the keywords with higher weight ratios are at the front of the sequence.

[0145] In a specific implementation, the preset knowledge base may be a database used to pre-train the intelligent assistant, which contains a large number of data samples constructed according to the keyword weight sequence and the modification intention. After the intelligent assistant determines the keyword weight sequence corresponding to the follow-up information, it can match the keyword weight sequence in the preset knowledge base to determine at least one data sample, and make a prediction based on the data sample to determine the information modification intention corresponding to the follow-up information.

[0146] This embodiment ensures that unnecessary calculations can be reduced as much as possible by selecting target question and answer data from related question and answer data based on the type keyword when detecting whether there is a type keyword in the follow-up question information; and when the type keyword does not exist, the data matching degree of each related question and answer data is calculated, and the target question and answer data is selected based on the data matching degree, to ensure that the target question and answer data with the highest degree of relevance to the follow-up question information can be accurately obtained.

[0147] refer to Figure 4 , Figure 4 It is a flowchart diagram of a third embodiment of a result generating method of the present invention.

[0148] Based on the first embodiment, step S30 of the result generation method of this embodiment includes:

[0149] Step S301: Obtain the information modification intention corresponding to the inquiry information.

[0150] It should be noted that the method of obtaining the information modification intention corresponding to the follow-up information is consistent with the obtaining method described in the above second embodiment, and will not be repeated here.

[0151] Step S302: Obtain question information and result type corresponding to the target question and answer data, and extract question information from the target question and answer data.

[0152] It should be noted that the result type may be the data type of the generated result data, and the result type may include multiple types such as text type, image type, and video type.

[0153] Among them, if the result data only contains text data, the result type of the result data can be determined to be a text type; if the result data also contains image data on the basis of the text data, the result type of the result data can be determined to be a picture type; if the result data also contains video data on the basis of the text data and / or image data, the result type of the result data can be determined to be a video type.

[0154] In actual use, obtaining the question information and result type corresponding to the target question and answer data may be to extract the result data and question information from the target question and answer information, and obtain the data type of the result data, thereby obtaining the result type.

[0155] Step S303: adjusting the question information according to the result type and the information modification intention to obtain follow-up question information.

[0156] It should be noted that the question information is adjusted according to the result type and the information modification intention, and the follow-up question information is obtained by adjusting the keywords in the question information according to the result type and the information modification intention, so as to obtain the follow-up question information. Among them, adjusting the keywords in the question information may be replacing keywords in the question information or adding keywords, etc., and this embodiment does not limit this.

[0157] Further, in order to ensure that the generated follow-up question information meets the user's true intention, step S303 in this embodiment may include:

[0158] Determine the problem modification location and information modification instructions based on the information modification intention;

[0159] Determining the keywords to be modified in the question information according to the question modification position;

[0160] Determine a replacement keyword according to the result type, the keyword to be modified and the information modification instruction;

[0161] The keywords to be modified in the question information are adjusted according to the replacement keywords to obtain follow-up question information.

[0162] It should be noted that determining the question modification position and question modification instructions based on the information modification intention may be to determine the keyword position related to the information modification intention in the question information, determine the question modification position, and search for question modification instructions that match the information modification intention. The question modification instructions may include the adjustment direction of the keywords, such as replacing them with more detailed synonyms or replacing them with professional terms similar to them.

[0163] In a specific implementation, determining the keywords to be modified in the question information according to the question modification position may be to obtain the keywords located at the question modification position in the question information, thereby determining the keywords to be modified. Determining the replacement keywords according to the result type, the keywords to be modified, and the information modification indication may be to search for corresponding keywords according to the adjustment direction of the keywords in the question modification indication, the keywords to be modified, and the result type, and to construct a keyword set according to the found keywords, and to extract a keyword from the keyword set as the replacement keyword.

[0164] In actual use, the keywords to be modified in the question information are adjusted according to the replacement keywords, and the follow-up question information can be obtained by modifying the keywords to be modified in the question information into the replacement keywords, thereby obtaining the follow-up question information.

[0165] Step S304: Generate follow-up question result data based on the target question and answer data and the follow-up question information.

