Data processing method and apparatus, storage medium, and computer program product

Abstract

Embodiments of the present invention relate to the technical field of computers, and provide a data processing method and apparatus, a storage medium, and a computer program product. The data processing method comprises: determining a target text, and processing the target text to obtain a database query text and a network search text which correspond to the target text; performing database search on the database query text to obtain a database search result; performing network search on the network search text to obtain a network search result; and on the basis of the target text, the database query text, the database search result, and the network search result, using a data processing model to obtain a data processing result corresponding to the target text.

Description

[0001] Data processing method, device, storage medium and computer program product

[0002]

[0001] Embodiments of the present disclosure relate to the technical field of computer technology, and in particular to three data processing methods. One or more embodiments of the present disclosure also relate to a data processing device, a computing device, a computer-readable storage medium, and a computer program product. BACKGROUND

[0003]

[0002] In live broadcast, customer service, media and other scenarios, a user can prepare more content materials related to work content or collect more materials related to work content, for example, a match commentator can prepare information about a game, information about a sports event, information about members of the event, and related topics before the game, and for another example, a customer service personnel can collect performance parameters, product comparison data, and selling points of a product related to the service before work.

[0004]

[0003] However, in actual applications, because there are many current information sources, such as books, web searches, historical databases, articles, and the like, manual searching requires more manpower and time, and the obtained data content is of low quality, and the user experience is poor.

[0005]

[0004] Therefore, there is an urgent need for a technical solution to automatically obtain high-quality content to solve the above technical problems. SUMMARY

[0006]

[0005] In view of this, embodiments of the present disclosure provide three data processing methods. One or more embodiments of the present disclosure also relate to a question and answer method, two data processing devices, a computing device, a computer-readable storage medium, and a computer program product, to solve the technical defects in the related art that manual searching requires more manpower, time, and the obtained data content is of low quality due to many information sources.

[0007]

[0006] According to a first aspect of embodiments of the present disclosure, a data processing method is provided, including: determining a target text, and processing the target text to obtain a database query text and a network search text corresponding to the target text; performing database searching on the database query text to obtain a database search result; performing network searching on the network search text to obtain a network search result; and obtaining a data processing result corresponding to the target text by using a data processing model according to the target text, the database query text, the database search result, and the network search result, wherein the data processing model is a machine learning model.

[0008]

[0007] According to a second aspect of the embodiments of the present disclosure, another data processing method is provided, including: determining a target event material generation text, and processing the target event material generation text to obtain a target event material database query text and a target event material network search text corresponding to the target event material generation text, wherein the target event material generation text is used to generate a target event material content material; performing database search on the target event material database query text to obtain a target event material database search result; performing network search on the target event material network search text to obtain a target event material network search result; and according to the target event material generation text, the target event material database query text, the target event material database search result and the target event material network search result, using a data processing model to obtain a target event material data processing result corresponding to the target event material generation text, wherein the data processing model is a machine learning model, and the target event material data processing result contains the target event material content material.

[0009]

[0008] According to a third aspect of the embodiments of the present disclosure, a question and answer method is provided, including: determining a target question, and processing the target question to obtain a database query question and a network search question corresponding to the target question; performing database search on the database query question to obtain a database search result; performing network search on the network search question to obtain a network search result; and according to the target question, the database query question, the database search result and the network search result, using a data processing model to obtain a target answer corresponding to the target question, wherein the data processing model is a machine learning model.

[0010]

[0009] According to a fourth aspect of the embodiments of the present disclosure, another data processing method is provided, which is applied to a data processing platform, including: determining a target text, and processing the target text to obtain a database query text and a network search text corresponding to the target text; performing database search on the database query text to obtain a database search result; performing network search on the network search text to obtain a network search result; and according to the target text, the database query text, the database search result and the network search result, using a data processing model to obtain a data processing result corresponding to the target text, wherein the data processing model is a machine learning model.

[0011]

[0010] According to a fifth aspect of the embodiments of the present disclosure, a data processing apparatus is provided, comprising: a target text processing module configured to determine a target text, and process the target text to obtain a database query text and a network search text corresponding to the target text; a database search module configured to perform database search on the database query text to obtain a database search result; a network search module configured to perform network search on the network search text to obtain a network search result; and a data processing result obtaining module configured to obtain a data processing result corresponding to the target text by using a data processing model according to the target text, the database query text, the database search result, and the network search result, wherein the data processing model is a machine learning model.

[0012]

[0011] According to a sixth aspect of the embodiments of the present disclosure, another data processing apparatus is provided, comprising: a target event material generation text processing module configured to determine a target event material generation text, and process the target event material generation text to obtain a target event material database query text and a target event material network search text corresponding to the target event material generation text, wherein the target event material generation text is used to generate a target event material content material; a target event material database search module configured to perform database search on the target event material database query text to obtain a target event material database search result; a target event material network search module configured to perform network search on the target event material network search text to obtain a target event material network search result; and a target event material data processing result obtaining module configured to obtain a target event material data processing result corresponding to the target event material generation text by using a data processing model according to the target event material generation text, the target event material database query text, the target event material database search result, and the target event material network search result, wherein the data processing model is a machine learning model, and the target event material data processing result contains the target event material content material.

[0013]

[0012] According to a seventh aspect of the embodiments of the present disclosure, a computing device is provided, comprising: a memory and a processor; wherein the memory is configured to store computer programs / instructions, and the processor is configured to execute the computer programs / instructions, and the computer programs / instructions, when executed by the processor, implement the steps of the above data processing method.

[0014]

[0013] According to an eighth aspect of an embodiment of the present disclosure, a computer readable storage medium is provided, which stores computer programs / instructions, which, when executed by a processor, implement the steps of the above data processing method.

[0015]

[0014] According to a ninth aspect of an embodiment of the present disclosure, a computer program product is provided, which includes computer programs / instructions, which, when executed by a processor, implement the steps of the above data processing method.

[0016]

[0015] One embodiment of the present disclosure provides a data processing method, which automatically provides a data processing result corresponding to a target text to a user through data processing, so that the user does not need to perform manual retrieval, thereby reducing labor cost and time cost. The data processing method realizes multi-data source data search through database search and network search, obtains a database search result and a network search result, so that the data processing model can refer to data processing results of database query text, the database search result and the network search result in the process of answering the target text, thereby improving the comprehensiveness and accuracy of the data processing result obtained by the data processing model, i.e., improving the quality of the content generated for the target text, thereby improving the user experience.

[0017]

[0016] FIG. 1 is a specific application scenario diagram of a data processing method according to one embodiment of the present disclosure;

[0018]

[0017] FIG. 2 is a flowchart of a data processing method according to one embodiment of the present disclosure;

[0019]

[0018] FIG. 3 is a flowchart of another data processing method according to one embodiment of the present disclosure;

[0020]

[0019] FIG. 4 is a first interaction interface diagram of a data processing method according to one embodiment of the present disclosure;

[0021]

[0020] FIG. 5 is a second interaction interface diagram of a data processing method according to one embodiment of the present disclosure;

[0022]

[0021] FIG. 6 is a processing flowchart of a data processing method according to one embodiment of the present disclosure;

[0023]

[0022] FIG. 7 is a comparison diagram of processing flows of a data processing method according to one embodiment of the present disclosure;

[0024]

[0023] FIG. 8 is a third interaction interface diagram of a data processing method according to one embodiment of the present disclosure;

[0025]

[0024] FIG. 9 is a fourth interaction interface diagram of a data processing method according to an embodiment of the present disclosure;

[0026]

[0025] FIG. 10 is a structural diagram of a data processing apparatus according to an embodiment of the present disclosure;

[0027]

[0026] FIG. 11 is a structural diagram of another data processing apparatus according to an embodiment of the present disclosure;

[0028]

[0027] FIG. 12 is a structural block diagram of a computing device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0029]

[0028] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure. However, the present disclosure can be practiced without the specific details, which are not necessary for the understanding of the present disclosure, and it is understood that the present disclosure is not limited to the embodiments here described and might also be used in conjunction with other apparatuses. Those skilled in the art, having the benefit of the present description, will appreciate the various ways in which the present disclosure might be practiced and implemented.

[0030]

[0029] The terminology used in the one or more embodiments of the present disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of the present disclosure. As used in the one or more embodiments of the present disclosure and the accompanying claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in the one or more embodiments of the present disclosure, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0031]

[0030] It is to be understood that the terms first, second, etc. can be employed in this disclosure to describe various information. Such information should not be limited by these terms. These terms are only used to distinguish one category of information from another. For example, without departing from the scope of the one or more embodiments of the present disclosure, first can be termed second, and similarly, second can be termed first. Depending on the context, the word "if' can be interpreted to mean "when" or "upon" or "in response to determining." Depending on the context, the word "if' can be interpreted to mean "when" or "upon" or "in response to determining."

[0032]

[0031] In addition, it should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in one or more embodiments of the present disclosure are information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of the country and region, and provide corresponding operation portal for user to choose authorization or refusal.

