Data processing method, query method, device, equipment, medium and program product

Abstract

The present disclosure provides a data processing method, a query method, an apparatus, a device, a medium and a program product, relating to the technical field of artificial intelligence and computer technology. The method comprises: collecting user question and answer content from different intelligent question and answer applications based on a preset collection interface, the user question and answer content comprising question content and question answers of user questions; in response to a query request for the question content of the user questions, querying whether there is first target data matching the question content from a target data set stored in a distributed database according to the question content of the user questions; in response to the existence of the first target data matching the question content in the target data set, generating second target data and a first target file according to the first target data, the question content of the user questions and the question answers; and updating the target data set based on the second target data and the first target file to obtain a new target data set.

Description

Technical Field

[0001] This disclosure relates to the fields of artificial intelligence and computer technology, and more specifically, to a data processing method, query method, apparatus, device, medium, and program product. Background Technology

[0002] The intelligent question-and-answer application system integrates technologies such as knowledge representation, information retrieval, and natural language processing. It can receive questions from users in natural language and return concise and accurate answers.

[0003] In the process of realizing the present invention, the inventors discovered that the related technology has at least the following problems: different business applications in an enterprise have corresponding intelligent question and answer systems, which are stored in their respective applications in units of user-initiated sessions. When a user needs to query a question that was asked before, the user needs to switch between different intelligent question and answer applications multiple times and search multiple times to find the answer to the previous question, resulting in low user operation efficiency. Summary of the Invention

[0004] In view of this, the present disclosure provides a data processing method, query method, apparatus, device, medium and program product.

[0005] One aspect of this disclosure provides a data processing method, comprising: collecting user question-and-answer content from different intelligent question-and-answer applications based on a preset collection interface, wherein the user question-and-answer content includes the content of a question asked by a user and the answer to the question; responding to a query request for the content of a question asked by a user, querying whether first target data matching the content of the question asked exists in a target data set stored in a distributed database, based on the content of the question asked by the user; responding to the existence of first target data matching the content of the question in the target data set, generating second target data and a first target file based on the first target data, the content of the question asked by the user, and the answer to the question, wherein the text content of the first target file is used to represent the answer to the question related to the second target data; and updating the target data set based on the second target data and the first target file to obtain a new target data set.

[0006] According to embodiments of this disclosure, in response to the existence of first target data matching the question content in the target data set, generating second target data and a first target file based on the first target data, the question content asked by the user, and the question answer includes: in response to the existence of first target data matching the question content, generating second target data based on the first target data and the question content asked by the user; obtaining a second target file from a cloud disk system based on the file identifier of the second target file associated with the first target data; and generating a first target file based on the question content asked by the user, the question answer, and the second target file.

[0007] According to embodiments of this disclosure, generating second target data based on first target data and user-asked questions includes: constructing first preset prompt words based on the first target data and user-asked questions; performing multimodal feature fusion on the first target data, question content, and first preset prompt words to obtain multimodal fusion features; performing hierarchical feature extraction on the multimodal features to obtain hierarchical features; and generating second target data based on the hierarchical features.

[0008] According to embodiments of this disclosure, generating a first target file based on a user-asked question, the answer to the question, and a second target file includes: extracting text features from the second target file to obtain the text content of the second target file; constructing a second preset prompt word based on the user-asked question, the answer to the question, and the text content of the second target file; aggregating the user-asked question, the answer to the question, the text content of the second target file, and the second preset prompt word to obtain the first target text content; and performing format conversion processing on the first target text content to generate the first target file.

[0009] According to embodiments of this disclosure, the method further includes: in response to the absence of first target data matching the question content in the target data set, generating third target data based on the question content asked by the user; generating a third target file associated with the third target data based on the question content and the answer to the question asked by the user, wherein the text content of the third target file is used to represent the question answer related to the third target data; and updating the target data set based on the third target data and the third target file to obtain a new target data set.

[0010] According to embodiments of this disclosure, generating third target data based on the content of a user's question includes: constructing a third preset prompt word based on the content of the user's question; performing semantic analysis on the question content and the third preset prompt word to generate the third target data.

[0011] According to embodiments of this disclosure, generating a third target file based on the content and answer of a user's question includes: constructing a fourth preset prompt word based on the content and answer of the user's question; performing semantic analysis on the content, answer, and fourth preset prompt word to generate second target text content; and performing file format conversion processing on the second target text content to generate the third target file.

[0012] According to embodiments of this disclosure, updating a target data set based on second target data and a first target file to obtain a new target data set includes: querying a fourth target data that is consistent with the second target data from the target data set, and determining parameter information related to the fourth target data, wherein the parameter information includes at least one of the following: a data identifier of the fourth target data, a file status of the first target file, and a file identifier of the first target file; and updating the target data set based on at least one of the data identifier, the file status of the first target file, and the file identifier of the first target file to obtain a new target data set.

[0013] According to embodiments of this disclosure, the above method further includes: generating a new data identifier when the data identifier is empty; or updating the data identifier when the data identifier is not empty; determining that the file state of the first target file is saturated when the number of characters in the first target file is greater than or equal to a preset character threshold; or determining that the file state of the first target file is unsaturated when the number of characters in the first target file is less than a preset character threshold.

[0014] According to embodiments of this disclosure, the user question-and-answer content further includes a user identifier; based on the question content asked by the user, querying a target data set stored in a distributed database to see if there is first target data matching the question content includes: querying a target data set corresponding to the user identifier from the distributed database; querying at least one candidate data matching the question content from the target data set based on the question content asked by the user; extracting semantic features from the question content asked by the user to obtain a first feature vector; extracting semantic features from the at least one candidate data to obtain at least one second feature vector; calculating the similarity between the first feature vector and each second feature vector; and determining whether there is first target data matching the question content from the at least one candidate data based on the similarity calculation result.

[0015] Another aspect of this disclosure provides a data query method, comprising: responding to a query request for historical question content posed by a user, querying a data set corresponding to the target user identifier from a distributed database based on a target user identifier carried in the query request, the data set being obtained according to the aforementioned data processing method; querying target data associated with the historical question content from the data set based on the historical question content; and obtaining a target file from a cloud storage system based on a file identifier of a target file associated with the target data, wherein the text content of the target file is used to represent the question answer related to the historical question content.

[0016] Another aspect of this disclosure provides a data processing apparatus, comprising: a question-and-answer content acquisition module, configured to acquire user question-and-answer content from different intelligent question-and-answer applications based on a preset acquisition interface, wherein the user question-and-answer content includes the question content and the answer to the question; a first target data query module, configured to, in response to a query request for the question content of the user, query whether there is first target data matching the question content in a target data set stored in a distributed database; a first file generation module, configured to, in response to the existence of first target data matching the question content in the target data set, generate second target data and a first target file based on the first target data, the question content and the answer to the question, wherein the text content of the first target file is used to represent the question answer related to the second target data; and a data acquisition module, configured to update the target data set based on the second target data and the first target file to obtain a new target data set.

[0017] Another aspect of this disclosure provides a data query apparatus, comprising: a query request module, configured to, in response to a query request for historical question content posed by a user, query a data set corresponding to the target user identifier from a distributed database based on a target user identifier carried in the query request, the data set being obtained using the aforementioned data processing apparatus; a target data query module, configured to, based on the historical question content, query target data associated with the historical question content from the data set; and a target file acquisition module, configured to, based on a file identifier of a target file associated with the target data, acquire a target file from a cloud storage system, wherein the text content of the target file represents the answer to a question related to the historical question content.