[0166] It should be noted that generating follow-up question result data based on the target question and answer data and follow-up question information can be to search for the corresponding answer corpus based on the follow-up question information, and generate follow-up question result data whose result type is consistent with the result type of the result data in the target question and answer data according to the answer corpus.

[0167] This embodiment obtains the information modification intention corresponding to the follow-up information; obtains the question information and result type corresponding to the target question and answer data, and extracts the question information from the target question and answer data; adjusts the question information according to the result type and the information modification intention to obtain the follow-up question information; and generates the follow-up question result data based on the target question and answer data and the follow-up question information. Since the follow-up information is parsed, the corresponding information modification intention is determined, and then the question information is modified according to the information modification intention and the result type, it is ensured that the final generated follow-up question information meets the actual result acquisition needs of the user, thereby improving the accuracy of the result generation method of this application.

[0168] refer to Figure 5 , Figure 5 It is a flowchart diagram of a fourth embodiment of a result generating method of the present invention.

[0169] Based on the third embodiment, the question result data generated in the result generation method of this embodiment is text type data. In this case, step S304 of this embodiment includes:

[0170] Step S3041: searching in a preset corpus according to the follow-up question information to obtain at least one piece of answer corpus information.

[0171] It should be noted that searching in a preset knowledge base based on the follow-up question information to obtain at least one piece of answer corpus information may be searching in a preset corpus for answer corpus information that matches the follow-up question information, thereby obtaining at least one piece of answer corpus information. The manager of the result generating device may pre-set corresponding data tags for the answer corpus information in the preset corpus, and if the tag of the answer corpus information matches the keyword in the follow-up question information, it may be determined that the answer corpus information matches the follow-up question information.

[0172] Step S3042: extract keywords from the answer corpus information to obtain result keywords.

[0173] It should be noted that the answer corpus information is generally the answers or suggestions obtained on the Internet, and the content of some answer corpus information is bloated and contains a lot of useless information. In order to exclude this useless information, keyword extraction can be performed on the answer corpus information to obtain keywords related to the answer question information, that is, result keywords.

[0174] Further, in order to accurately extract keywords from the answer corpus information, step S3042 in this embodiment may include:

[0175] Obtaining question keywords corresponding to the follow-up question information;

[0176] Find the answer keywords corresponding to the question keywords;

[0177] Keyword extraction is performed on the answer corpus information according to the answer keywords to obtain result keywords.

[0178] It should be noted that obtaining the question keyword corresponding to the follow-up question information may be to parse the follow-up question information to obtain at least one keyword contained in the follow-up question information, thereby obtaining the question keyword. Finding the answer keyword corresponding to the question keyword may be to find the answer keyword corresponding to the question keyword in a preset keyword mapping table, wherein the answer keyword may be a keyword that may appear when the question corresponds to the question keyword, and the preset keyword mapping table may include a mapping relationship between the answer keyword and the question keyword, and the mapping relationship may be pre-set by the administrator of the result generation device, or may be constructed by the intelligent assistant when training the intelligent assistant.

[0179] In a specific implementation, keyword extraction is performed on the answer corpus information according to the answer keyword, and the result keyword can be obtained by extracting keywords that are the same as the answer keyword or have a similar part of speech to the answer keyword from the answer corpus information to obtain the result keyword.

[0180] Step S3043: constructing follow-up question result data according to the result keywords and the target question and answer data.

[0181] In a specific implementation, constructing follow-up result data based on result keywords and historical result data can be to use the historical result data as context information, make predictions based on the context information, determine the connection order and connecting words of each result keyword, and then connect the result keywords according to the connection order and connecting words, thereby generating follow-up result data.

[0182] This embodiment searches in a preset corpus according to the follow-up question information to obtain at least one answer corpus information; extracts keywords from the answer corpus information to obtain result keywords; and constructs follow-up question result data according to the result keywords and the target question and answer data. Since keyword extraction is performed on the answer corpus information after it is obtained, and follow-up question result data is constructed according to the extracted result keywords and historical result data, it is ensured that the final result data will not be too redundant, which is convenient for users to read, thereby improving the user experience.