[0033]

[0032] In one or more embodiments of the present disclosure, a large model refers to a deep learning model with a large number of model parameters, usually containing hundreds of millions, tens of billions, hundreds of billions, tens of billions or even more than one hundred billion model parameters. The large model can also be called a cornerstone model / foundation model

[0034]

[0035]

[0033] In actual application, the large model only needs a small amount of samples to fine-tune the pre-trained model and can be applied to different tasks. The large model can be widely applied to natural language processing (NLP, Natural Language Processing) and computer vision fields, and can be applied to computer vision field tasks such as visual data processing (VQA, Visual Question Answering), image description (IC, Image Caption), image generation, and natural language processing field tasks such as text-based sentiment classification, text summary generation, and machine translation. The main application scenarios of the large model include digital assistants, intelligent robots, search, online education, office software, e-commerce, intelligent design, etc.

[0036]

[0034] First, the nomenclature involved in one or more embodiments of the present disclosure is explained.

[0037] ​

[0035] Structured Query Language (SQL): A computer language used to manage and manipulate databases. Structured Query Language is designed to access, query, update and manage database systems.

[0038]

[0036] In this disclosure, two data processing methods are provided. One or more embodiments of this disclosure also involve two data processing devices, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail in the following embodiments.

[0039]

[0037] Considering the large number of model parameters in large models and the limited computing resources of mobile terminals, the data processing method provided in this application embodiment can be applied to the application scenario shown in Figure 1, but is not limited thereto.

[0040]

[0038] Referring to Figure 1, Figure 1 shows a specific application scenario diagram of a data processing method provided by an embodiment of the present disclosure.

[0041]

[0039] In the application scenario shown in Figure 1, the large model is deployed on server 104, which includes, but is not limited to, physical servers, cloud servers, etc. Server 104 can connect to one or more clients 102 via local area network (LAN), wide area network (WAN), Internet, or other types of data networks. Clients 102 may include, but are not limited to, smartphones, tablets, laptops, PDAs, personal computers, smart home devices, in-vehicle devices, etc. Clients 102 can interact with users through a graphical user interface to call the large model, thereby realizing the data processing method provided in this embodiment of the present disclosure.

[0042]

[0040] Specifically, the data processing method includes: determining target text, and processing the target text to obtain database query text and web search text corresponding to the target text; performing a database search on the database query text to obtain database search results; performing a web search on the web search text to obtain web search results; and using a data processing model based on the target text, the database query text, the database search results, and the web search results to obtain data processing results corresponding to the target text, wherein the data processing model is a machine learning model.

[0043]

[0041] In the embodiments of the present disclosure, the system composed of the client 102 and the server 104 can perform the following steps: the client 102 generates target text in response to the user's starting operation on the user interaction interface of the client, and sends the target text to the server 104; the server 104 processes the target text into database query text and network search text corresponding to the target text, performs database search on the database query text to obtain database search results, and performs network search on the network search text to obtain network search results, so as to obtain data processing results corresponding to the target text by using a data processing model according to the target text, the database query text, the database search results and the network search results; further, the server 104 sends the data processing results to the client 102, and the client 102 displays the data processing results to the user through the user interaction interface.

[0044]

[0042] It should be noted that, in the case that the running resources of the client can meet the deployment and running conditions of the large model, the embodiments of the present disclosure can be performed in the client, and in addition, the large model can also be deployed in a third-party server, which includes but is not limited to a third-party physical server, a third-party cloud server, etc.

[0045]

[0043] One embodiment of the present disclosure provides a data processing method, in which the client sends target text to the server through data processing, the server automatically generates data processing results corresponding to the target text, and returns the data processing results to the client to display to the user, so that the user does not need to perform manual retrieval, thereby reducing the labor cost and time cost, and in addition, the server realizes data search of multiple data sources through database search and network search to obtain database search results and network search results, so that the data processing model can refer to the data processing results of the database query text, the database search results and the network search results in the process of answering the target text, thereby improving the comprehensiveness and accuracy of the data processing results obtained by the data processing model, i.e., improving the quality of the content generated for the target text, thereby improving the user experience.

[0046]

[0044] Referring to FIG. 2, FIG. 2 shows a flowchart of a data processing method according to one embodiment of the present disclosure, which specifically includes the following steps.

[0047]

[0045] Step 202: determining target text and processing the target text to obtain database query text and network search text corresponding to the target text.

[0048]

[0046] Wherein, the target text can be understood as a text to be processed, including but not limited to questions, dialogues, etc., for example, the target text can be understood as an event commentary type target text, which is used to provide materials to users who have event commentary needs, i.e., target event material generation text, and the event commentary type includes but is not limited to a sports event commentary type in a sports scene, a match commentary type in a sports scene, a game content commentary type in a game scene, and an item commentary type in a live broadcast scene, or for example, the target text can also be understood as a dialogue in a dialogue scene, including but not limited to question answering in a customer service dialogue scene, replies to user dialogues in a live broadcast scene, chat answers of a chat robot, etc.

[0049]

[0047] In the case where the target text is understood as a target event material generation text, the target text can be used to generate target event materials, i.e., the data processing result corresponding to the subsequently generated target text contains target event materials, wherein the target event materials can be understood as materials related to the content that the user wants to explain, for example, the target text can be understood as "the competition result situation of the project member in the xx type match", "the team information of the xx type match in xxxx year", etc., and accordingly, the data processing result can contain historical competition result situation of the project member, competition result records, etc.

[0050]

[0048] In the case where the target text is understood as a dialogue question in a data processing scene, the target text can be used to generate data related to the target text, and the data processing result corresponding to the subsequently generated target text can contain data related to the target text, wherein the target text can be understood as a question raised by a user in a customer service dialogue live broadcast process, for example, the target text can be understood as "the performance of xx item", "which is better, xx item or yy item", etc., and accordingly, the data processing result can contain item models, item parameters, comparison tables between items, etc.

[0051]

[0049] The database query text can be understood as a query text used to query data related to the target text from a database, including but not limited to SQL query instructions, keyword matching query texts, vector matching query vectors, etc.

[0052]

[0050] The network search text can be understood as a question for retrieving a webpage related to the target text from the network, including but not limited to keywords, wildcards, etc. For example, the target text is understood as "XX type competition team information in XXXX year", the network search text can be understood as "xx type competition, xxxx year, competition result", "xx type competition, xxxx year, competition result information", "xx type competition, xxxx year, team information", etc.

[0053]

[0054]

[0051] In practical applications, the target text is processed to obtain the database query text and the network search text corresponding to the target text, which can be achieved by a preset text conversion rule. For example, a preset mapping table is used to map the target text to obtain the database query text and the network search text corresponding to the target text, or a machine learning model is used to process the target text to obtain the database query text and the network search text corresponding to the target text.

[0055]

[0052] Specifically, the machine learning model is used to process the target text to obtain the database query text and the network search text corresponding to the target text, and the specific implementation manner is as follows.

[0056]

[0053] The target text is processed to obtain the database query text and the network search text corresponding to the target text, including: determining a preset prompt text, wherein the preset prompt text is used to instruct a text conversion model to convert the target text into a database query text and a network search text; inputting the target text and the preset prompt text into the text conversion model to obtain the database query text and the network search text corresponding to the target text, wherein the text conversion model is a machine learning model.

[0057] ​

[0054] The preset prompt text can be understood as a text preset for guiding the output direction of the text conversion model. Specifically, the preset prompt text is used to instruct the text conversion model to convert the target text input into the text conversion model into a database query text and a network search text. For example, the preset prompt text can be understood as "convert the following input text into two output forms: 1. database query text; 2. network search text", "convert the following input content into two texts, one is a database query text, and the other is a network search text", and the like.

[0058]

[0056] Based on this, the target text and the preset prompt text are input into the text conversion model, and the text conversion model converts the target text according to the preset prompt text, and outputs the database query text corresponding to the target text and the network search text corresponding to the target text.

[0059]

[0057] The data processing method provided by one or more embodiments of the present disclosure improves the processing efficiency of data processing by automatically converting the target text using the text conversion model, reduces the labor cost caused by manual conversion, and increases the preset prompt text in the process of converting the target text using the text conversion model, thereby improving the accuracy of the text conversion model in converting the target text, and improving the accuracy of the obtained database query text and network search text corresponding to the target text.

[0060]

[0058] In practical applications, the text conversion model can be deployed locally on the server or in a third-party server. In the case of deploying the text conversion model locally on the server, the target text and the preset prompt text can be directly input into the locally deployed text conversion model to obtain the output result of the text conversion model, i.e., the database query text and the network search text corresponding to the target text, to improve the security and speed of converting the target text.

[0061]

[0059] In the case of deploying the text conversion model in a third-party server, the target text and the preset prompt text are input into the text conversion model to obtain the specific implementation manner of the database query text and the network search text corresponding to the target text as follows.

[0062]

[0060] The target text and the preset prompt text are input into the text conversion model to obtain the database query text and the network search text corresponding to the target text.

[0063]

[0061] The text conversion model calling interface can be understood as an interface for calling the text conversion model, including but not limited to a physical interface, an application program calling interface, a data interface, a network interface, and the like.

[0064]

[0062] Therefore, the text conversion model can be called through the text conversion model calling interface, and then the target text and the preset prompt text are input into the called text conversion model to obtain the database query text and the network search text corresponding to the question output by the text conversion model.

