Search task processing method, device and equipment and readable storage medium

Abstract

The invention discloses a search task processing method and device, equipment and a readable storage medium. The method comprises the steps that a search interface is displayed; receiving target search content input in the search interface; in response to input of target search content, content associated data and target task data are displayed in the search interface, the content associated data comprise the target search content, and the target task data are used for indicating task requirements of search tasks corresponding to the target search content. By adopting the method and the device, an interface presentation form and an interaction mode can be enriched in a search scene.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, device, and readable storage medium for processing search tasks. Background Technology

[0002] Search answer refers to the process where a user enters the content they want to search for and issues a search task, and the machine provides corresponding answer content based on the search task issued by the user.

[0003] When users use search applications, after entering their search queries, the application provides corresponding answers for the user to view. Currently, although there are many types and numbers of search applications, the interfaces presented to users during the search process are largely the same, resulting in a monotonous interface presentation and a limited interaction method between users and search applications. Summary of the Invention

[0004] This application provides a method, apparatus, device, and readable storage medium for processing search tasks, which can enrich the interface presentation and interaction methods in search scenarios.

[0005] This application provides a method for processing a search task, including:

[0006] Display the search interface;

[0007] Receive the target search content entered in the search interface;

[0008] In response to the input of target search content, the search interface displays content-related data and target task data. The content-related data contains the target search content, and the target task data indicates the task requirements of the search task corresponding to the target search content.

[0009] One embodiment of this application provides a processing apparatus for a search task, including:

[0010] The interface display module is used to display the search interface;

[0011] The content receiving module is used to receive the target search content entered in the search interface;

[0012] The task display module is used to respond to the input of target search content and display content-related data and target task data in the search interface. The content-related data contains the target search content, and the target task data is used to indicate the task requirements of the search task corresponding to the target search content.

[0013] In one embodiment, after the task display module displays content-related data and target task data in the search interface, the task display module is further specifically used for:

[0014] In response to the target task data being triggered, a results details page is displayed; the results details page includes a preliminary answer area and a task breakdown area.

[0015] The rough solution area displays a rough solution to the target task data, and the task breakdown area displays N subtask data associated with the target task data; N is a positive integer. A subtask data is used to indicate a search task corresponding to a breakdown execution step. The breakdown execution step refers to the steps that need to be performed when executing the search task indicated by the target task data.

[0016] In one embodiment, after the task display module displays a rough solution to the target task data in the rough solution area, the task display module is further specifically used for:

[0017] In response to the operation of viewing content for the rough answer to the target task data, the rough answer page is displayed;

[0018] The rough solution page displays the complete rough solution for the target task data.

[0019] In one embodiment, after the task display module displays N sub-task data associated with the target task data in the task decomposition area, the task display module is further specifically used for:

[0020] In response to the triggering of target subtask data, the breakdown and solution interface is displayed; target subtask data refers to any one of the N subtask data, and the breakdown and solution interface is displayed independently on the results details page;

[0021] The task display area and detailed solution area are shown in the disassembly and solution interface;

[0022] The task display area shows the target subtask data, and the detailed solution area shows the detailed solution content for the target subtask data.

[0023] In one embodiment, the number of disassembly and execution steps corresponding to the search task indicated by the target task data is M, and the total number of subtask data associated with the target task data is M, where M is a positive integer greater than or equal to N.

[0024] During the process of displaying the target subtask data in the task display area, the remaining subtask data is also displayed in the task display area. The display method of the target subtask data in the task display area is different from the display method of the remaining subtask data. The remaining subtask data refers to the subtask data other than the target subtask data among the M subtask data associated with the target task data.

[0025] In one embodiment, the target subtask data is displayed in the task display area using a highlighted mode, and the task display module is further specifically used for:

[0026] In response to the triggering of remaining subtask data, the target subtask data highlighted in the task display area is switched to the remaining subtask data.

[0027] Switch the display of detailed solutions for the target subtask data in the detailed solution area to detailed solutions for the remaining subtask data.

[0028] In one embodiment, the specific implementation method of the task display module displaying content-related data and target task data in the search interface includes:

[0029] While displaying related data in the search interface, the target task data is highlighted.

[0030] The display methods for highlighting target task data include at least one of the following:

[0031] Display the target task data at a specified brightness;

[0032] Display target task data in a dynamically flashing manner;

[0033] The identifier set for displaying target task data during the process of displaying target task data.