[0183] refer to Figure 6 , Figure 6 It is a flowchart diagram of a fifth embodiment of a result generating method of the present invention.

[0184] Based on the third embodiment, the question result data generated in the result generation method of this embodiment is picture type data. In this case, step S304 of this embodiment includes:

[0185] Step S3041': searching in a preset material library according to the follow-up question information to obtain a drawing template and picture materials.

[0186] It should be noted that searching in a preset knowledge base based on the follow-up question information to obtain at least one answer corpus information may be searching for a painting template and a picture material matching the follow-up question information in a preset material library. The manager of the result generation device may pre-set corresponding data labels for each painting template and picture material in the preset material library. If the label of the painting template or picture material matches the keyword in the follow-up question information, it can be determined that the painting template or picture material matches the follow-up question information.

[0187] Step S3042': Generate a picture background based on the drawing template.

[0188] In a specific implementation, the intelligent assistant can extract the background contour data, color data, and texture data recorded in the drawing template, and then make predictions based on the follow-up question information combined with the background contour data, color data, and texture data to determine the picture background required by the user.

[0189] Step S3043': adjusting the picture material based on the drawing template to obtain the material to be inserted.

[0190] In actual use, the picture material is adjusted based on the drawing template, and the material to be inserted can be obtained by adjusting the size and dimension of the picture material based on the material filling area reserved in the drawing template and the size of the material filling area, and performing style prediction by asking follow-up question information, and further adjusting the picture material according to the predicted style to obtain the material to be inserted.

[0191] Step S3044': fill the material to be inserted into the picture background to obtain the picture to be displayed.

[0192] It can be understood that filling the material to be inserted into the picture background to obtain the picture to be displayed may be filling the material to be inserted into a material filling area reserved in the picture background to obtain the picture to be displayed.

[0193] Step S3045': constructing question result data according to the to-be-displayed picture and the target question and answer data.

[0194] It should be noted that constructing the follow-up question result data based on the picture to be displayed and the target question and answer data can be to extract the picture data contained in the target answer data, and then construct the follow-up question result data based on the picture data and the picture to be displayed, so as to observe the difference between the picture data generated based on the follow-up question information and the previous picture data.

[0195] Furthermore, in order to allow the user to more intuitively observe the difference between the newly generated image and the previous image, step S3045' in this embodiment may include:

[0196] Compare the image to be displayed with the image data in the target question and answer data to obtain image difference information;

[0197] Constructing adjustment description information according to the information modification intention and the picture difference information;

[0198] Construct the inquiry result data according to the picture to be displayed, the target question and answer data and the adjustment description information.

[0199] In a specific implementation, constructing adjustment description information according to information modification intention and picture difference information can be to determine the user modification intention, picture difference position and picture difference type according to the information modification intention and picture difference information, and then generate a concise and fluent descriptive text in combination with the user modification intention, picture difference position and picture difference type, thereby obtaining the adjustment description information.

[0200] It should be understood that the picture to be displayed and the picture data in the target question and answer data are compared in advance, and then the adjustment description information is constructed based on the picture difference information and the information modification intention, and then the follow-up question result data is constructed based on the adjustment description information, the picture to be displayed, and the target question and answer data. The user can then quickly know the adjustment position of the picture directly based on the adjustment description information contained in the follow-up question result data, and compare it with the picture to be displayed and the picture data in the target question and answer data to determine whether the modification of the picture is in line with its true intention.

[0201] This embodiment obtains a drawing template and a picture material by searching in a preset material library according to the follow-up question information; generates a picture background based on the drawing template; adjusts the picture material based on the drawing template to obtain a material to be inserted; fills the material to be inserted into the picture background to obtain a picture to be displayed; and constructs the question result data according to the picture to be displayed and the target question and answer data. Since after searching for a drawing template and a picture material to generate a picture to be displayed according to the follow-up question information, the question result data will be constructed based on the picture to be displayed and the target question and answer data, ensuring that the user can directly compare the picture to be displayed with the picture data in the target question and answer data, thereby quickly determining whether the modification of the picture is in line with its true intention.

[0202] refer to Figure 7 , Figure 7 It is a flowchart diagram of a sixth embodiment of a result generating method of the present invention.