[0065]

[0063] The data processing method provided by one or more embodiments of the present disclosure can realize conversion of the target text by calling the text conversion model deployed by the third-party server, thereby reducing the local resource occupancy rate and the model training cost, and ensuring the running stability of the server.

[0066]

[0064] Step 204: performing database search on the database query text to obtain a database search result.

[0067]

[0065] The database search can be understood as a search method for searching data from a database, and the database search result can be understood as data associated with the database query text in the database searched.

[0068]

[0066] Specifically, performing database search on the database query text to obtain a database search result can be understood as searching data associated with the database query text from the database, that is, obtaining the database search result, for example, searching data completely matching the database query text from the database, searching data having a mapping relationship with the database query text from the database, and the like.

[0069]

[0067] In actual application, in order to improve the accuracy of database search, database search can be performed on a database related to the target application scenario corresponding to the target text, and the specific implementation manner is as follows.

[0070]

[0068] The database searching on the database query text, obtaining a database search result, comprises: determining a target database associated with the target text, wherein the target database is constructed through a target application scenario corresponding to the target text; and performing data searching on the target database according to the database query text, and obtaining the database search result.

[0071]

[0069] The target application scenario corresponding to the target text can be understood as a specific application scenario involved in the target text. In the above example, the target application scenario corresponding to the target text "xx type match team information in xxxx year" can be understood as a sports scenario, and the target application scenario corresponding to the target text "xx article performance" can be understood as a customer service dialogue scenario, and so on.

[0072]

[0070] Based on this, different databases can be constructed for different application scenarios, such as a sports database for a sports scenario, an article information database for a customer service dialogue scenario, and so on.

[0073]

[0071] In order to make the database searching on the target text obtain more accurate database search results, a target database associated with the target text and constructed through a target application scenario corresponding to the target text can be determined first, so that the data searching on the database query text is performed in the target database, and the database search result searched from the target database is obtained.

[0074]

[0072] The data processing method provided by one or more embodiments of the present disclosure can make the database search result obtained by performing data searching on the database query text in the target database constructed through the target application scenario corresponding to the target text more relevant to the target text, and improve the accuracy of the data processing result obtained by subsequently referring to the database search result.

[0073] In actual application, the database search result corresponding to the database query text can be queried according to the database query text. In order to enable the subsequent data processing model to answer more accurate data processing results, the context of the database search result can also be queried in the database. The specific implementation manner is described as follows.

[0075]

[0074] The method further comprises: searching data in the target database according to the database query text to obtain a first database search result; in a case where the first database search result meets a preset search condition, updating the database query text to obtain an updated database query text; searching data in the target database according to the updated database query text to obtain a second database search result; and obtaining the database search result according to the first database search result and the second database search result.

[0076]

[0075] The first database search result can be understood as a search result obtained by directly searching data in the target database according to the database query text. For example, the database query text is "xx type competition team information in xxxx year", and the corresponding SQL query instruction is "xx team" "xx team participates in xx type competition n times in xxxx year", etc.

[0077]

[0076] In a specific implementation, the first search result can also be null, that is, no content is searched in the target database according to the database query text. In this case, the null value can be directly used as the database search result.

[0078]

[0077] The preset search condition can be understood as the first search result being not null, that is, in a case where the first search result is not null, the step of updating the database query text to obtain the updated database query text is performed.

[0079]

[0078] Specifically, the updated database query text can be understood as a database query text used to search data associated with the first database search result in the target database, and the data associated with the first database search result is the second database search result.

[0080]

[0079] In a specific implementation, the step of updating the database query text to obtain the updated database query text can be understood as rewriting the database query text according to the first database search result to obtain a database query text used to search data associated with the first database search result, that is, the updated database query text.

[0081]

[0080] Based on this, the updated database query text can be searched again in the target database to obtain a second database search result. For example, the second database search result can be understood as the context of the first database search result, the database table to which the first database search result belongs, and the like.

[0082]

[0081] Further, after obtaining the first database search result and the second database search result, the first database search result and the second database search result can be combined as the database search result.

[0083]

[0082] One or more embodiments of the present disclosure provide a data processing method. After database searching according to a database query text, the database query text is updated to obtain an updated database query text for searching related data of the first database search result, and the database is searched again. The comprehensiveness of the database search result is improved, and the quality of the data processing result obtained by subsequently referring to the database search result is further improved.

[0084]

[0083] Step 206: performing a network search on the network search text to obtain a network search result

[0085]

[0084] The network search can be understood as a search operation in the Internet through a web directory, a web classification index, a network search engine, or the like.

[0086]

[0085] The network search result can be understood as web page data obtained by performing a network search on the network search text. Specifically, the network search result is compliant and desensitized web page data.

[0087]

[0086] In practical applications, user information (including but not limited to user equipment information, user personal information, and the like) and data (including but not limited to data for analysis, stored data, displayed data, and the like) involved in the network search data are information and data authorized by the user or authorized by all parties, and the collection, use, and processing of the related data need to comply with relevant laws, regulations, and standards of relevant countries and regions, and provide corresponding operation portals for the user to select authorization or rejection.

[0088]

[0087] In the case where the network search text is understood as “xx type match, xxxx year, match result”, the network search result can be understood as an article link, a match notification web page, or the like containing one or more of “xx type match”, “xxxx year”, and “match result”.

[0089]

[0088] In practical applications, the specific implementation of network search using a network search engine is as follows.

[0090]

[0089] The network search on the network search text to obtain a network search result includes: according to the network search text, using a target network search engine to obtain network search data output by the target network search engine and / or data attribute information of the network search data, wherein the data attribute information includes data sources of the network search data and / or a website address of a webpage to which the network search data belongs; and according to the network search data and / or the data attribute information of the network search data, obtaining the network search result.

[0091]

[0090] The target network search engine can be understood as a network search engine to be searched, and can be a pre-set network search engine, or a plurality of network search engines can be pre-set, and one of the plurality of network search engines can be selected as the target network search engine, and the like. The determination manner of the network search engine is not limited in the embodiments of the present disclosure.

[0092]

[0091] The network search data can be understood as webpage data output by the target network search engine searching the network search text, and the data attribute information of the network search data includes but is not limited to a webpage address of a webpage where the network search data is located (i.e., a website address of a webpage to which the network search data belongs), a data source of the network search data, and the like.

[0093]

[0092] For example, the network search data can be understood as a tweet written by an xx user of an xx website, the website address of the webpage to which the network search data belongs can be understood as a webpage address that can link to the tweet written by the xx user of the xx website, and the data source of the network search data can be understood as an xx tweet title published by the xx user of the xx website at an xx time.

[0094]

[0093] For another example, the network search data can be understood as an xx interpretation published by an xx person at an xx time on an xx website, the website address of the webpage to which the network search data belongs can be understood as a webpage address of the interpretation, and the data source of the network search data can be understood as an xx interpretation article title of the xx website, the xx person, the xx time, and the like.

[0095]

[0094] One or more embodiments of the present disclosure provide a data processing method, which improves the comprehensiveness of the network search result by adding the data source of the network search data and / or the URL of the webpage to which the network search data belongs in the network search result, and the data source of the network search data and / or the URL of the webpage to which the network search data belongs can be added in the data processing result generated by the data processing model subsequently, so that the user can directly view the original data of the network search data and / or the webpage to which the network search data belongs through the data processing result, and the user experience is improved.

[0096]

[0095] Step 208: obtaining the data processing result corresponding to the target text by using a data processing model according to the target text, the database query text, the database search result, and the network search result.

[0097]

[0096] The data processing model is a machine learning model.

[0098]

[0097] The data processing model can be understood as a machine learning model for data processing of the target text. The data processing model can be obtained based on a model trained by a language model, a large language model, a multi-modal pre-training model, etc., and has the ability to analyze the target text and give the data processing result corresponding to the target text. For example, the data processing model can be understood as a text processing model (for text processing), a data processing model (for question answering), a dialogue model (for dialogue processing), etc.

[0099]

[0098] In actual application, after the multi-source data search of the target text is completed, i.e., the database search result and the network search result are obtained, the data processing result corresponding to the target text can be obtained by using a data processing model according to the target text, the database query text, the database search result, and the network search result.

[0100]

[0099] Specifically, the data processing model can be deployed locally on the server or on a third-party server. In the case of deploying the data processing model locally on the server, the target text, the database query text, the database search result, and the network search result can be input into the data processing model deployed locally to obtain the data processing result corresponding to the target text output by the data processing model, so as to improve the security of answering the target text and improve the answering speed of the target text.

[0101]

[0100] In the case of deploying the data processing model on the third-party server, the data processing model performs the specific implementation of the answer to the target text as follows.

[0102]

[0101] According to the target text, the database query text, the database search result, the network search result, the data processing model is used to obtain the data processing result corresponding to the target text, including: calling the data processing model according to the data processing model calling interface; inputting the target text, the database query text, the database search result, and the network search result into the data processing model to obtain the data processing result corresponding to the target text.