[0018] Another aspect of this disclosure provides an electronic device comprising:

[0019] One or more processors;

[0020] Memory, used to store one or more programs.

[0021] Specifically, when one or more programs are executed by one or more processors, the one or more processors implement the above method.

[0022] Another aspect of this disclosure provides a computer-readable storage medium storing computer-executable instructions that, when executed, are used to implement the methods described above.

[0023] Another aspect of this disclosure provides a computer program product including computer-executable instructions that, when executed, are used to implement the methods described above.

[0024] According to embodiments of this disclosure, user question-and-answer content, including user-asked questions and answers, is collected from different intelligent question-and-answer applications using a preset acquisition interface. Based on the user-asked questions, when a first target data matching the question content is found in a target data set stored in a distributed database, the first target data, the question content, and the question answer are used to generate second target data and a first target file containing the question answers related to the second target data. The target data set is then updated based on the first target data and the first target file. This approach at least partially solves the technical problem of low user efficiency caused by users needing to switch between different intelligent question-and-answer applications and search multiple times to find previously asked questions. Furthermore, it achieves the technical effect of aggregating and classifying the question-and-answer content of different intelligent question-and-answer applications into a unified target data set, allowing users to search for historical questions in a unified manner and improving user efficiency. Attached Figure Description

[0025] The above and other objects, features and advantages of this disclosure will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0026] Figure 1 The illustration schematically shows an exemplary system architecture to which data processing methods, query methods, apparatuses, devices, media, and program products can be applied according to embodiments of the present disclosure;

[0027] Figure 2 A flowchart illustrating a data processing method according to an embodiment of the present disclosure is shown schematically.

[0028] Figure 3 A flowchart illustrating a data query method according to an embodiment of the present disclosure is shown schematically;

[0029] Figure 4 This illustration schematically shows a use case structure for aggregating information carried by question-and-answer content from different intelligent question-and-answer applications according to embodiments of the present disclosure;

[0030] Figure 5 A schematic diagram illustrating a data processing method according to an embodiment of the present disclosure is shown.

[0031] Figure 6 The message queue for generating user question-and-answer content according to an embodiment of this disclosure is illustrated schematically;

[0032] Figure 7 A schematic diagram illustrating the synthesis of a new problem title and a first object file according to an embodiment of this disclosure is shown.

[0033] Figure 8This illustration schematically shows a diagram illustrating the generation of a new problem title and a third object file according to an embodiment of this disclosure;

[0034] Figure 9 A schematic diagram illustrating a data query method according to an embodiment of the present disclosure is shown.

[0035] Figure 10 A block diagram of a data processing apparatus according to an embodiment of the present disclosure is shown schematically;

[0036] Figure 11 A block diagram schematically illustrates a data query apparatus according to embodiments of the present disclosure; and

[0037] Figure 12 A block diagram schematically illustrates an electronic device suitable for implementing a data processing method and a data query method according to embodiments of the present disclosure. Detailed Implementation

[0038] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0039] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0040] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0041] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0042] In the embodiments disclosed herein, the collection, updating, analysis, processing, use, transmission, provision, disclosure, and storage of data (e.g., including but not limited to user personal information) comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures have been taken to prevent unauthorized access to user personal information data and to safeguard user personal information security and network security.

[0043] In the embodiments disclosed herein, user authorization or consent is obtained before acquiring or collecting user personal information.

[0044] In related technologies, different business units within an enterprise have corresponding intelligent question-and-answer application systems. These systems store user sessions in their respective applications on a per-session basis. However, this approach has several drawbacks: identical questions from different sessions are stored repeatedly, and similar questions from the user's historical questions are also stored separately. Furthermore, user questions from different business units are scattered across their respective business applications. When a user needs to find answers to similar questions or previously asked questions, they must switch between different intelligent question-and-answer applications and search multiple times to find the answers, thus reducing user efficiency.

[0045] In view of this, embodiments of the present disclosure provide a data processing method, comprising: collecting user question and answer content from different intelligent question-and-answer applications based on a preset collection interface, wherein the user question and answer content includes the question content asked by the user and the answer to the question; responding to a query request for the question content asked by the user, querying whether there is first target data matching the question content from a target data set stored in a distributed database, based on the question content asked by the user; responding to the existence of first target data matching the question content in the target data set, generating second target data and a first target file based on the first target data, the question content asked by the user, and the answer to the question, wherein the text content of the first target file is used to represent the question answer related to the second target data; and updating the target data set based on the second target data and the first target file to obtain a new target data set.

[0046] Figure 1 This illustration schematically depicts an exemplary system architecture to which data processing methods, query methods, apparatuses, devices, media, and program products can be applied according to embodiments of this disclosure. It should be noted that... Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this disclosure, in order to help those skilled in the art understand the technical content of this disclosure, but do not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments or scenarios.

[0047] like Figure 1As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0048] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social media platform software, etc. (for example only).

[0049] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0050] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0051] It should be noted that the data processing method and data query method provided in the embodiments of this disclosure can generally be executed by server 105. Correspondingly, the data processing device and data query device provided in the embodiments of this disclosure can generally be located in server 105. The data processing method and data query method provided in the embodiments of this disclosure can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the data processing device and data query device provided in the embodiments of this disclosure can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Alternatively, the data processing method and data query method provided in the embodiments of this disclosure can also be executed by the first terminal device 101, the second terminal device 102, or the third terminal device 103, or by other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103. Accordingly, the data processing device and data query device provided in the embodiments of this disclosure may also be disposed in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or disposed in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0052] For example, the user question and answer content may originally be stored in any one of the first terminal device 101, the second terminal device 102, or the third terminal device 103 (e.g., the first terminal device 101, but not limited thereto), or it may be stored on an external storage device and imported into the first terminal device 101. Then, the first terminal device 101 may locally execute the data processing method and data query method provided in the embodiments of this disclosure, or send the user question and answer content to other terminal devices, servers, or server clusters, and have the other terminal devices, servers, or server clusters that receive the image to be processed execute the data processing method and data query method provided in the embodiments of this disclosure.

[0053] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0054] Figure 2 A flowchart illustrating a data processing method according to an embodiment of the present disclosure is shown schematically.

[0055] like Figure 2 As shown, the data processing method 200 includes operations S210 to S240.

[0056] When operating S210, user question and answer content is collected from different intelligent question and answer applications based on the preset collection interface. The user question and answer content includes the question content and the answer to the question asked by the user.

[0057] According to embodiments of this disclosure, the preset data collection interface can be an interface provided by a cloud storage system, used to collect user question-and-answer content from intelligent question-and-answer applications corresponding to different services. The user question-and-answer content can be questions raised by users through different intelligent question-and-answer applications and the answers provided to those questions.

[0058] According to embodiments of this disclosure, each business application is configured with its own intelligent question-and-answer (ICA) function. Users can ask questions to the corresponding ICA function for different business applications and obtain answers. For example, if a user wants to ask a question about a meeting topic, they can ask the question and obtain an answer based on the intelligent ICA function in the meeting application system; if a user wants to ask a question about personnel management, they can ask the question and obtain an answer based on the intelligent ICA function in the personnel management system, and so on.

[0059] According to embodiments of this disclosure, user question-and-answer content can be the content of historical questions asked by the user and their corresponding historical answers; or it can be the content of questions asked by the user in real time and their corresponding answers. After obtaining answers by asking questions through their respective intelligent question-and-answer applications configured in different business applications, the user's question-and-answer content is stored in their respective intelligent question-and-answer applications.