[0203] Based on the third embodiment, the question result data generated in the result generation method of this embodiment is video type data. In this case, step S304 of this embodiment includes:

[0204] Step S3041": searching for corresponding video data in a preset knowledge base according to the follow-up question information to obtain result video data.

[0205] In a specific implementation, the corresponding video data is searched in the preset knowledge base according to the follow-up question information, and the result video data is obtained by searching the preset knowledge base for the video data that matches the follow-up question information, thereby obtaining the result video data. There is a corresponding content identification tag for each video data in the preset knowledge base, and the content identification tag can be pre-calibrated by the administrator of the result generation device, or can be set by the intelligent assistant after parsing the video explanation content. If the content identification tag of the video data matches the keyword of the follow-up question information, it can be determined that the video data matches the follow-up question information.

[0206] Step S3042": perform speech recognition on the result video data to obtain video explanation text.

[0207] It should be noted that performing speech recognition on the result video data to obtain the video explanation text may be to extract the audio file corresponding to the result video data, perform speech recognition on the audio file, and thus obtain the video explanation text.

[0208] Step S3043": compare the video explanation text with the video explanation text in the target question and answer data to obtain text difference information.

[0209] In a specific implementation, the video explanation text is compared with the video explanation text in the target question and answer data, and obtaining the text difference information can be performed by extracting the text summary data of the video explanation text and extracting the text summary data of the video explanation text in the target question and answer data, and comparing the text summary data of the two to obtain the text difference information.

[0210] Step S3044": construct question result data according to the text difference information, the result video data and the target question and answer data.

[0211] It should be noted that the follow-up question result data is constructed based on the text difference information, result video data and target question and answer data. When the user views the follow-up question result data, he can determine the difference between the video data in the regenerated follow-up question result data and the previous video data based on the text difference information therein, so that the user can quickly determine whether the newly generated result data meets his actual needs.

[0212] Further, in order to enable the user to quickly identify the core content of the newly generated video data, step S3044" in this embodiment may include:

[0213] Extract keywords from the video explanation text to obtain explanation keywords;

[0214] Construct a video content summary based on the explanation keywords;

[0215] The question result data is constructed according to the video content summary, the text difference information, the result video data and the target question and answer data.

[0216] It should be noted that keyword extraction is performed on the video explanation text to obtain explanation keywords, which can be obtained by obtaining question keywords corresponding to follow-up question information, finding answer keywords corresponding to the question keywords, and extracting keywords from the video explanation text based on the answer keywords to obtain explanation keywords.

[0217] In a specific implementation, constructing a video content summary based on explanation keywords can be to perform context prediction on the explanation keywords, determine the connection order and connecting words of each explanation keyword, and then connect the explanation keywords according to the connection order and connecting words to generate a video content summary.

[0218] It can be understood that by constructing the follow-up question result data based on the video content summary, text difference information, result video data and target question and answer data, the user can not only determine the difference between the video data in the regenerated follow-up question result data and the previous video data based on the text difference information, but also know the core content of the feedback video data based on the video content summary, thereby quickly determining whether the feedback result data meets their actual needs.

[0219] This embodiment searches for the corresponding video data in the preset knowledge base according to the follow-up question information to obtain the result video data; performs voice recognition on the result video data to obtain the video explanation text; compares the video explanation text with the video explanation text in the target question and answer data to obtain text difference information; and constructs the follow-up question result data according to the text difference information, the result video data, and the target question and answer data. After the corresponding video data is found, voice recognition will be performed on the video data to obtain the corresponding video explanation text, which will be compared with the video explanation text in the target question and answer data, and the follow-up question result data will be constructed according to the text difference information obtained by the comparison, the result video data, and the Mubao question and answer data, so as to ensure that the user can quickly determine whether the generated video data meets his true intention according to the text difference information in the follow-up question result data, and further simplify the user's operation.

[0220] In addition, an embodiment of the present invention further provides a storage medium, on which a result generation program is stored. When the result generation program is executed by a processor, the steps of the result generation method described above are implemented.

[0221] Reference Figure 8 , Figure 8 It is a structural block diagram of the first embodiment of the result generating device of the present invention.