[0103]

[0102] Specifically, the specific implementation of the embodiment of the present disclosure can refer to the above-mentioned implementation of the embodiment of the specification. In the case of deploying the text conversion model on the third-party server, the target text and the preset prompt text are input into the text conversion model to obtain the specific implementation of the database query text and the network search text corresponding to the target text. The difference lies in the different models called and the different data input into the model, and accordingly, different output results of the model output are obtained, which will not be described here.

[0104]

[0103] One or more embodiments of the present disclosure provide a data processing method, which calls a question model deployed by a third-party server, and inputs a target text, a database query text, a database search result, and a network search result to realize an answer to the target text and obtain a data processing result corresponding to the target text, thereby reducing local resource occupancy and model training cost and ensuring server operation stability.

[0105]

[0104] In practical applications, in order to improve the quality of the data processing result output by the data processing model, a data processing model can be trained in advance using high-quality samples. For example, in the sports commentary scenario, in addition to the accuracy of the target text returned by the data processing model, the comprehensiveness of the returned information, the integrity of the reasoning process, the ability to provide additional effective information, and the requirement that the data processing model not output illegal information can also be considered.

[0105] Based on this, high-quality data processing samples and simulated data processing data samples with noise (such as illegal information, invalid data, etc.) can be constructed in advance. According to the high-quality data processing samples and the simulated data processing data samples, a data processing model is trained to improve the comprehensiveness and integrity of the data processing result output by the data processing model, and to reduce the probability of invalid data processing results and illegal information output by the data processing model.

[0106]

[0106] In practical applications, the data processing model can also combine the data in the pre-constructed sample library to answer the target text, to further improve the quality of the data processing result. The specific implementation is as follows.

[0107]

[0107] After determining the target text, the method further includes: performing a data processing sample database search on the target text to obtain a sample library search result, wherein the data processing sample database is constructed by data processing sample pairs formed by a text sample and a data processing result sample corresponding to the text sample; and obtaining the data processing result corresponding to the target text by using the data processing model according to the target text, the database query text, the database search result, the network search result, and the sample library search result, including: obtaining the data processing result corresponding to the target text by using the data processing model according to the target text, the database query text, the database search result, the network search result, and the sample library search result.

[0108]

[0108] The text sample can be understood as a pre-constructed question associated with the above-mentioned target application scenario, and the data processing result sample corresponding to the text sample can be understood as a high-quality data processing result sample obtained by replying to the text sample according to the requirements of answering the question in the target application scenario, for the reference of the data processing model.

[0109]

[0109] Based on this, a data processing example pair can be formed according to the text example and the data processing result example corresponding to the text example, and a data processing example database is constructed according to the data processing example pair. In the data processing example database, the target text is searched to obtain a sample library search result. The sample library search result contains a text example matched with the target text and a data processing result example corresponding to the text example matched with the target text.

[0110]

[0110] Further, the data processing model can be used to obtain the data processing result corresponding to the target text according to the target text, the database query text, the database search result, the network search result, and the sample library search result. The data processing model increases the sample library search result as a reference in the case of the above-mentioned database query text, database search result, and network search result.

[0111]

[0111] One or more embodiments of the present disclosure provide a data processing method. The data processing model increases the sample library search result as a reference, so that the data processing result output by the data processing model is more biased towards the data processing result example in the sample library search result. Since the data processing result example is a pre-constructed data processing result example of higher quality, the quality of the data processing result output by the data processing model is greatly improved.

[0112]

[0112] In practical applications, in order to improve the search efficiency of the data processing example database, the target text can be subjected to semantic search. The specific implementation manner is described as follows.

[0113]

[0113] The data processing example database search on the target text to obtain the sample library search result includes: determining the text example and the data processing result example of each data processing example pair in the data processing example database; calculating the semantic similarity between the target text and each text example, and determining a target text example according to the semantic similarity between the target text and each text example; and obtaining the sample library search result according to the data processing result example corresponding to the target text example.

[0114]

[0114] The semantic similarity can be understood as the similarity between the semantics of the target text and the semantics of each text example.

[0115]

[0115] In practice, first, the text samples in the data processing sample database to be searched are determined, as well as the data processing result samples corresponding to each text sample, and the semantic similarity between the target text and each text sample is calculated, so as to search for a text sample with more similar semantics to the target text, that is, according to the semantic similarity between the target text and each text sample, the target text sample is determined, and the target text sample is a text sample with more similar semantics to the target text.

[0116]

[0116] Further, according to the data processing result sample corresponding to the target text sample, the sample library search result can be obtained, so as to improve the quality of the data processing result, for example, the data processing result sample corresponding to the target text sample is taken as the sample library search result, or for example, the data processing result sample corresponding to the target text sample is segmented, and the segmentation result is taken as the sample library search result, and the like.

[0117]

[0117] One or more embodiments of the present disclosure provide a data processing method, according to the semantic similarity between the target text and each text sample, the target text sample is determined, so that the target text sample has more similar semantics to the target text, and according to the data processing result sample corresponding to the target text sample, the sample library search result is determined, which improves the accuracy of the sample library search result.

[0118]

[0118] In one or more embodiments of the present disclosure, after determining the text sample and the data processing result sample of each data processing sample pair in the data processing sample database, the target text sample matching the target text vector or the keyword matching target text can be determined from each data processing sample in the data processing sample database through vector retrieval and keyword retrieval, and then the sample library search result can be obtained according to the data processing result sample corresponding to the target text sample, so as to improve the quality of the data processing result, which will not be described in detail in the embodiments of the present disclosure.

[0119] In practical applications, in order to improve the accuracy of the semantic similarity calculation, a more accurate semantic similarity can be calculated by calculating the vector similarity between the target text and the text sample, and the specific implementation manner is as follows.

[0119]

[0120] The calculating the semantic similarity of the target text and the text samples comprises: performing vectorization processing on the target text to obtain a target vector corresponding to the target text; determining a text sample vector of each text sample and calculating a vector distance between the target vector and the text sample vector of each text sample; and calculating the semantic similarity of the target text and each text sample according to the vector distance between the target vector and the text sample vector of each text sample.

[0120]

[0121] The vectorization processing can be understood as a processing manner of converting the target text into a vector, and can be implemented by a vector conversion table, a vectorization model (which can be understood as a machine learning model), or the like. For example, the target text can be vectorized according to a preset mapping relationship between a text and a vector to obtain a target vector corresponding to the target text, or the target text can be input into a vectorization model to obtain a target vector corresponding to the target text output by the vectorization model, and the like.

[0121]

[0122] The text sample vector of each text sample can be determined by the above vectorization processing manner or by a preset mapping relationship between a text sample and a text sample vector, and the disclosure embodiments do not limit this.

[0122]

[0123] The vector distance can be understood as a spatial distance between vectors, and includes but is not limited to a cosine distance, a Manhattan distance, and the like.

[0123]

[0124] In a specific implementation, the vector distance between the target vector corresponding to the target text and the text sample vector of each text sample can be calculated, and then the semantic similarity of the target text and each text sample can be calculated according to the vector distance between the target vector and the text sample vector of each text sample.

[0124]

[0125] For example, the vector distance between the target vector and the text sample vector of each text sample is determined as the semantic similarity of the target text and each text sample.

[0125]

[0126] For another example, the semantic similarity of the target text and each text sample can be calculated according to the vector distance between the target vector and the text sample vector of each text sample and a preset coefficient factor (i.e., a coefficient factor between the vector distance and the semantic similarity, for example, 0.5, i.e., in a case where the vector distance is 200, the semantic similarity is 200*0.5, i.e., 100).

[0126]

[0127] The data processing method provided by one or more embodiments of the present disclosure can obtain semantic similarity between a target text and each text sample by calculating vector distance between a target vector and a text sample vector of each text sample, so that the semantic similarity between the target text and each text sample is more accurate, and the obtained target text sample is more matched with the target text.

[0127]

[0128] In actual application, the target text sample can be obtained by screening according to the semantic similarity between the target text and each text sample, so as to improve the reference value of the target text sample. The specific implementation manner is as follows.

[0128]

[0129] The target text sample is determined according to the semantic similarity between the target text and each text sample, including: the each text sample is sorted according to the semantic similarity between the target text and the each text sample, and the target text sample is determined from the each text sample according to the sorting result and a preset sample number; or the target text sample is determined from the each text sample according to the association relationship between the semantic similarity between the target text and the each text sample and a preset similarity threshold.

[0129]

[0130] The preset sample number can be understood as a preset number of target text samples to be obtained by screening, for example, the preset sample number can be understood as 2, 3, etc. In actual application, the preset sample number can be adjusted or determined according to the answer quality of the data processing model, or can be determined according to the actual demand of the data processing model. The present disclosure does not limit the preset sample number and the data source of the preset sample number.

[0130]

[0131] The preset similarity threshold can be understood as a preset threshold of semantic similarity. For example, in the case that the range of semantic similarity is 0 to 100, the preset similarity threshold can be understood as 80, 90, etc. For another example, in the case that the range of semantic similarity is 0 to 500, the preset similarity threshold can be understood as 450, 480, etc. The preset similarity threshold is used to screen out the text sample with higher semantic similarity (i.e., more similar semantics) with the target text and determine it as the target text sample, so as to ensure that the obtained target text sample has more similar semantics with the target text.