[0060] According to embodiments of this disclosure, the preset collection interface can be called to collect user question and answer content from different intelligent question and answer applications and store it in a cloud storage system. The user question and answer content may include at least one question and answer pair, and a message queue is generated based on the cloud storage system for at least one user question and answer content.

[0061] In operation S220, in response to a query request for the content of a question asked by a user, the system queries the target data set stored in the distributed database to see if there is first target data that matches the question content.

[0062] According to embodiments of this disclosure, a query request can be a request to query the content of a user's question within each user question-and-answer content in the message queue. The target dataset can be a clustering table formed by clustering the user's question content and answers in different intelligent question-and-answer applications. The data structure information of this target dataset includes question title, data identifier, file size, file saturation status, questioner, file identifier, creation time, and modification time. The question title is the title after clustering the user's question content, the text content in the file is the question answer corresponding to the clustered question title, and the data identifier is the title number corresponding to the question title.

[0063] According to embodiments of this disclosure, based on a cloud storage system subscribing to the aforementioned message queue, for each user question and answer content in the message queue, a query is performed from a distributed database to find first target data that matches the user question and answer content. The first target data may represent question titles in a target data set that are similar to the user question and answer content.

[0064] In operation S230, in response to the existence of first target data in the target data set that matches the question content, second target data and a first target file are generated based on the first target data, the question content asked by the user, and the question answer. The text content of the first target file is used to represent the question answer related to the second target data.

[0065] According to embodiments of this disclosure, the second target data may represent a new title generated by aggregating the aforementioned similar question titles with the content of the questions asked by the user; the first target file may represent a file of question answers related to the generated new title, which may be generated by aggregating the text content of the file associated with the first target data with the question answers.

[0066] According to embodiments of this disclosure, the generated first target file can be stored in a cloud storage system and a corresponding file identifier can be generated.

[0067] In operation S240, the target data set is updated based on the second target data and the first target file to obtain a new target data set.

[0068] According to embodiments of this disclosure, the data structure information of the row containing the first target data in the target data set includes the question title, data identifier, file size, file saturation status, questioner, file ID, creation time, and modification time.

[0069] According to embodiments of this disclosure, the issue title of the second target data, the file identifier of the first target file associated with the second target data, the file size of the first target file, and the file saturation status can be updated at the corresponding data structure information of the row containing the first target data, so that the data structure of the row containing the first target data is updated to the data structure of the second target data. Simultaneously, the data structure information of the row containing the first target data is stored in a target historical data set. The target historical data set is used to store the data structure before aggregation, so that it is possible to trace which data clusters resulted in the aggregated target data structure.

[0070] According to embodiments of this disclosure, the query time, modification time, and other information of the row containing the second target data are also updated simultaneously.

[0071] According to embodiments of this disclosure, user question-and-answer content, including user-asked questions and answers, is collected from different intelligent question-and-answer applications using a preset acquisition interface. Based on the user-asked questions, when a first target data matching the question content is found in a target data set stored in a distributed database, the first target data, the question content, and the question answer are used to generate second target data and a first target file containing the question answers related to the second target data. The target data set is then updated based on the first target data and the first target file. This approach at least partially solves the technical problem of low user efficiency caused by users needing to switch between different intelligent question-and-answer applications and search multiple times to find previously asked questions. Furthermore, it achieves the technical effect of aggregating and classifying the question-and-answer content of different intelligent question-and-answer applications into a unified target data set, allowing users to search for historical questions in a unified manner and improving user efficiency.

[0072] According to a disclosed embodiment, querying a target data set stored in a distributed database to determine whether there is first target data matching the question content based on the question content asked by a user includes: querying a target data set corresponding to a user identifier from the distributed database; querying at least one candidate data matching the question content from the target data set based on the question content asked by the user; extracting semantic features from the question content asked by the user to obtain a first feature vector; extracting semantic features from the at least one candidate data to obtain at least one second feature vector; calculating the similarity between the first feature vector and each second feature vector; and determining whether there is first target data matching the question content from the at least one candidate data based on the similarity calculation result.

[0073] According to embodiments of this disclosure, the user question-and-answer content also includes a user identifier. The user identifier is unique information used to identify a user. A list of question titles associated with historical questions asked by that user can be retrieved from a distributed database based on the user identifier, i.e., the target data set.

[0074] According to embodiments of this disclosure, each row in the target dataset contains the following data structure information: question title, data identifier, file size, file saturation status, questioner, file identifier, creation time, and modification time.

[0075] According to the embodiments of this disclosure, before sequentially matching the question titles of each row in the target data set with the question content asked by the user, it is first determined whether the target file corresponding to the question title of that row is in a saturated state. If it is in a saturated state, it means that the number of characters written to the target file has reached the preset number of characters, so the question title of that row will no longer be matched with the question content asked by the user, and the next row will be executed. If it is in an unsaturated state, it means that the number of characters written to the target file has reached the preset number of characters, and then the question content asked by the user and the question title in that row are calculated for similarity. If the similarity result reaches the similarity threshold, that row can be used as a candidate list, and the question title in that row can be used as a candidate title, i.e., candidate data. Each row in the target data set is matched sequentially to obtain at least one candidate data.

[0076] According to an embodiment of this disclosure, based on a large language model, at least one candidate data and the content of the user's question are input into the large language model, and semantic feature extraction is performed on the content of the user's question and at least one candidate data to obtain a first feature vector and at least one second feature vector.

[0077] According to embodiments of this disclosure, the first feature vector is compared with at least one second feature vector to calculate the vector similarity, thereby obtaining at least one similarity.

[0078] According to embodiments of this disclosure, candidate data corresponding to the second feature vector with the largest similarity value in at least one similarity score is determined as the first target data, i.e., the question title that is similar to the content of the question asked by the user.

[0079] According to embodiments of this disclosure, in response to the existence of first target data matching the question content in the target data set, generating second target data and a first target file based on the first target data, the question content asked by the user, and the question answer includes: in response to the existence of first target data matching the question content, generating second target data based on the first target data and the question content asked by the user; obtaining a second target file from a cloud disk system based on the file identifier of the second target file associated with the first target data; and generating a first target file based on the question content asked by the user, the question answer, and the second target file.

[0080] According to embodiments of this disclosure, the first target data, the content of the question asked by the user, and the prompt words entered by the user can be used as input to a large language model to process the first target data, the content of the question asked by the user, and the prompt words entered by the user, and output the second target data.

[0081] According to an embodiment of this disclosure, the first target data is a new question title related to the content of the question asked by the user.

[0082] According to embodiments of this disclosure, the text content of the second target file is a question title retrieved from the target data set that is similar to the question asked by the user, i.e., the answer to the question in the first target data, which is stored in a cloud storage system. The second target file associated with the first target data can be retrieved from the cloud storage system based on the file identifier of the row containing the first target data.

[0083] According to embodiments of this disclosure, the user's question content, the question answer, the second target file, and the prompt words are input into a large language model for aggregation, the text content of the first target file is output, and then the text content of the first target file is converted to generate a first target file in a preset format.

[0084] According to embodiments of this disclosure, the prompt words can be instructions on how a large language model can use the content of a user's question, first target data, question answer, and second target file to generate second target data and first target file.