[0222] like Figure 8 As shown, the result generating device proposed in the embodiment of the present invention includes:

[0223] An acquisition module 10, used to acquire follow-up information;

[0224] A selection module 20, configured to select target question and answer data from the relevant question and answer data corresponding to the follow-up question information;

[0225] The generating module 30 is used to generate the inquiry result data based on the inquiry information and the target question and answer data.

[0226] This embodiment obtains follow-up information; selects target question and answer data from the relevant question and answer data corresponding to the follow-up information; and generates follow-up question result data based on the follow-up information and the target question and answer data. Since the target question and answer data with the highest degree of relevance to the follow-up information will be searched from the relevant question and answer data corresponding to the follow-up information, and new follow-up question result data will be generated based on the follow-up information and the target question and answer data, it is ensured that the user does not need to modify the question or keywords by himself, and only needs to ask questions to try to obtain the actual required results, which reduces the user's operation complexity and improves the user's actual use experience.

[0227] Furthermore, the selection module 20 is also used to search for relevant question and answer data corresponding to the follow-up information; detect whether there is a type keyword in the follow-up information; if so, extract the type keyword from the follow-up information; and select target question and answer data from the relevant question and answer data based on the type keyword.

[0228] Furthermore, the selection module 20 is also used to obtain the information modification intention corresponding to the follow-up information if it does not exist; filter the relevant question and answer data according to the information modification intention to obtain the question and answer data to be selected; match the information modification intention with the question and answer data to be selected to determine the data matching degree corresponding to each question and answer data to be selected; and select the target question and answer data from the question and answer data to be selected based on the data matching degree.

[0229] Furthermore, the selection module 20 is also used to extract keywords from the follow-up information to obtain a follow-up keyword set; perform weight analysis on the follow-up keyword set to obtain a keyword weight sequence; and match the keyword weight sequence in a preset knowledge base to determine the information modification intention corresponding to the follow-up information.

[0230] Furthermore, the generation module 30 is also used to obtain the information modification intention corresponding to the follow-up information; obtain the question information and result type corresponding to the target question and answer data, and extract the question information from the target question and answer data; adjust the question information according to the result type and the information modification intention to obtain follow-up question information; and generate follow-up question result data based on the target question and answer data and the follow-up question information.

[0231] Furthermore, the inquiry result data includes text type data;

[0232] The generation module 30 is further used to search in a preset corpus according to the follow-up question information to obtain at least one answer corpus information; extract keywords from the answer corpus information to obtain result keywords; and construct follow-up question result data according to the result keywords and the target question and answer data.

[0233] Furthermore, the generation module 30 is also used to select target corpus information from the at least one answer corpus information according to the target result data; obtain question keywords corresponding to the follow-up question information; search for answer keywords corresponding to the question keywords; and perform keyword extraction on the target corpus information according to the answer keywords to obtain result keywords.

[0234] Further, the inquiry result data includes picture type data;

[0235] The generation module 30 is also used to search in a preset material library according to the follow-up question information to obtain a drawing template and picture material; generate a picture background based on the drawing template; adjust the picture material based on the drawing template to obtain a material to be inserted; fill the material to be inserted into the picture background to obtain a picture to be displayed; and construct follow-up question result data based on the picture to be displayed and the target question and answer data.

[0236] Furthermore, the generation module 30 is also used to compare the picture data to be displayed and the picture data in the target question and answer data to obtain picture difference information; construct adjustment description information according to the information modification intention and the picture difference information; and construct follow-up question result data according to the picture to be displayed, the target question and answer data and the adjustment description information.

[0237] Further, the inquiry result data includes video type data;

[0238] The generation module 30 is further used to search for corresponding video data in a preset knowledge base according to the follow-up question information to obtain result video data; perform voice recognition on the result video data to obtain video explanation text; compare the video explanation text with the video explanation text in the target question and answer data to obtain text difference information; and construct follow-up question result data according to the text difference information, the result video data and the target question and answer data.

[0239] Furthermore, the generation module 30 is also used to extract keywords from the video explanation text to obtain explanation keywords; construct a video content summary based on the explanation keywords; and construct follow-up question result data based on the video content summary, the text difference information, the result video data and the target question and answer data.