[0131]

[0132] In a specific implementation, the target text and each text sample are ranked according to the semantic similarity, and the target text sample is determined from the text samples according to the ranking result and a preset sample number. In this case, the target text and each text sample are ranked according to the semantic similarity, and the text sample that meets the preset sample number is determined from the ranked text samples from high to low, and the target text sample is determined.

[0132]

[0133] According to the association between the semantic similarity of the target text and each text sample and the preset similarity threshold, the target text sample is determined from the text samples. In this case, in the case where the higher the semantic similarity, the more similar the semantics of the target text and the text sample, the text sample corresponding to the semantic similarity that is greater than or equal to (or greater than) the preset similarity threshold in the semantic similarity of the target text and each text sample is determined as the target text sample. Or in the case where the lower the semantic similarity, the more similar the semantics of the target text and the text sample, the text sample corresponding to the semantic similarity that is less than or equal to (or less than) the preset similarity threshold in the semantic similarity of the target text and each text sample is determined as the target text sample.

[0134] The data processing method provided by one or more embodiments of the present disclosure ensures the matching degree between the target text sample and the target text by screening and determining the text sample with more similar semantics to the target text in the text sample as the target text sample, improves the accuracy of the target text sample obtained subsequently, and further improves the accuracy of the search result of the sample library obtained according to the target text sample.

[0133]

[0135] In actual application, the data processing result corresponding to the target text is obtained by using a data processing model according to the target text, the database query text, the database search result, the network search result, and the sample library search result. The data processing model can be implemented locally on the server, or a data processing model in a third-party server can be called to implement the data processing model.

[0134]

[0136] The data processing result corresponding to the target text is obtained by calling a data processing model according to the target text, the database query text, the database search result, the network search result, and the sample library search result.

[0135]

[0137] Specifically, the specific implementation of the embodiments of the present disclosure can refer to the above-described embodiments, and will not be repeated here.

[0136]

[0138] The data processing method provided by one or more embodiments of the present disclosure can obtain the data processing result corresponding to the target text by calling the question model deployed by the third-party server and inputting the target text, the database query text, the database search result, the network search result, and the sample library search result to answer the target text, thereby reducing the local resource occupancy rate and the model training cost and ensuring the running stability of the server.

[0137]

[0139] In actual application, in order to improve the user experience, the data processing method provided by the embodiments of the present disclosure can interact with the user, and the specific implementation is as follows.

[0138]

[0140] The target text is determined by receiving the target text sent by the client, wherein the target text is generated by the user's trigger operation on the user interaction interface of the client; after obtaining the data processing result corresponding to the target text, the data processing result is returned to the client, so that the client displays the data processing result to the user through the user interaction interface.

[0139]

[0141] The user interaction interface can be understood as an interface for the user to interact with the client, and the user interaction interface includes interactive buttons, text input boxes, and the like.

[0140]

[0142] The client can be understood as a client corresponding to the server applying the above data processing method, that is, the user can input the target text through the client, the client can send the target text to the server, and then the server can obtain the data processing result corresponding to the target text by using the above data processing method, and the client can display the data processing result to the user.

[0141]

[0143] The data processing method provided by one or more embodiments of the present disclosure can improve the user interaction and the user experience.

[0142]

[0144] The data processing method provided by one or more embodiments of the present disclosure can automatically provide the data processing result corresponding to the target text to the user in a data processing manner, so that the user does not need to perform manual search, and the labor cost and the time cost are reduced. The data search of multiple data sources is implemented through the database search and the network search, and the database search result and the network search result are obtained. In the process of answering the target text by the data processing model, the data processing result of the database query text, the database search result and the network search result can be referred to, so that the comprehensiveness and the accuracy of the data processing result obtained by the data processing model are improved, the quality of the content generated for the target text is improved, and the user experience is improved.

[0143]

[0145] Referring to FIG. 3, FIG. 3 shows a flowchart of another data processing method according to one embodiment of the present disclosure, which specifically includes the following steps.

[0144]

[0146] Step 302: determining a target event material generation text, and processing the target event material generation text to obtain a target event material database query text and a target event material network search text corresponding to the target event material generation text.

[0145]

[0147] The target event material generation text is used to generate a target event material.

[0146]

[0148] Step 304: performing database search on the target event material database query text to obtain a target event material database search result.

[0147]

[0149] Step 306: performing network search on the target event material network search text to obtain a target event material network search result.

[0148]

[0150] Step 308: according to the target event material generation text, the target event material database query text, the target event material database search result and the target event material network search result, a data processing model is used to obtain a target event material data processing result corresponding to the target event material generation text.

[0149]

[0151] The data processing model is a machine learning model, and the target event material data processing result includes the target event material.

[0150]

[0152] In one or more embodiments of the present disclosure, the target event material generation text is determined by receiving the target event material generation text sent by the client, wherein the target event material generation text is generated by the user through a trigger operation of a user interaction interface of the client. After obtaining the target event material data processing result corresponding to the target event material generation text, the target event material data processing result is returned to the client, so that the client displays the target event material data processing result to the user through the user interaction interface.

[0151]

[0153] Specifically, the specific implementation of the embodiments of the present disclosure can refer to the above-mentioned description of the embodiments, which will not be repeated here.

[0152]

[0154] The data processing method provided by one or more embodiments of the present disclosure automatically provides the target event material data processing result corresponding to the target event material generation text to the user in need of the target event material, so that the user does not need to perform manual retrieval, thereby reducing the labor cost and time cost. Through database search and network search, multi-data source data search is realized to obtain target event material database search results and target event material network search results. In the process of answering the target event material generation text by the data processing model, the target event material data processing result can be obtained by referring to the target event material database query text, the target event material database search result and the target event material network search result, thereby improving the quality of the target event material data processing result obtained by the data processing model, i.e. improving the quality of the target event material content material, thereby improving the user experience of the user in need of the target event material.

[0153]

[0155] The data processing method provided by the present disclosure will be further described below by taking the application of the data processing method provided by the present disclosure in the sports event commentary scene for realizing question and answer as an example in combination with FIG. 4 to FIG. 9.

[0154]

[0156] Referring to FIG. 4, FIG. 4 shows a first interaction interface schematic diagram of a data processing method provided by one embodiment of the present disclosure.

[0155]

[0157] As shown in FIG. 4, FIG. 4 shows an interaction interface diagram of an interaction interface displayed to a user by a client, which includes a title area, an answer display area, and a question input area (for example, a control showing "Hello, I am your dialogue assistant, do you have any questions?"). The user can input a target question in the question input area, and view a target answer to the target question obtained by the server through the title area and the answer display area. Specifically, the title area is used to display a title index of the target answer, and the answer display area is used to display specific content under each title index.

[0156]

[0158] The client obtains a trigger operation of the user on the question input area, and displays a dialogue interface to the user. As shown in FIG. 5, FIG. 5 shows a second interaction interface diagram of the data processing method provided by an embodiment of the present disclosure. Compared with FIG. 4, FIG. 5 adds a dialogue display area to display the dialogue record with the user. For example, as shown in FIG. 5, the client first displays the dialogue "Hello, I am your dialogue assistant, do you have any questions?" to the user and waits for the user to input a target question.

[0157] The client obtains a trigger operation of the user on the question input area, and displays a dialogue interface to the user. As shown in FIG. 5, FIG. 5 shows a second interaction interface diagram of the data processing method provided by an embodiment of the present disclosure. Compared with FIG. 4, FIG. 5 adds a dialogue display area to display the dialogue record with the user. For example, as shown in FIG. 5, the client first displays the dialogue "Hello, I am your dialogue assistant, do you have any questions?" to the user and waits for the user to input a target question.

[0158]

[0159] Based on this, the client obtains a target question input by the user in the question input area, and sends the target question to the server. The server answers the target question, and displays a target answer corresponding to the target question to the user through the title area and the answer display area.

[0159]

[0160] In one or more embodiments of the present disclosure, the client can also be understood as a front end, and the server can also be understood as a back end, which are not limited in the embodiments of the present disclosure.

[0160]

[0161] As shown in FIG. 6, FIG. 6 shows a process flow diagram of the data processing method provided by an embodiment of the present disclosure, which specifically includes the following steps.

[0161]

[0162] Step 602: Obtain a user question.

[0162]

[0163] Specifically, the client can obtain a target question input by the user through the question input area, and send the target question to the server.

[0163]

[0164] Specifically, the client can obtain a target question input by the user through the question input area, and send the target question to the server.

[0164]

[0165] Step 604: Question conversion.

[0165]

[0166] The problem conversion can be understood as the server converting the target problem into a SQL query instruction and a network search problem.

[0166]

[0167] Specifically, after receiving the target problem, the server inputs the target problem into the table question and answer module. The table question and answer module calls a text conversion model (for example, a large language model), inputs the target problem and a preset prompt text for guiding the language model to convert the target problem into a SQL query instruction and a network search problem into the text conversion model, and obtains the target SQL query instruction (i.e., the above database query text) and the target network search problem (i.e., the above network search text) corresponding to the target problem output by the text conversion model.

[0167]

[0168] Step 606: SQL execution.

[0168]

[0169] The SQL execution can be understood as executing the above SQL query instruction.