[0085] According to embodiments of this disclosure, a new target file is generated by aggregating the question content, the answer to the question, and the first target file associated with the first target data. A new question title is generated by aggregating the question content and the first target data. This solves the problem of repeated storage of repeated sessions in multiple sessions of a user in an intelligent question-and-answer application or in intelligent question-and-answer applications from different sources. Multiple similar or identical question content and answers are aggregated into a question title and a target file, making it easier for users to find.

[0086] According to embodiments of this disclosure, generating second target data based on first target data and user-asked questions includes: constructing first preset prompt words based on the first target data and user-asked questions; performing multimodal feature fusion on the first target data, question content, and first preset prompt words to obtain multimodal fusion features; performing hierarchical feature extraction on the multimodal features to obtain hierarchical features; and generating second target data based on the hierarchical features.

[0087] According to embodiments of this disclosure, the first preset prompt word may be pre-constructed by the user or generated by a large language model based on the style type of the first target data and the domain characteristics of the question content asked by the user.

[0088] According to embodiments of this disclosure, a matching basic template can be retrieved from a preset prompt word template library based on the domain of the user's question content and the style type of the first target data. Then, the keywords in the user's question content and the style type of the first target data are filled into the basic template to generate a first preset prompt word.

[0089] According to embodiments of this disclosure, the style type of the first target data may be the basic sentence form, sentence length, semantic and sentiment features, etc. of the first target data; the domain to which the question content belongs may be the domain to which the question content asked by the user belongs, such as science and technology, finance, etc.

[0090] According to embodiments of this disclosure, the first target data, question content, and first preset prompt word can be encoded separately using the input encoding layer of a large language model, resulting in their respective encoded feature vectors. A cross-membrane attention mechanism is then used to interactively focus on the feature vectors of the user's question and the first target data, dynamically adjusting the fusion weights to obtain a fused vector. The feature vector of the first preset prompt word is then incorporated into the fused vector, forming a multimodal fused feature with guiding feature representation.

[0091] According to embodiments of this disclosure, hierarchical feature extraction may include surface feature extraction, semantic feature extraction, and style feature extraction. Surface feature extraction can determine the basic framework of the generated new question title, i.e., the second target data, such as the length range, sentence structure, and keyword distribution of the new question title; semantic feature extraction can determine that the generated second target data is consistent with the user's question in intent and sentiment, and conforms to the expression habits of the relevant domain; style feature extraction can mimic the rhetoric of the first target data, such as font size, so that the second target data has a similar style.

[0092] According to embodiments of this disclosure, second target data is generated based on the parsing of hierarchical features extracted from multimodal features.

[0093] According to embodiments of this disclosure, by taking a first target data similar to the content of a user's question as a reference and using the content of the user's question as the core, the first target data and the question content are aggregated through a large language model to generate a new question title, namely the second target data, so that multiple similar questions can be combined into one question, which is convenient for unified recording and easy for users to find.

[0094] According to embodiments of this disclosure, generating a first target file based on a user-asked question, the answer to the question, and a second target file includes: extracting text features from the second target file to obtain the text content of the second target file; constructing a second preset prompt word based on the user-asked question, the answer to the question, and the text content of the second target file; aggregating the user-asked question, the answer to the question, the text content of the second target file, and the second preset prompt word to obtain the first target text content; and performing format conversion processing on the first target text content to generate the first target file.

[0095] According to embodiments of this disclosure, a second target file associated with the first target data can be read from a cloud storage system, and the text content of the second target file can be extracted. This text content consists of question-and-answer responses to question titles similar to those in the first target data.

[0096] According to embodiments of this disclosure, the second preset prompt can be constructed based on the content of the user's question, the answer to the question, and the text content of the second target file. The construction process and basis are similar to the aforementioned process of constructing the first preset prompt, and will not be repeated here.

[0097] According to embodiments of this disclosure, the user's question content, the question answer, the text content of the second target file, and the second preset prompt words are preprocessed, including deduplication, and then used as input to a large language model for aggregation processing, outputting the text content of the first target text. The text content of the first target file is the question answer associated with the aforementioned second target data.

[0098] According to embodiments of this disclosure, the text content of the generated first target file is converted into a second target file using a preset text format. The second target file and its identifier are then stored in a cloud storage system.

[0099] According to embodiments of this disclosure, a first target file is generated by aggregating the user's question content, the question answer, and the question answer related to the first target data retrieved from the target data aggregation using a large language model. The question answers of similar questions are then integrated to generate a second target file. This allows similar questions from different intelligent question-answering applications to be generated and their corresponding target files to be integrated into a unified target file, making it easier for users to find the question answers of previously asked questions and improving operational efficiency.

[0100] According to embodiments of this disclosure, the method further includes: in response to the absence of first target data matching the question content in the target data set, generating third target data based on the question content asked by the user; generating a third target file associated with the third target data based on the question content and the answer to the question asked by the user, wherein the text content of the third target file is used to represent the question answer related to the third target data; and updating the target data set based on the third target data and the third target file to obtain a new target data set.

[0101] According to embodiments of this disclosure, if no question title similar to the content of the user's question is found in the target data set, a new question title, i.e., a third target data, can be generated based on the content of the user's question.

[0102] According to embodiments of this disclosure, generating third target data based on the content of a user's question includes: constructing a third preset prompt word based on the content of the user's question; performing semantic analysis on the question content and the third preset prompt word to generate the third target data.

[0103] According to embodiments of this disclosure, a large language model can be invoked, and the content of the user's question and the constructed third preset prompt words can be used as input to the large language model. Semantic analysis is performed on the content of the user's question, and a new question title, i.e., the third target data, is output.

[0104] According to embodiments of this disclosure, generating a third target file based on the content and answer of a user's question includes: constructing a fourth preset prompt word based on the content and answer of the user's question; performing semantic analysis on the content, answer, and fourth preset prompt word to generate second target text content; and performing file format conversion processing on the second target text content to generate the third target file.

[0105] According to embodiments of this disclosure, a fourth preset prompt word is used to guide the large language model on how to generate a third target text based on the content of the user's question and the answer to the question.

[0106] According to embodiments of this disclosure, the user's question content, the answer to the question, and a fourth preset prompt word are used as input to a large language model. Semantic analysis is performed on the user's question content and the answer to the question, and after being fused with the fourth preset prompt word, the text content of the third target file is output.

[0107] According to embodiments of this disclosure, the text content of the third target file is used to represent the answer to the question regarding the generation of the new title, i.e., the third target data.

[0108] According to embodiments of this disclosure, the text content of a third target file can be converted into a preset text format based on a preset text format to generate a third target file, which is then stored in a cloud storage system, and a file identifier for the third target file is generated.

[0109] According to embodiments of this disclosure, if it is determined that there is no question title in the target data set that is similar to the question content asked by the user, a new title is generated from the question content asked by the user and stored in the target data set. A third target file in a preset format is generated based on the question content asked by the user and the question-and-answer content and stored in the cloud storage system. This allows the question content asked by the user in different intelligent question-and-answer applications to be recorded in the target data set, and the generated third target file is stored in the cloud storage system. This makes it easier for the user to avoid switching between different intelligent question-and-answer applications multiple times when they need to query the answer to a question they asked before, thus improving operational efficiency.