[0240] Furthermore, the generation module 30 is also used to determine the question modification position and the information modification indication based on the information modification intention; determine the keywords to be modified in the question information according to the question modification position; determine the replacement keywords according to the result type, the keywords to be modified and the information modification indication; adjust the keywords to be modified in the question information according to the replacement keywords to obtain follow-up question information.

[0241] Furthermore, the acquisition module 10 is also used to acquire the follow-up question information input by the user through the interactive input box, or to acquire the follow-up question information input by the user through the follow-up question prompt button.

[0242] Furthermore, the interactive input box and / or the follow-up prompt button include at least one of the following situations:

[0243] The interactive input box is displayed independently;

[0244] The interactive input box and the interactive topic are displayed correspondingly;

[0245] Displayed in association with question and answer pairs;

[0246] Display in association with the question;

[0247] Display in association with the answer;

[0248] After receiving the user's operation of selecting page content, it is displayed in association with the selected content.

[0249] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0250] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0251] In addition, for technical details that are not described in detail in this embodiment, reference can be made to the result generation method provided in any embodiment of the present invention, and will not be repeated here.

[0252] In addition, it should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0253] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0254] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0255] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

[0256] The present application discloses A1, a result generation method, the result generation method comprising the following steps:

[0257] Get follow-up information;

[0258] Selecting target question and answer data from the relevant question and answer data corresponding to the follow-up question information;

[0259] Generate question result data based on the question information and the target question and answer data.

[0260] A2. The result generation method as described in A1, wherein the step of selecting target question and answer data from the relevant question and answer data corresponding to the follow-up question information comprises:

[0261] Searching for relevant question and answer data corresponding to the follow-up question information;

[0262] Detecting whether a type keyword exists in the inquiry information;

[0263] If it exists, extracting the type keyword from the query information;

[0264] Target question and answer data are selected from the related question and answer data based on the type keyword.

[0265] A3. The result generation method as described in A2, after the step of detecting whether the type keyword exists in the inquiry information, further comprising:

[0266] If not, obtaining the information modification intention corresponding to the inquiry information;

[0267] Screening the relevant question and answer data according to the information modification intention to obtain question and answer data to be selected;

[0268] Matching the information modification intention with the question and answer data to be selected, and determining the data matching degree corresponding to each question and answer data to be selected;

[0269] Target question and answer data are selected from the candidate question and answer data based on the data matching degree.

[0270] A4. In the result generation method described in A3, the step of obtaining the information modification intention corresponding to the inquiry information comprises:

[0271] Extracting keywords from the follow-up question information to obtain a set of follow-up question keywords;

[0272] Performing weight analysis on the query keyword set to obtain a keyword weight sequence;

[0273] Matching is performed in a preset knowledge base according to the keyword weight sequence to determine the information modification intention corresponding to the follow-up information.

[0274] A5. The result generation method as described in A1, wherein the step of generating the follow-up question result data based on the follow-up question information and the target question and answer data comprises:

[0275] Obtaining the information modification intention corresponding to the inquiry information;

[0276] Obtaining question information and result type corresponding to the target question and answer data, and extracting question information from the target question and answer data;

[0277] Adjust the question information according to the result type and the information modification intention to obtain follow-up question information;

[0278] Generate follow-up question result data based on the target question and answer data and the follow-up question information.

[0279] A6. The result generating method as described in A5, wherein the question result data includes text type data;

[0280] The step of generating follow-up question result data based on the target question and answer data and the follow-up question information includes:

[0281] Searching in a preset corpus according to the follow-up question information to obtain at least one answer corpus information;

[0282] Perform keyword extraction on the answer corpus information to obtain result keywords;

[0283] Construct the follow-up question result data according to the result keywords and the target question and answer data.

[0284] A7. The result generation method as described in A6, wherein the step of extracting keywords from the answer corpus information to obtain result keywords comprises:

[0285] Selecting target corpus information from the at least one answer corpus information according to the target result data;

[0286] Obtaining question keywords corresponding to the follow-up question information;

[0287] Find the answer keywords corresponding to the question keywords;

[0288] Keyword extraction is performed on the target corpus information according to the answer keywords to obtain result keywords.