[0169]

[0170] Specifically, after the server obtains the target SQL query instruction corresponding to the target problem through the table question and answer module, the target SQL query instruction is executed in the database to obtain a first database search result.

[0171] Further, the target SQL query instruction can be updated according to the first database search result to obtain an updated SQL query instruction (i.e., the above updated database query text) for querying the context of the first database search result. Then, the updated SQL query instruction is executed in the database to obtain a second database search result.

[0170]

[0172] Step 608: Obtain SQL retrieval result.

[0171]

[0173] The obtaining of the SQL retrieval result can be understood as obtaining the above database search result.

[0172]

[0174] Specifically, the database search result can be obtained by combining the first database search result and the second database search result.

[0173]

[0175] Step 610: Network search.

[0174]

[0176] The network search can be understood as performing network search on the network search problem.

[0175]

[0177] Specifically, after the server obtains the network search question corresponding to the target question through the table question and answer module, the network search engine can be used to perform network search on the network search question.

[0176]

[0178] Step 612: Obtain network search results.

[0177]

[0179] Wherein, obtaining network search results can be understood as obtaining network search results output by the network search engine.

[0178]

[0180] In specific implementation, the network search switch can also be set for the network search engine, and steps 610 to 612 can be executed when the network search switch is in the on state, and the database search results can be directly used to answer according to the above database search results when the network search switch is in the off state.

[0179]

[0181] And in the case that the network search switch is in the on state, the above network search results and database search results can be used to answer according to the data processing model

[0180]

[0182] Step 614: Answer using a data processing model.

[0181]

[0183] Wherein, using a data processing model to answer can be understood as using a data processing model to answer the above target question to obtain a target answer.

[0182]

[0184] Specifically, after obtaining the above network search results and database search results, the target question, target SQL query instruction, database search results, network search results can be input into the data processing model, and the data processing model can refer to the target SQL query instruction, database search results, network search results to answer the target question, and output the target answer corresponding to the target question (i.e. the above data processing result).

[0183]

[0185] In actual application, in order to improve the content quality of the target answer, after obtaining the network search result and the database search result, the question examples associated with the target question and meeting the preset example quantity can be searched from the example library (i.e., the above-mentioned question and answer example database), and the answer examples corresponding to the question examples are taken as the example library search result. Based on this, the target question, the target SQL query instruction, the database search result, the network search result, and the example library search result are input into the data processing model, the data processing model refers to the target SQL query instruction, the database search result, the network search result, and the example library search result, answers the target question, and outputs the target answer corresponding to the target question.

[0184]

[0186] In specific implementation, searching the question examples associated with the target question and meeting the preset example quantity from the example library (i.e., the above-mentioned question and answer example database) can be implemented through vector retrieval or keyword retrieval, and the present disclosure does not limit this.

[0185]

[0187] Taking the target question "xx type competition xxxx year participating team information" as an example, the question and answer method provided by the embodiments of the present disclosure is exemplarily explained.

[0186]

[0188] Referring to FIG. 7, FIG. 7 shows a comparison schematic diagram of a processing flow of a data processing method provided by an embodiment of the present disclosure.

[0187]

[0189] As shown in FIG. 7, the specific implementation mode of steps 702 to 716 shown in FIG. 7 can refer to the specific implementation mode of steps 602 to 614 described above, and will not be repeated here. Through steps 702 to 716, the target answer "xx type competition xxxx year participating team includes xx team, in which xxx obtains xx competition result, yyy obtains yy competition result" corresponding to "xx type competition xxxx year participating team information" can be obtained.

[0188]

[0190] And through the step of step 718, that is, inputting "xx type competition xxxx year participating team information" directly into the data processing model, obtaining that the data processing model outputs "xx type competition xxxx year participating team includes xx team", based on comparison, the question and answer method provided by the embodiments of the present disclosure can obtain a more comprehensive and more accurate target answer.

[0189]

[0191] Step 616: showing the answer to the user.

[0190]

[0192] Specifically, after the server obtains the target answer corresponding to the target question, the target answer corresponding to the target question can be sent to the client. The client displays the target answer to the user through the dialogue interface, the title area, and the answer display area.

[0191]

[0193] Referring to FIG. 8, FIG. 8 shows a third interaction interface schematic diagram of a data processing method according to an embodiment of the present disclosure.

[0192]

[0194] As shown in FIG. 8, the client displays the target answer corresponding to the target question to the user through the interaction interface, including “reference information 1” and “reference information 2”. The “reference information 1” contains “xx type competition xxxx year participating team includes xx team, where xxx obtains xx competition result, yyy obtains yy competition result” information, and displays detailed table information of xx type competition xxxx year competition result to the user (see FIG. 8, the table information contains 2004, A team, 22 times of participation, etc.). The table information is obtained through the second database search result.

[0193]

[0195] In actual application, in order to facilitate the user, the target answer can also carry the data source (reference mark) and the web address (link) of the network search result. Referring to FIG. 9, FIG. 9 shows a fourth interaction interface schematic diagram of a data processing method according to an embodiment of the present disclosure.

[0194]

[0196] As shown in FIG. 9, the client displays the target answer to the user in the dialogue interface as follows:

[0195] “XX type competition XXXX year participating team includes XX team… Network search result: …… Web data source: …… Web address:

[0196]

[0197] Based on this, the user can directly access the network search result through the data source and the web address of the network search result to obtain more material data.

[0197]

[0198] One embodiment of the present disclosure provides a data processing method. The data processing method automatically provides a data processing result corresponding to a target text to a user in a question and answer manner, so that the user does not need to perform manual search, and the labor cost and time cost are reduced. The data processing method realizes data search of multiple data sources through database search, network search, sample library search and the like. In the process of answering the target text, the data processing model can refer to data processing results of database query text, database search results and network search results, so that the comprehensiveness and accuracy of the data processing result obtained by the data processing model are improved, the quality of the data processing result generated for the target text is improved, and the user experience is improved.

[0198]

[0199] The above is a schematic scheme of the data processing method of the present embodiment. It should be noted that the technical scheme of the data processing method and the technical scheme of the data processing method described above belong to the same concept. Details of the technical scheme of the data processing method that are not described in detail can be referred to the description of the technical scheme of the data processing method.

[0199]

[0200] The present disclosure also provides another data processing method embodiment. The data processing method is applied to a data processing platform, and includes: receiving a target event material generation text sent by a client, wherein the target event material generation text is generated by a user through a trigger operation of a user interaction interface of the client; and after obtaining a target event material data processing result corresponding to the target event material generation text, the data processing method further includes: returning the target event material data processing result to the client, so that the client displays the target event material data processing result to the user through the user interaction interface.

[0200]

[0201] The data processing platform can be any platform capable of processing data.

[0201]

[0202] Specifically, the specific implementation mode of the embodiment of the present disclosure can be referred to the above described embodiments, and will not be described here.

[0202]

[0203] The data processing method provided by one embodiment of the present disclosure is applied to a data processing platform, and automatically provides a data processing result corresponding to a target text to a user in a data processing manner, so that the user does not need to perform manual retrieval, and the labor cost and time cost are reduced. In addition, the data search of multiple data sources is implemented through database search and network search, and the database search result and the network search result are obtained. In the process of answering the target text by the data processing model, the data processing result of the database query text, the database search result and the network search result can be referred to, so that the comprehensiveness and accuracy of the data processing result obtained by the data processing model are improved, that is, the quality of the content generated for the target text is improved, and thus the user experience is improved.

[0203]

[0204] The above is a schematic scheme of the data processing method of the embodiment. It should be noted that the technical scheme of the data processing method belongs to the same concept as the technical scheme of the data processing method described above, and details of the technical scheme of the data processing method that are not described in detail can be referred to the description of the technical scheme of the data processing method.

[0204]

[0205] The data processing method provided by one embodiment of the present disclosure is applied to a data processing platform, and automatically provides a data processing result corresponding to a target text to a user in a data processing manner, so that the user does not need to perform manual retrieval, and the labor cost and time cost are reduced. In addition, the data search of multiple data sources is implemented through database search and network search, and the database search result and the network search result are obtained. In the process of answering the target text by the data processing model, the data processing result of the database query text, the database search result and the network search result can be referred to, so that the comprehensiveness and accuracy of the data processing result obtained by the data processing model are improved, that is, the quality of the content generated for the target text is improved, and thus the user experience is improved.

[0205]

[0206] The data processing method provided by one embodiment of the present disclosure is applied to a data processing platform, and automatically provides a data processing result corresponding to a target text to a user in a data processing manner, so that the user does not need to perform manual retrieval, and the labor cost and time cost are reduced. In addition, the data search of multiple data sources is implemented through database search and network search, and the database search result and the network search result are obtained. In the process of answering the target text by the data processing model, the data processing result of the database query text, the database search result and the network search result can be referred to, so that the comprehensiveness and accuracy of the data processing result obtained by the data processing model are improved, that is, the quality of the content generated for the target text is improved, and thus the user experience is improved.

[0206]

[0207] The above is a schematic solution of the question-answering method of the embodiment. It should be noted that the technical solution of the question-answering method and the technical solution of the data processing method described above belong to the same concept, and details of the technical solution of the question-answering method that are not described in detail can be found in the description of the technical solution of the data processing method.