[0110] According to embodiments of this disclosure, updating a target data set based on second target data and a first target file to obtain a new target data set includes: querying a fourth target data that is consistent with the second target data from the target data set, and determining parameter information related to the fourth target data, wherein the parameter information includes at least one of the following: a data identifier of the fourth target data, a file status of the first target file, and a file identifier of the first target file; and updating the target data set based on at least one of the data identifier, the file status of the first target file, and the file identifier of the first target file to obtain a new target data set.

[0111] According to an embodiment of this disclosure, the second target data is a new question title generated by aggregating similar question titles with the question content when a question title similar to the question content asked by the user is found in the target data set.

[0112] According to embodiments of this disclosure, based on the determined second target data, a query is performed from the target data set to determine whether there is fourth target data that is completely consistent with the second target data. If it exists, all data structure information of the row containing the fourth target data is returned; if it does not exist, an empty table row representing the fourth target data is returned, so that the aggregated new question title, i.e., the relevant information of the second target data, can be filled into the empty table row to obtain all data structure information for the row containing the second target data.

[0113] According to embodiments of this disclosure, when the data identifier is empty, a new data identifier is generated; or when the data identifier is not empty, the data identifier is updated; when the number of characters in the first target file is greater than or equal to a preset character threshold, the file state of the first target file is determined to be saturated; or when the number of characters in the first target file is less than the preset character threshold, the file state of the first target file is determined to be unsaturated.

[0114] According to the embodiments of this disclosure, if fourth target data exists, and after returning all data structure information of the row containing the fourth target data, the data identifier of the row containing the fourth target data is incremented by 1 to generate a new data identifier; if fourth target data does not exist, after returning an empty table row, the second target data is written into the empty table row as the problem title, and the data identifier is 1.

[0115] According to embodiments of this disclosure, the number of characters in the first target file corresponding to the question answer of the second target data is calculated. If the number of characters is greater than or equal to a preset character threshold, it indicates that the first target file has reached a saturated state, and the file status of the row containing the second target data is updated to a saturated state. If the number of characters in the first target file is less than the preset character threshold, it is updated to an unsaturated state. Simultaneously, the file identifier, file size, and modification time of the first target file corresponding to the question answer of the second target data in the row containing the second target data are updated.

[0116] According to embodiments of this disclosure, all data structure information of the row containing question titles similar to the user's question is stored in a historical target data set, forming a correspondence with the second target data. When a user wants to query which questions were aggregated from the second target data, they can do so through the second target data.

[0117] According to embodiments of this disclosure, by merging and storing user-asked questions in a target dataset, and by matching the merged question titles with the original question content, users can trace their past questions through the merged question titles, thereby improving user efficiency.

[0118] It should be noted that there are no specific limitations on the large language model used in the aforementioned data processing process; any large language model capable of performing the aforementioned data processing is acceptable.

[0119] Figure 3 A flowchart illustrating a data query method according to an embodiment of the present disclosure is shown schematically.

[0120] like Figure 3 As shown, the data query method 300 may include operations S310 to S330.

[0121] In operation S310, in response to a query request for the content of a user’s historical questions, the system retrieves a data set corresponding to the target user identifier from the distributed database based on the target user identifier carried in the query request. The data set is obtained according to the data processing method described above.

[0122] In operation S320, target data related to the historical problem content is queried from the dataset based on the historical problem content.

[0123] In operation S330, the target file is retrieved from the cloud disk system based on the file identifier of the target file associated with the target data. The text content of the target file is used to represent the answer to the question related to the historical question content.

[0124] According to embodiments of this disclosure, when a target user queries historical questions previously asked from a cloud storage system, the system can retrieve, page by page, a data set associated with that target user from a distributed database based on the target user's user identifier. This data set records the data structure information corresponding to all historical questions previously asked by the target user.

[0125] According to embodiments of this disclosure, a target user queries a dataset for the title of a question related to a historical question, i.e., the relevant target data, as well as other data structure information of the row containing the target data, based on the historical question content to be queried. Based on the file identifier in the other data structure information, the target user retrieves the target file of the question answer associated with the target data from the cloud storage system.

[0126] According to embodiments of this disclosure, target users can read, share, or download the target file online. Simultaneously, target users can view historical data sets related to the target data in the dataset to trace the historical data from which the target data was merged.

[0127] According to embodiments of this disclosure, a target user can also transmit a target file containing the question answers associated with the target data to an intelligent question answering application, ask related questions based on the transmitted target file, and generate corresponding question answers.

[0128] According to embodiments of this disclosure, by collecting the question-and-answer content of different smart question-and-answer applications into a unified cloud storage system and aggregating the collected question-and-answer content, when a target user queries historical questions, the question-and-answer content of different smart question-and-answer applications is aggregated into a unified target data set for searching. This avoids the low operational efficiency caused by users switching between different applications and searching multiple times, thereby improving the user's search efficiency.

[0129] Figure 4The illustration schematically depicts a use case structure for aggregating information carried by question-and-answer content from different intelligent question-and-answer applications according to embodiments of the present disclosure.

[0130] like Figure 4 As shown, the use case structure may include intelligent question-and-answer applications, which are mainly used to generate answers to user-asked questions; aggregated message queues; target file information in cloud storage, mainly including a list of target files, and based on these target files, intelligent question-and-answer, reading / sharing / downloading of target files, and querying of historical target files; target files; data structure information synthesis of the rows containing the target data, mainly including title synthesis strategies to synthesize new titles, title numbering strategies to generate new title numbers, target file saturation calculation, and target file synthesis strategies; and data sets, mainly including target data sets and target historical data sets.

[0131] Figure 5 A schematic diagram of a data processing method according to an embodiment of the present disclosure is shown.

[0132] like Figure 5 As shown, the user asks a question through the intelligent question-and-answer application, receives an answer, and generates user question-and-answer content 501; the cloud storage system calls a preset collection interface to collect the user question-and-answer content and generate a message queue (MQ, Message Queue) 502. The cloud storage system subscribes to the message queue 503; based on the user's question content, it queries at least one (e.g., K) candidate question titles that are similar to the user's question content from the target data set stored in the distributed database 504; it calls a large language model to calculate the similarity between the user's question content and the K question titles, and determines whether there is a most similar question title among the aforementioned K similar candidate question titles based on the similarity calculation result 505.

[0133] If the most similar question title exists, call the large language model to synthesize a new question title based on the user's question content and the similar question title 506; call the large language model to generate a first target file based on the second target file with similar question titles, the user's question content and the question answer 507; and upload the first target file to the cloud storage system for storage 508; and record all data structure information of the row where the similar question title is located in the target historical data set 509; according to the synthesized new question title, query whether there is a question title in the target data set that is consistent with the new question title 510; if it exists, update the title number of the question title that is consistent with the new question title 511; if it does not exist, the title number is 1 (512); at the same time, update the file identifier, file size, file saturation status and modification time of the row where the question title that is consistent with the new question title is located 513.

[0134] If no similar question title exists, the large language model is invoked to generate a new title based on the content of the user's question 514; a third target file is generated based on the content of the user's question and the answer 515; and the third target file is stored in the cloud storage system 516; based on the generated new question title, the target data set is queried to see if there is a question title that matches the new question title 510; if it exists, the title number of the question title that matches the new question title is updated 511; if it does not exist, the title number is 1 (512); at the same time, the file identifier, file size, file saturation status, and creation time of the line containing the question title that matches the new question title are written 517.

[0135] Figure 6 The message queue for generating user question-and-answer content according to an embodiment of this disclosure is illustrated schematically; Figure 7 A schematic diagram illustrating the synthesis of a new problem title and a first object file according to an embodiment of this disclosure is shown.