[0289] A8, the result generating method as described in A5, wherein the questioning result data includes picture type data;

[0290] The step of generating follow-up question result data based on the target question and answer data and the follow-up question information includes:

[0291] Searching in a preset material library according to the follow-up question information to obtain a drawing template and picture material;

[0292] Generate a picture background based on the drawing template;

[0293] Adjust the picture material based on the drawing template to obtain the material to be inserted;

[0294] Fill the material to be inserted into the picture background to obtain the picture to be displayed;

[0295] Construct question result data according to the to-be-displayed picture and the target question and answer data.

[0296] A9. The result generating method as described in A8, wherein the step of constructing the follow-up question result data according to the to-be-displayed picture and the target question and answer data comprises:

[0297] Compare the image to be displayed with the image data in the target question and answer data to obtain image difference information;

[0298] Constructing adjustment description information according to the information modification intention and the picture difference information;

[0299] Construct the inquiry result data according to the picture to be displayed, the target question and answer data and the adjustment description information.

[0300] A10, the result generating method as described in A5, wherein the questioning result data includes video type data;

[0301] The step of generating follow-up question result data based on the target question and answer data and the follow-up question information includes:

[0302] Searching for corresponding video data in a preset knowledge base according to the follow-up question information to obtain result video data;

[0303] Performing speech recognition on the result video data to obtain a video explanation text;

[0304] Compare the video explanation text with the video explanation text in the target question and answer data to obtain text difference information;

[0305] The question result data is constructed according to the text difference information, the result video data and the target question and answer data.

[0306] A11. The result generation method as described in A10, wherein the step of constructing the follow-up question result data according to the text difference information, the result video data and the target question and answer data comprises:

[0307] Extract keywords from the video explanation text to obtain explanation keywords;

[0308] Construct a video content summary based on the explanation keywords;

[0309] The question result data is constructed according to the video content summary, the text difference information, the result video data and the target question and answer data.

[0310] A12. The result generating method as described in A5, wherein the step of adjusting the question information according to the result type and the information modification intention to obtain the follow-up question information comprises:

[0311] Determine the problem modification location and information modification instructions based on the information modification intention;

[0312] Determining the keywords to be modified in the question information according to the question modification position;

[0313] Determine a replacement keyword according to the result type, the keyword to be modified and the information modification instruction;

[0314] The keywords to be modified in the question information are adjusted according to the replacement keywords to obtain follow-up question information.

[0315] A13. In the result generating method as described in any one of A1 to A12, the step of obtaining follow-up information comprises:

[0316] Obtain the follow-up information input by the user through the interactive input box, or obtain the follow-up information input by the user through the follow-up prompt button.

[0317] A14. In the result generation method as described in any one of A1 to A12, the interactive input box and / or the follow-up prompt button include at least one of the following situations:

[0318] The interactive input box is displayed independently;

[0319] The interactive input box and the interactive topic are displayed correspondingly;

[0320] Displayed in association with question and answer pairs;

[0321] Display in association with the question;

[0322] Display in association with the answer;

[0323] After receiving the user's operation of selecting page content, it is displayed in association with the selected content.

[0324] The present application also discloses B15, a result generating device, the result generating device comprising the following modules:

[0325] An acquisition module is used to obtain follow-up information;

[0326] A selection module, used to select target question and answer data from the relevant question and answer data corresponding to the follow-up question information;

[0327] A generation module is used to generate question result data based on the question information and the target question and answer data.

[0328] B16. In the result generating device as described in B15, the selection module is further used to obtain the information modification intention corresponding to the follow-up information, and search for relevant question and answer data corresponding to the follow-up information; detect whether there is a type keyword in the follow-up information; if so, extract the type keyword from the follow-up information; and select target question and answer data from the relevant question and answer data based on the type keyword and the information modification intention.

[0329] B17. The result generating device as described in B16, wherein the selection module is further used to extract keywords from the follow-up information to obtain a follow-up keyword set; perform weight analysis on the follow-up keyword set to obtain a keyword weight sequence; and perform matching in a preset knowledge base according to the keyword weight sequence to determine the information modification intention corresponding to the follow-up information.