[0207]

[0208] Corresponding to the method embodiment, the disclosure also provides a data processing device embodiment. FIG. 10 shows a structural schematic diagram of a data processing device according to an embodiment of the disclosure. As shown in FIG. 10, the device comprises: a target text processing module 1002 configured to determine a target text and process the target text to obtain a database query text and a network search text corresponding to the target text; a database search module 1004 configured to perform database search on the database query text to obtain a database search result; a network search module 1006 configured to perform network search on the network search text to obtain a network search result; and a data processing result obtaining module 1008 configured to obtain a data processing result corresponding to the target text by using a data processing model according to the target text, the database query text, the database search result, and the network search result, wherein the data processing model is a machine learning model.

[0208]

[0209] Optionally, the target text processing module 1002 is further configured to: determine a preset prompt text, wherein the preset prompt text is used to instruct a text conversion model to convert the target text into a database query text and a network search text; and input the target text and the preset prompt text into the text conversion model to obtain the database query text and the network search text corresponding to the target text, wherein the text conversion model is a machine learning model.

[0209]

[0210] Optionally, the target text processing module 1002 is further configured to: call the text conversion model according to a text conversion model calling interface; and input the target text and the preset prompt text into the text conversion model to obtain the database query text and the network search text corresponding to the target text.

[0210]

[0211] Optionally, the database search module 1004 is further configured to: determine a target database associated with the target text, wherein the target database is constructed through a target application scenario corresponding to the target text; and perform data search in the target database according to the database query text to obtain the database search result.

[0211]

[0212] Optionally, the database search module 1004 is further configured to: perform data search in the target database according to the database query text to obtain a first database search result; update the database query text to obtain updated database query text in a case where the first database search result meets a preset search condition; perform data search in the target database according to the updated database query text to obtain a second database search result; and obtain the database search result according to the first database search result and the second database search result.

[0212]

[0213] Optionally, the network search module 1006 is configured to: obtain network search data output by a target network search engine and / or data attribute information of the network search data according to the network search text by using the target network search engine, wherein the data attribute information includes data sources of the network search data and / or a website address of a webpage to which the network search data belongs; and obtain the network search result according to the network search data and / or the data attribute information of the network search data.

[0213]

[0214] Optionally, the apparatus further includes a sample library search module configured to: perform data processing sample database search on the target text to obtain a sample library search result, wherein the data processing sample database is constructed through a data processing sample pair formed by a text sample and a data processing result sample corresponding to the text sample; and the data processing result obtaining module 1008 is further configured to: obtain a data processing result corresponding to the target text by using a data processing model according to the target text, the database query text, the database search result, the network search result, and the sample library search result.

[0214]

[0215] Optionally, the example library searching module is further configured to: determine the text examples and the data processing result examples of each data processing example pair in the data processing example database; calculate the semantic similarity between the target text and each text example, and determine the target text example according to the semantic similarity between the target text and each text example; and obtain the example library searching result according to the data processing result example corresponding to the target text example.

[0215]

[0216] Optionally, the example library searching module is further configured to: perform vectorization processing on the target text to obtain a target vector corresponding to the target text; determine the text example vectors of each text example, and calculate the vector distance between the target vector and the text example vectors of each text example; and calculate the semantic similarity between the target text and each text example according to the vector distance between the target vector and the text example vectors of each text example.

[0216]

[0217] Optionally, the example library searching module is further configured to: sort each text example according to the semantic similarity between the target text and each text example, and determine the target text example from each text example according to the sorting result and a preset number of examples; or determine the target text example from each text example according to the association relationship between the semantic similarity between the target text and each text example and a preset similarity threshold.

[0217]

[0218] Optionally, the data processing result obtaining module 1008 is further configured to: call the data processing model according to a data processing model calling interface; and input the target text, the database query text, the database searching result, and the network searching result into the data processing model to obtain the data processing result corresponding to the target text.

[0218]

[0219] Optionally, the target text is a target event material generation text, and the data processing result includes target event material.

[0219]

[0220] Optionally, the target text processing module 1002 is further configured to: receive the target text sent by the client, wherein the target text is generated by a user in a trigger operation of a user interaction interface of the client; and the apparatus further includes a data processing result returning module configured to: return the data processing result to the client, so that the client displays the data processing result to the user through the user interaction interface.

[0220]

[0221] The data processing apparatus provided by one or more embodiments of the present disclosure automatically provides a data processing result corresponding to target text to a user, without the user needing to perform manual retrieval, thereby reducing labor costs and time costs. The data processing apparatus implements data search of multiple data sources through database search and network search, obtains database search results and network search results, and enables the data processing model to refer to data processing results of database query text, database search results and network search results in the process of answering the target text, thereby improving the comprehensiveness and accuracy of the data processing result obtained by the data processing model, improving the quality of the content generated for the target text, and improving the user experience.

[0221]

[0222] The above is a schematic scheme of the data processing apparatus of the present embodiment. It should be noted that the technical scheme of the data processing apparatus belongs to the same concept as the technical scheme of the data processing method described above, and details of the technical scheme of the data processing apparatus that are not described in detail can be referred to the description of the technical scheme of the data processing method.

[0222]

[0223] Corresponding to the method embodiments described above, the present disclosure also provides another data processing apparatus embodiment. FIG. 11 shows a structural schematic diagram of another data processing apparatus according to an embodiment of the present disclosure. As shown in FIG. 11, the apparatus includes: a target event material generation text processing module 1102 configured to determine a target event material generation text, and process the target event material generation text to obtain a target event material database query text and a target event material network search text corresponding to the target event material generation text, wherein the target event material generation text is used to generate a target event material content material; a target event material database search module 1104 configured to perform database search on the target event material database query text to obtain a target event material database search result; a target event material network search module 1106 configured to perform network search on the target event material network search text to obtain a target event material network search result; and a target event material data processing result obtaining module 1108 configured to obtain a target event material data processing result corresponding to the target event material generation text by using a data processing model according to the target event material generation text, the target event material database query text, the target event material database search result and the target event material network search result, wherein the data processing model is a machine learning model, and the target event material data processing result contains the target event material content material.

[0223]

[0224] The one or more embodiments of the present disclosure provide a data processing device. The data processing device automatically provides a target event material data processing result corresponding to a target event material generation text to a user in need of the target event material by means of data processing, so that the user does not need to perform manual retrieval, and the labor cost and time cost are reduced. The data processing device realizes multi-data source data search through database search and network search, obtains a target event material database search result and a target event material network search result, and enables the data processing model to refer to the target event material database query text, the target event material database search result and the target event material network search result to obtain the target event material data processing result in the process of answering the target event material generation text, thereby improving the quality of the target event material data processing result obtained by the data processing model, i.e., improving the quality of the target event material content material, and thereby improving the use experience of the user in need of the target event material.

[0224]

[0225] The above is a schematic scheme of the data processing device of the present embodiment. It should be noted that the technical scheme of the data processing device belongs to the same concept as the technical scheme of the data processing method described above, and the details of the technical scheme of the data processing device that are not described in detail can be referred to the description of the technical scheme of the data processing method.

[0225]

[0226] FIG. 12 shows a structural block diagram of a computing device 1200 according to an embodiment of the present disclosure. The components of the computing device 1200 include but are not limited to a memory 1210 and a processor 1220. The processor 1220 is connected to the memory 1210 through a bus 1230, and a database 1250 is used to save data.

[0226]

[0227] The computing device 1200 also includes an access device 1240 that enables the computing device 1200 to communicate via one or more networks 1260. Examples of such networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or an intranet, such as a company's internal network. The access device 1240 can include one or more of any type of network interface (for example, a network interface card (NIC)) such as a wireless Local Area Network (WLAN) interface, a Worldwide Interoperability for Microwave Access (WiMAX) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a Bluetooth® interface, a ZigBee® interface, a Near Field Communication (NFC) interface, or the like.

[0228] In one embodiment of the disclosure, the above-described components of the computing device 1200, as well as other components not shown in FIG. 12, can be connected to each other by a bus. It should be understood that the structure block diagram of the computing device shown in FIG. 12 is merely for the purpose of example, and is not a limitation on the scope of the disclosure. Other components can be added or replaced as needed by those skilled in the art.

[0227]

[0229] The computing device 1200 can be any type of stationary or mobile computing device including a mobile computer or mobile computing device (for example, a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, and the like), a mobile phone (for example, a smartphone), a wearable computing device (for example, a smart watch, smart glasses, and the like), or other type of mobile device, or a stationary computing device such as a desktop computer or a personal computer (PC). The computing device 1200 can also be a mobile or stationary server.

[0228]

[0230] The memory 1210 is configured to store computer programs / instructions, and the processor 1220 is configured to execute the computer programs / instructions stored in the memory 1210. The computer programs / instructions, when executed by the processor, implement the steps of the data processing method.

[0229]

[0231] The above describes a schematic scheme of the computing device of the embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the data processing method described above belong to the same concept, and details of the technical scheme of the computing device that are not described in detail can be referred to the description of the technical scheme of the data processing method.

[0230]

[0232] The embodiment of the present disclosure further provides a computer readable storage medium, which stores computer programs / instructions. The computer programs / instructions, when executed by a processor, implement the steps of the data processing method.