[0136] like Figure 6 As shown, the user asks a question 601 through the intelligent question and answer application, and the intelligent question and answer application generates an answer 602 for the question. The cloud storage system calls a preset collection interface to collect input parameter information: the user's identifier, the content of the question asked by the user, and the answer to the question 603. The cloud storage application generates a message queue 604 from the user identifier, the content of the question asked by the user, and the answer to the question, and sends the message queue to the MQ service 605.

[0137] like Figure 7As shown, the cloud storage service subscribes to the message queue 701; the MQ service sends a message queue to the cloud storage service 702; the cloud storage service calls the data generation interface, and based on the user's question content in the user Q&A content in the message queue, and based on the filtering condition: file saturation state = unsaturation state, queries the target data set stored in the distributed database 703; returns a list of rows containing the top K similar titles to the cloud storage service 704, where K is an integer greater than or equal to 1; calls the large language model to determine whether similar titles exist from the list of rows containing the top K similar titles 705; if they exist, calls the large language model to synthesize a new question title 706 based on the user's question content and similar question titles, and returns the new question title 707; downloads the second target file 707-1 from the cloud storage based on the file identifier of the second target file associated with the similar question titles; and returns the second target file 707-2 to the cloud storage service; aggregates the second target file, the user's question content and answer, and the prompt words to generate the first target file 708; and returns the first target file 708 to the cloud storage service. File 709; after uploading the first target file to the cloud storage system, return the file identifier of the first target file to the cloud storage service 710; at the same time, according to the synthesized new question title, accurately query the data with the same title from the target data set 711; and return the data structure information of the row where the same title is located to the cloud storage service 712; determine the title number N1 of the same title, if it is empty, write the title number to 1; if it is not empty, update the original title number to the original title number + 1 (713); calculate the saturation state of the first target file associated with the synthesized new question title, if the number of characters in the file is greater than or equal to the preset character threshold, set it to the saturation state, otherwise it is the unsaturated state 714; at the same time, record the data structure information of the row where the question title is similar to the question content asked by the user to the target historical data aggregation related to the aforementioned same title 715; and return the execution result to the cloud storage service 716; at the same time, update the title number, file identifier, file size, file saturation state and modification time of the row where the same title is located 717, to obtain the updated target data set; and return the execution result to the cloud storage service 718. The data structure information of the row containing the new question title in the updated target dataset should include at least the following: question title, title number, file identifier, questioner, question time, file saturation status, and modification time.

[0138] Figure 8 The illustration shows a schematic diagram of generating a new problem title and a third object file according to an embodiment of the present disclosure.

[0139] Combination Figure 7As shown, after executing steps 701-705, if no question title similar to the user's question content exists, the large language model is invoked to generate a new question title 801 based on the user's question content and prompts, and the generated new question title 802 is returned; the large language model is then invoked to generate a third target file 803 based on the user's question content, answer, and prompts, and the third target file 804 is returned to the cloud storage service; after uploading the third target file to the cloud storage system, the file identifier of the third target file 805 is returned to the cloud storage service; simultaneously, based on the generated new question title, data with the same title is precisely queried from the target data set 806; and... Return the data structure information of the row containing the same title to the cloud storage service 807; determine the title number N1 of the same title, if it is empty, write the title number to 1; if it is not empty, update the original title number to the original title number + 1 (808); calculate the saturation state of the first target file associated with the generated new question title, if the number of characters in the file is greater than or equal to the preset character threshold, set it to saturation state, otherwise set it to unsaturation state 809; at the same time, write the title number, file identifier, file size, file saturation state and modification time of the row containing the same title into the target data set 810, generate a new target data set; and return the execution result to the cloud storage service 811. The data structure information of the row containing the new question title in the generated new target data set includes at least: question title, title number, file identifier, questioner, question time, file saturation state, creation time, etc.

[0140] Figure 9 A schematic diagram of a data query method according to an embodiment of the present disclosure is shown.

[0141] like Figure 9 As shown, users can send a request to the cloud storage service to query historical questions. Based on the received target user identifier and the condition that the questioner = the current user, the cloud storage service reads the target data set 901 related to the target user identifier from the cloud storage service. According to the content of the historical questions queried by the user, it determines the file identifier associated with the historical question content from the row containing the question title related to that historical question in the target data set, retrieves the target file corresponding to the file identifier from the cloud storage 902, and returns the query result 903 to the cloud storage service. Users can then read / share / download the target file online 904. Simultaneously, users can also query the historical data corresponding to the target file using the business primary key corresponding to the current historical question 905, and return the historical data set 906 to the cloud storage service. Furthermore, users can upload the target file they read / share / download online to the intelligent question-answering application 907 and ask a question based on the uploaded target file 908; the intelligent question-answering application generates a corresponding answer 909 for the question asked about the target file.

[0142] Figure 10 A block diagram of a data processing apparatus according to an embodiment of the present disclosure is shown schematically.

[0143] like Figure 10 As shown, the data processing device 1000 includes: a question and answer content acquisition module 1010, a first target data query module 1020, a first file generation module 1030, and a data acquisition module 1040.

[0144] The question and answer content collection module 1010 is used to collect user question and answer content from different intelligent question and answer applications based on a preset collection interface. The user question and answer content includes the question content and the answer to the question asked by the user.

[0145] The first target data query module 1020 is used to respond to a query request for the content of a question asked by a user, and to query whether there is first target data that matches the content of the question from the target data set stored in the distributed database.

[0146] The first file generation module 1030 is used to generate second target data and a first target file in response to the existence of first target data in the target data set that matches the question content. The second target data is used to generate second target data and a first target file based on the first target data, the question content asked by the user, and the question answer. The text content of the first target file is used to represent the question answer related to the second target data.

[0147] The data acquisition module 1040 is used to update the target data set based on the second target data and the first target file to obtain a new target data set.

[0148] According to embodiments of this disclosure, the first file generation module 1030 includes: a data generation submodule, a target acquisition submodule, and a file generation submodule.

[0149] The data generation submodule is used to generate second target data in response to the existence of first target data that matches the question content, based on the first target data and the question content asked by the user.

[0150] The target acquisition submodule is used to obtain the second target file from the cloud disk system based on the file identifier of the second target file associated with the first target data.

[0151] The file generation submodule is used to generate the first target file based on the user's question, the answer to the question, and the second target file.

[0152] According to embodiments of this disclosure, the data generation submodule includes: a prompt word first construction unit, a feature fusion unit, a feature extraction unit, and a data generation unit.

[0153] The first prompt word construction unit is used to construct the first preset prompt word based on the first target data and the content of the user's question.

[0154] The feature fusion unit is used to perform multimodal feature fusion on the first target data, the question content and the first preset prompt words to obtain multimodal fused features.

[0155] The feature extraction unit is used to perform hierarchical feature extraction on multimodal features to obtain hierarchical features.

[0156] The data generation unit is used to generate second target data based on hierarchical features.

[0157] According to embodiments of this disclosure, the file generation submodule includes: a first content acquisition unit, a second prompt word construction unit, a second content acquisition unit, and a file generation unit.

[0158] The first content acquisition unit is used to extract text features from the second target file to obtain the text content of the second target file.

[0159] The second prompt word construction unit is used to construct a second preset prompt word based on the content of the user's question, the answer to the question, and the text content of the second target file.