[0330] B18. The result generating device as described in B16, wherein the selection module is further used to determine the modification target type and the information modification indication according to the information modification intention if they do not exist; to filter the relevant question and answer data according to the modification target type to obtain the question and answer data to be selected; to match the information modification indication with the question and answer data to be selected to determine the data matching degree corresponding to each question and answer data to be selected; and to select the target question and answer data from the question and answer data to be selected based on the data matching degree.

[0331] The present application also discloses C19, a result generating device, which includes: a processor, a memory, and a result generating program stored in the memory and executable on the processor, wherein the result generating program implements the steps of the result generating method described above when executed.

[0332] The present application also discloses D20, a computer-readable storage medium, on which a result generation program is stored, and when the result generation program is executed, the steps of the result generation method described above are implemented.

Claims

1. A result generation method, characterized in that: The result generation method comprises the following steps: Get follow-up information; Selecting target question and answer data from the relevant question and answer data corresponding to the follow-up question information; Generate question result data based on the question information and the target question and answer data.

2. The result generation method according to claim 1, characterized in that: The step of selecting target question and answer data from the relevant question and answer data corresponding to the follow-up question information includes: Searching for relevant question and answer data corresponding to the follow-up question information; Detecting whether a type keyword exists in the inquiry information; If it exists, extracting the type keyword from the query information; Target question and answer data are selected from the related question and answer data based on the type keyword.

3. The result generation method according to claim 2, characterized in that: After the step of detecting whether the type keyword exists in the inquiry information, the method further includes: If not, obtaining the information modification intention corresponding to the inquiry information; Screening the relevant question and answer data according to the information modification intention to obtain question and answer data to be selected; Matching the information modification intention with the question and answer data to be selected, and determining the data matching degree corresponding to each question and answer data to be selected; Target question and answer data are selected from the candidate question and answer data based on the data matching degree.

4. The result generating method according to claim 3, characterized in that: The step of obtaining the information modification intention corresponding to the inquiry information includes: Extracting keywords from the follow-up question information to obtain a set of follow-up question keywords; Performing weight analysis on the query keyword set to obtain a keyword weight sequence; Matching is performed in a preset knowledge base according to the keyword weight sequence to determine the information modification intention corresponding to the follow-up information.

5. The result generation method according to claim 1, characterized in that: The step of generating the follow-up question result data based on the follow-up question information and the target question and answer data includes: Obtaining the information modification intention corresponding to the inquiry information; Obtaining question information and result type corresponding to the target question and answer data, and extracting question information from the target question and answer data; Adjust the question information according to the result type and the information modification intention to obtain follow-up question information; Generate follow-up question result data based on the target question and answer data and the follow-up question information.

6. The result generating method according to claim 5, characterized in that: The inquiry result data includes text type data; The step of generating follow-up question result data based on the target question and answer data and the follow-up question information includes: Searching in a preset corpus according to the follow-up question information to obtain at least one answer corpus information; Perform keyword extraction on the answer corpus information to obtain result keywords; Construct the follow-up question result data according to the result keywords and the target question and answer data.

7. The result generating method according to claim 6, characterized in that: The step of extracting keywords from the answer corpus information to obtain result keywords includes: Selecting target corpus information from the at least one answer corpus information according to the target result data; Obtaining question keywords corresponding to the follow-up question information; Find the answer keywords corresponding to the question keywords; Keyword extraction is performed on the target corpus information according to the answer keywords to obtain result keywords.

8. A result generating device, characterized in that: The result generating device comprises the following modules: An acquisition module is used to obtain follow-up information; A selection module, used to select target question and answer data from the relevant question and answer data corresponding to the follow-up question information; A generation module is used to generate question result data based on the question information and the target question and answer data.

9. A result generating device, characterized in that: The result generation device comprises: a processor, a memory, and a result generation program stored in the memory and executable on the processor, wherein the result generation program implements the steps of the result generation method according to any one of claims 1 to 7 when executed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a result generation program, which, when executed, implements the steps of the result generation method according to any one of claims 1 to 7.

Images