[0231]

[0233] The above describes a schematic scheme of the computer readable storage medium of the embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the data processing method described above belong to the same concept, and details of the technical scheme of the storage medium that are not described in detail can be referred to the description of the technical scheme of the data processing method.

[0232]

[0234] The embodiment of the present disclosure further provides a computer program product, which includes computer programs / instructions. The computer programs / instructions, when executed by a processor, implement the steps of the data processing method.

[0233]

[0235] The above describes a schematic scheme of the computer program product of the embodiment. It should be noted that the technical scheme of the computer program product and the technical scheme of the data processing method described above belong to the same concept, and details of the technical scheme of the computer program product that are not described in detail can be referred to the description of the technical scheme of the data processing method.

[0234]

[0236] The above describes specific embodiments of the present disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in a different order and still achieve desirable results. Additionally, the processes depicted in the figures do not necessarily require the particular order shown or sequential order in order to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous.

[0235]

[0237] The computer program / instructions can include a computer program code, which can be in a form of source code, object code, executable file, or some intermediate form etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier wave signal, telecommunication signal, software distribution medium, etc. It should be noted that the computer readable medium can include contents which can be appropriately added or deleted according to the requirements of patent practice. For example, according to the patent practice in some regions, the computer readable medium does not include electric carrier wave signal and telecommunication signal.

[0236]

[0238] It should be noted that, for the foregoing method embodiments, in order to facilitate description, each is described as a combination of a series of acts, but those skilled in the art should know that the embodiments of the present disclosure are not limited to the order of the acts described, because according to the embodiments of the present disclosure, certain steps can be performed in other orders or at the same time. Secondly, those skilled in the art should know that the embodiments described in the specification are all preferred embodiments, and the acts and modules involved are not necessarily essential to the embodiments of the present disclosure.

[0237]

[0239] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0238]

[0240] The preferred embodiments of the present disclosure disclosed above are only used to help explain the present disclosure. The alternative embodiments do not describe all the details, nor limit the present disclosure to the specific embodiments described. Obviously, according to the content of the embodiments of the present disclosure, many modifications and changes can be made. The present disclosure selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of the present disclosure, so that those skilled in the art can well understand and utilize the present disclosure. The present disclosure is limited only by the claims and their full scope and equivalents.

Claims

Claims 1. A data processing method, comprising: Identify the target text and process it to obtain the database query text and web search text corresponding to the target text; A database search is performed on the database query text to obtain database search results; a web search is performed on the web search text to obtain web search results; based on the target text, the database query text, the database search results, and the web search results, a data processing model is used to obtain the data processing result corresponding to the target text, wherein the data processing model is a machine learning model.

2. The data processing method according to claim 1, wherein processing the target text to obtain the database query text and web search text corresponding to the target text includes: A preset prompt text is determined, wherein the preset prompt text is used to instruct the text conversion model to convert the target text into database query text and web search text; the target text and the preset prompt text are input into the text conversion model to obtain the database query text and the web search text corresponding to the target text, wherein the text conversion model is a machine learning model.

3. The data processing method according to claim 2, wherein inputting the target text and the preset prompt text into the text conversion model to obtain the database query text and the web search text corresponding to the target text includes: According to the text conversion model call interface, the text conversion model is called; the target text and the preset prompt text are input into the text conversion model to obtain the database query text and the web search text corresponding to the target text.

4. The data processing method according to claim 1 or 2, wherein performing a database search on the database query text to obtain database search results includes: A target database associated with the target text is determined, wherein the target database is constructed through a target application scenario corresponding to the target text; based on the database query text, a data search is performed in the target database to obtain the database search results.

5. The data processing method according to claim 4, wherein the step of performing a data search in the target database based on the database query text to obtain the database search results includes: Based on the database query text, a data search is performed in the target database to obtain the first database search result; If the search results of the first database meet the preset search conditions, the database query text is updated to obtain the updated database query text. Based on the updated database query text, a data search is performed in the target database to obtain a second database search result; based on the first database search result and the second database search result, the database search result is obtained.

6. The data processing method according to claim 1 or 2, wherein performing a web search on the web search text to obtain web search results includes: Based on the web search text, using a target web search engine, obtain the web search data output by the target web search engine and / or the data attribute information of the web search data, wherein the data attribute information includes the data source of the web search data and / or the URL of the webpage to which it belongs; based on the web search data and / or the data attribute information of the web search data, obtain the web search results.

7. The data processing method according to claim 1 or 2, further comprising, after determining the target text: The target text is searched using a data processing sample database to obtain sample database search results. The data processing sample database is constructed using data processing sample pairs formed by text samples and corresponding data processing result samples. Obtaining the data processing result corresponding to the target text using a data processing model based on the target text, the database query text, the database search results, and the web search results includes: obtaining the data processing result corresponding to the target text using a data processing model based on the target text, the database query text, the database search results, the web search results, and the sample database search results.

8. The data processing method according to claim 7, wherein performing a data processing sample database search on the target text to obtain sample database search results includes: Determine the text sample and data processing result sample for each data processing sample pair in the data processing sample database; Calculate the semantic similarity between the target text and each text example, and determine the target text example based on the semantic similarity between the target text and each text example; obtain the search results of the example library based on the data processing result example corresponding to the target text example.

9. The data processing method according to claim 8, wherein calculating the semantic similarity between the target text and each text sample includes: The target text is vectorized to obtain the target vector corresponding to the target text; Determine the text sample vector of each text sample, and calculate the vector distance between the target vector and the text sample vector of each text sample; based on the vector distance between the target vector and the text sample vector of each text sample, calculate the semantic similarity between the target text and each text sample.

10. The data processing method according to claim 8, wherein determining the target text sample based on the semantic similarity between the target text and each text sample includes: The text samples are sorted according to their semantic similarity to the target text, and the target text sample is determined from the text samples based on the sorting results and a preset number of samples; or the target text sample is determined from the text samples based on the semantic similarity between the target text and the text samples and the correlation with a preset similarity threshold.

11. The data processing method according to claim 1, wherein obtaining the data processing result corresponding to the target text using a data processing model based on the target text, the database query text, the database search result, and the web search result includes: The data processing model is invoked according to the data processing model call interface; The target text, the database query text, the database search result, and the web search result are input into the data processing model to obtain the data processing result corresponding to the target text.

12. The data processing method according to claim 1 or 2, wherein the target text is text generated from target event material, and the data processing result includes the target event material.

13. The data processing method according to claim 1 or 2, wherein determining the target text includes: The method includes receiving the target text sent by the client, wherein the target text is generated by a user's trigger operation on the client's user interface; after obtaining the data processing result corresponding to the target text, the method further includes: returning the data processing result to the client, so that the client can display the data processing result to the user through the user interface.

14. A data processing method, comprising: The process involves determining the target event material generation text, processing it to obtain the target event material database query text and the target event material web search text corresponding to the target event material generation text, wherein the target event material generation text is used to generate target event materials; performing a database search on the target event material database query text to obtain target event material database search results; performing a web search on the target event material web search text to obtain target event material web search results; and using the target event material generation text, the target event material database query text, the target event material database search results, and the target event material web search results, a data processing model is employed to obtain the target event material data processing result corresponding to the target event material generation text, wherein the data processing model is a machine learning model, and the target event material data processing result includes the target event materials.

15. The data processing method according to claim 14, wherein determining the target event material to generate text includes: The method includes receiving the target event material generated text sent by the client, wherein the target event material generated text is generated by a user's trigger operation on the user interface of the client; after obtaining the target event material data processing result corresponding to the target event material generated text, the method further includes: returning the target event material data processing result to the client, so that the client can display the target event material data processing result to the user through the user interface.

16. A question-and-answer method, comprising: Identify the target problem, process the target problem, and obtain the corresponding database query problem and web search problem; Perform a database search on the database query question to obtain database search results; perform a web search on the web search question to obtain web search results; Based on the target question, the database query question, the database search results, and the online search results, a data processing model is used to obtain the target answer corresponding to the target question, wherein the data processing model is a machine learning model.

17. A data processing method, applied to a data processing platform, the method comprising: Identify the target text and process it to obtain the database query text and web search text corresponding to the target text; Perform a database search on the database query text to obtain database search results; perform a web search on the web search text to obtain web search results; 19. Based on the target text, the database query text, the database search result, and the network search result, a data processing model is used to obtain the data processing result corresponding to the target text, wherein the data processing model is a machine learning model.

18. A data processing apparatus, comprising: The target text processing module is configured to determine the target text and process the target text to obtain the database query text and web search text corresponding to the target text; The database search module is configured to perform a database search on the database query text to obtain the database search results; The web search module is configured to perform a web search on the web search text to obtain web search results; The data processing result acquisition module is configured to obtain the data processing result corresponding to the target text based on the target text, the database query text, the database search result, and the network search result, using a data processing model, wherein the data processing model is a machine learning model.

19. A computing device, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the data processing method according to any one of claims 1 to 17.

20. A computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the data processing method according to any one of claims 1 to 17.

21. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of the data processing method according to any one of claims 1 to 17.

Images