[0160] The second content acquisition unit is used to aggregate the user's question content, the question answer, the text content of the second target file, and the second preset prompt words to obtain the first target text content.

[0161] The file generation unit is used to perform format conversion processing on the first target text content to generate the first target file.

[0162] According to embodiments of this disclosure, the apparatus 1000 further includes: a data generation module, a second file generation module, and a data set acquisition module.

[0163] The data generation module is used to generate third target data based on the user's question content in response to the absence of first target data matching the question content in the target data set.

[0164] The second file generation module is used to generate a third target file associated with the third target data based on the content of the user's question and the answer to the question. The text content of the third target file is used to represent the answer to the question related to the third target data.

[0165] The data set acquisition module is used to update the target data set based on the third target data and the third target file to obtain a new target data set.

[0166] According to embodiments of this disclosure, the data generation module includes: a prompt word first construction submodule and a semantic analysis submodule.

[0167] The first prompt word construction submodule is used to construct a third preset prompt word based on the content of the user's question.

[0168] The semantic analysis submodule is used to perform semantic analysis on the question content and the third preset prompt words to generate third target data.

[0169] According to embodiments of this disclosure, the second document generation module includes: a second prompt word construction submodule, a text content generation submodule, and a conversion processing submodule.

[0170] The second prompt word construction submodule is used to construct a fourth preset prompt word based on the content of the user's question and the answer to the question.

[0171] The text content generation submodule is used to perform semantic analysis on the user's question, the answer to the question, and the fourth preset prompt word to generate the second target text content;

[0172] The conversion processing submodule is used to convert the file format of the second target text content to generate a third target file.

[0173] According to embodiments of this disclosure, the data acquisition module 1040 includes: a parameter information determination submodule and a data set update submodule.

[0174] The parameter information determination submodule is used to query the fourth target data that is consistent with the second target data from the target data set based on the second target data, and determine the parameter information related to the fourth target data. The parameter information includes at least one of the following: the data identifier of the fourth target data, the file status of the first target file, and the file identifier of the first target file.

[0175] The data set update submodule is used to update the target data set based on at least one of the data identifier, the file status of the first target file, and the file identifier of the first target file, so as to obtain a new target data set.

[0176] According to embodiments of this disclosure, the data acquisition module 1040 further includes an identifier update submodule and a status determination submodule.

[0177] The identifier update submodule is used to generate a new data identifier when the data identifier is empty, or to update the data identifier when the data identifier is not empty.

[0178] The state determination submodule is used to determine the file state of the first target file as saturated if the number of characters in the first target file is greater than or equal to a preset character threshold; or to determine the file state of the first target file as unsaturated if the number of characters in the first target file is less than the preset character threshold.

[0179] According to embodiments of this disclosure, the first target data query module 1020 includes: a data query submodule, a candidate data matching submodule, a vector acquisition submodule, a similarity calculation submodule, and a data determination submodule.

[0180] The data query submodule is used to query the target data set corresponding to the user identifier from the distributed database.

[0181] The candidate data matching submodule is used to query at least one candidate data that matches the content of the question asked by the user from the target data set.

[0182] The vector acquisition submodule is used to extract semantic features from the content of the user's question to obtain a first feature vector; and to extract semantic features from at least one candidate data to obtain at least one second feature vector.

[0183] The similarity calculation submodule is used to calculate the similarity between the first feature vector and each second feature vector.

[0184] The data determination submodule is used to determine, based on the similarity calculation results, whether there is a first target data that matches the content of the question from at least one candidate data.

[0185] Figure 11 A block diagram of a data query apparatus according to an embodiment of the present disclosure is shown schematically.

[0186] like Figure 11 As shown, the data processing device 1100 includes: a query request module 1110, a target data query module 1120, and a target file acquisition module 1130.

[0187] The query request module 1110 is used to respond to a query request for the content of a user’s historical questions, and to query a data set corresponding to the target user identifier from the distributed database based on the target user identifier carried in the query request. The data set is obtained from the aforementioned data processing device.

[0188] The target data query module 1120 is used to query target data related to the historical problem content from the data set based on the historical problem content.

[0189] The target file acquisition module 1130 is used to acquire the target file from the cloud disk system based on the file identifier of the target file associated with the target data. The text content of the target file is used to represent the answer to the question related to the historical question content.

[0190] Any one or more of the modules, submodules, and units according to embodiments of this disclosure, or at least part of the functions of any one or more of them, can be implemented in one module. Any one or more of the modules, submodules, and units according to embodiments of this disclosure can be implemented by dividing them into multiple modules. Any one or more of the modules, submodules, and units according to embodiments of this disclosure can be at least partially implemented as hardware circuitry, such as a Field-Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a System-on-Chip, a System-on-a-Substrate, a System-on-Package, an Application-Specific Integrated Circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging the circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, one or more of the modules, submodules, and units according to embodiments of this disclosure can be at least partially implemented as computer program modules, which, when run, can perform corresponding functions.

[0191] For example, any and multiple modules among the question-and-answer content acquisition module 1010, the first target data query module 1020, the first file generation module 1030, and the data acquisition module 1040, or the query request module 1110, the target data query module 1120, and the target file acquisition module 1130, can be combined into one module / unit / subunit, or any one of these modules / submodules / units can be split into multiple modules / submodules / units. Alternatively, at least some of the functions of one or more of these modules / submodules / units can be combined with at least some of the functions of other modules / submodules / units and implemented in one module / submodule / unit. According to embodiments of this disclosure, at least one of the question-and-answer content acquisition module 1010, the first target data query module 1020, the first file generation module 1030, and the data acquisition module 1040, or the query request module 1110, the target data query module 1120, and the target file acquisition module 1130, can be at least partially implemented as hardware circuits, such as field-programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems-on-a-chip, systems-on-a-substrate, systems-on-package, application-specific integrated circuits (ASICs), or any other reasonable means of integrating or packaging circuits, or implemented in hardware or firmware, or in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, at least one of the question-and-answer content acquisition module 1010, the first target data query module 1020, the first file generation module 1030, and the data acquisition module 1040, or the query request module 1110, the target data query module 1120, and the target file acquisition module 1130, can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0192] It should be noted that the data processing device and data query device parts in the embodiments of this disclosure correspond to the data processing method and data query method parts in the embodiments of this disclosure. For a detailed description of the data processing device and data query device parts, please refer to the data processing method and data query method parts, which will not be repeated here.

[0193] Figure 12 A block diagram schematically illustrates an electronic device suitable for implementing a data processing method and a data query method according to embodiments of the present disclosure. Figure 12 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0194] like Figure 12As shown, an electronic device 1200 according to an embodiment of the present disclosure includes a processor 1201, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1202 or a program loaded from a storage portion 1208 into a random access memory (RAM) 1203. The processor 1201 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 1201 may also include onboard memory for caching purposes. The processor 1201 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0195] RAM 1203 stores various programs and data required for the operation of electronic device 1200. Processor 1201, ROM 1202, and RAM 1203 are interconnected via bus 1204. Processor 1201 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 1202 and / or RAM 1203. It should be noted that programs may also be stored in one or more memories other than ROM 1202 and RAM 1203. Processor 1201 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in one or more memories.

[0196] According to embodiments of this disclosure, the electronic device 1200 may further include an input / output (I / O) interface 1205, which is also connected to a bus 1204. The electronic device 1200 may also include one or more of the following components connected to the input / output (I / O) interface 1205: an input section 1206 including a keyboard, mouse, etc.; an output section 1207 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card such as a LAN card, modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the input / output (I / O) interface 1205 as needed. A removable medium 1211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 1210 as needed so that computer programs read from it can be installed into the storage section 1208 as needed.

[0197] According to embodiments of this disclosure, the method flow according to embodiments of this disclosure can be implemented as a computer software program. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1209, and / or installed from removable medium 1211. When the computer program is executed by processor 1201, it performs the functions defined in the system of embodiments of this disclosure. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0198] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0199] According to embodiments of this disclosure, the computer-readable storage medium can be a non-volatile computer-readable storage medium. Examples include, but are not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0200] For example, according to embodiments of this disclosure, a computer-readable storage medium may include the ROM 1202 and / or RAM 1203 described above and / or one or more memories other than ROM 1202 and RAM 1203.

[0201] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods provided in the embodiments of this disclosure. When the computer program product is run on an electronic device, the program code enables the electronic device to implement the data processing method and data query method provided in the embodiments of this disclosure.

[0202] When the computer program is executed by the processor 1201, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0203] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 1209, and / or installed from the removable medium 1211. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0204] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on a user's computing device, partially on a user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0205] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions. Those skilled in the art will understand that the features described in the various embodiments of the present disclosure can be combined and / or combined in various ways, even if such combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments of this disclosure may be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0206] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A data processing method, comprising: User question and answer content is collected from different intelligent question and answer applications based on a preset collection interface. The user question and answer content includes the question asked by the user and the answer to the question. In response to a query request for the content of a question asked by the user, the system queries the target data set stored in the distributed database to see if there is first target data that matches the content of the question. In response to the existence of first target data in the target data set that matches the question content, a second target data and a first target file are generated based on the first target data, the question content asked by the user, and the question answer, wherein the text content of the first target file is used to represent the question answer related to the second target data; The target data set is updated based on the second target data and the first target file to obtain a new target data set.

2. The method according to claim 1, wherein, The step of responding to the existence of first target data matching the question content in the target data set, and generating second target data and a first target file based on the first target data, the question content asked by the user, and the question answer, includes: In response to the existence of first target data that matches the question content, second target data is generated based on the first target data and the question content asked by the user; Based on the file identifier of the second target file associated with the first target data, obtain the second target file from the cloud disk system; A first target file is generated based on the content of the user's question, the answer to the question, and the second target file.

3. The method according to claim 2, wherein, The step of generating second target data based on the first target data and the content of the user's question includes: Based on the first target data and the content of the question asked by the user, a first preset prompt word is constructed; Multimodal feature fusion is performed on the first target data, the question content, and the first preset prompt words to obtain multimodal fused features; Hierarchical feature extraction is performed on the multimodal features to obtain hierarchical features; Based on the hierarchical features, second target data is generated.

4. The method according to claim 2, wherein, The step of generating a first target file based on the user's question content, the question answer, and the second target file includes: Text features are extracted from the second target file to obtain the text content of the second target file; Based on the content of the user's question, the answer to the question, and the text content of the second target file, a second preset prompt word is constructed; The user-asked question, the answer to the question, the text content of the second target file, and the second preset prompt words are aggregated to obtain the first target text content; The first target text content is formatted and converted to generate the first target file.

5. The method according to claim 1, further comprising: In response to the absence of first target data matching the question content in the target data set, a third target data is generated based on the question content asked by the user. Based on the question asked by the user and the answer to the question, a third target file associated with the third target data is generated, wherein the text content of the third target file is used to represent the question and answer related to the third target data; The target data set is updated based on the third target data and the third target file to obtain a new target data set.

6. The method according to claim 5, wherein, The step of generating third target data based on the content of the user's question includes: Based on the content of the question asked by the user, a third preset prompt word is constructed; Semantic analysis is performed on the question content and the third preset prompt word to generate third target data.

7. The method according to claim 5, wherein, The step of generating a third target file based on the user's question and the answer to the question includes: Based on the content of the user's question and the answer to the question, a fourth preset prompt word is constructed; Semantic analysis is performed on the user's question, the answer, and the fourth preset prompt words to generate second target text content; The second target text content is converted into a file format to generate a third target file.

8. The method according to claim 1, wherein, The step of updating the target data set based on the second target data and the first target file to obtain a new target data set includes: Based on the second target data, query the target data set for fourth target data that is consistent with the second target data, and determine the parameter information related to the fourth target data. The parameter information includes at least one of the following: the data identifier of the fourth target data, the file status of the first target file, and the file identifier of the first target file. The target data set is updated based on at least one of the data identifier, the file status of the first target file, and the file identifier of the first target file to obtain a new target data set.

9. The method according to claim 8, further comprising: If the data identifier is empty, a new data identifier is generated; Alternatively, if the data identifier is not empty, update the data identifier; If the number of characters in the first target file is greater than or equal to a preset character threshold, the file status of the first target file is determined to be saturated; or if the number of characters in the first target file is less than the preset character threshold, the file status of the first target file is determined to be unsaturated.

10. The method according to claim 1, wherein, The user question and answer content also includes a user identifier; the step of querying the target data set stored in the distributed database to see if there is first target data matching the question content includes: Based on the user identifier, query the target data set corresponding to the user identifier from the distributed database; Based on the content of the question asked by the user, at least one candidate data matching the question content is queried from the target data set; Semantic features are extracted from the content of the user's question to obtain a first feature vector; semantic features are extracted from the at least one candidate data to obtain at least one second feature vector; Calculate the similarity between the first feature vector and each of the second feature vectors; Based on the similarity calculation results, determine whether there is a first target data that matches the content of the question from the at least one candidate data.

11. A data query method, comprising: In response to a query request for the content of a user’s historical questions, a data set corresponding to the target user identifier carried in the query request is retrieved from a distributed database, wherein the data set is obtained by the data processing method according to any one of claims 1 to 10. Based on the content of the historical questions, query the target data associated with the content of the historical questions from the dataset; The target file is retrieved from the cloud disk system based on the file identifier of the target file associated with the target data, wherein the text content of the target file is used to represent the answer to the question related to the historical question content.

12. A data processing apparatus, comprising: The question-and-answer content collection module is used to collect user question-and-answer content from different intelligent question-and-answer applications based on a preset collection interface. The user question-and-answer content includes the question asked by the user and the answer to the question. The first target data query module is used to respond to the query request for the content of the question asked by the user, and query whether there is first target data that matches the content of the question from the target data set stored in the distributed database. The first file generation module is configured to, in response to the existence of first target data in the target data set that matches the question content, generate second target data and a first target file based on the first target data, the question content asked by the user, and the question answer, wherein the text content of the first target file is used to represent the question answer related to the second target data; The data acquisition module is used to update the target data set based on the second target data and the first target file to obtain a new target data set.

13. A data query device, comprising: The query request module is used to respond to a query request for the content of a user’s historical questions, and to query a data set corresponding to the target user identifier from a distributed database based on the target user identifier carried in the query request, wherein the data set is obtained by the data processing device according to claim 12. The target data query module is used to query target data associated with the historical question content from the data set based on the historical question content. The target file acquisition module is used to acquire the target file from the cloud disk system based on the file identifier of the target file associated with the target data, wherein the text content of the target file is used to represent the answer to the question related to the historical question content.

14. An electronic device, comprising: One or more processors; Memory, used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 11.

15. A computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the method of any one of claims 1 to 11.

16. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 11.

Images