[0034] In one embodiment, the content-related data displayed in the search interface is obtained by text matching processing of the target search content; the way the target task data is displayed in the search interface is different from the way the content-related data is displayed in the search interface.

[0035] In one embodiment, the specific implementation of the task display module displaying N sub-task data associated with the target task data in the task decomposition area includes:

[0036] Retrieve the configuration decomposition results corresponding to the target task data from the configuration library. The configuration library contains the correspondence between the configuration task data set and the configuration decomposition result set. There is a correspondence between a configuration task data in the configuration task data set and a configuration decomposition result in the configuration decomposition result set. A configuration task data refers to the task data corresponding to a configuration search task. The configuration decomposition result corresponding to a configuration task data is obtained by decomposing the corresponding configuration search task through an optimized task decomposition model. The configuration decomposition result corresponding to each configuration search task includes at least one configuration decomposition execution step corresponding to the configuration search task, and the subtask data corresponding to each configuration decomposition execution step.

[0037] From at least one subtask data contained in the configuration decomposition result corresponding to the target task data, obtain N subtask data as N subtask data associated with the target task data;

[0038] The task breakdown area displays the data of N subtasks associated with the target task data.

[0039] In one embodiment, the optimization process for optimizing the task decomposition model includes:

[0040] Obtain the task data for the first search task;

[0041] Based on the task data of the first search task, a pre-trained task decomposition model is invoked to decompose the first search task, resulting in the task decomposition result of the first search task. The task decomposition result of the first search task includes at least one decomposition execution step obtained by decomposing the first search task and the sub-task data corresponding to each decomposition execution step. The pre-trained task decomposition model is obtained by pre-training the task decomposition model based on training samples. The training samples are generated based on at least one second search task executed by the sample object within a historical time period.

[0042] Based on the task decomposition results of the first search task, the task correction results for the first search task are obtained. The task correction results for the first search task are obtained after correcting the task decomposition results of the first search task according to the task decomposition criteria. The task correction results for the first search task include at least one labeled execution step obtained by the first search task after correction and the subtask data corresponding to each labeled execution step.

[0043] Based on the difference between the task decomposition results of the first search task and the task correction results of the first search task, the pre-trained task decomposition model is optimized to obtain the optimized task decomposition model.

[0044] In one embodiment, the apparatus further includes a step execution module.

[0045] The step execution module is used to obtain task data of historical search tasks executed by the search object within a historical time period; the search object refers to the object into which the target search content is input; both the target search content and the task data of historical search tasks are text data.

[0046] The step execution module is also used to perform word segmentation on the task data of historical search tasks to obtain the word set corresponding to the task data of historical search tasks.

[0047] The step execution module is also used to obtain the core words corresponding to the task data of the historical search task from the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task refer to words that can reflect the semantics of the task data of the historical search task.

[0048] The step execution module is also used to execute the step of displaying content-related data and target task data in the search interface if the target search content contains core words corresponding to the task data of historical search tasks.

[0049] In one embodiment, the specific implementation method of the step execution module displaying content-related data and target task data in the search interface includes:

[0050] Based on the target search content and historical search task data, recall at least one candidate task data from the configuration task data set;

[0051] Based on the task data of historical search tasks, filter at least one candidate task data to obtain a filter set;

[0052] Determine the estimated trigger rate for each candidate task data in the filter set;

[0053] Obtain the maximum estimated trigger rate from the estimated trigger rate set corresponding to the filter set, and determine the candidate task data corresponding to the maximum estimated trigger rate as the target task data;

[0054] Obtain content-related data for the target search content and display the content-related data and target task data in the search interface.

[0055] In one embodiment, the specific implementation method of the step execution module recalling at least one candidate task data from the configuration task data set based on the target search content and historical search task task data includes:

[0056] Calculate the word matching degree between the target search content and each configuration task data in the configuration task data set according to the keyword matching rules;

[0057] Calculate the semantic matching degree between each configuration task data and the target search content and the task data of historical search tasks according to the semantic matching rules;

[0058] Configuration task data with word matching degree greater than the first matching degree threshold is determined as the first candidate task data, and configuration task data with semantic matching degree greater than the second matching degree threshold is determined as the second candidate task data.

[0059] Both the first candidate task data and the second candidate task data are determined as candidate task data, resulting in at least one candidate task data.

[0060] In one embodiment, each configuration task data in the configuration task data set is text data;

[0061] The specific implementation method of the step execution module in calculating the word matching degree between the target search content and each configuration task data in the configuration task data set according to the keyword matching rules includes:

[0062] Select any one of the configuration task data from the configuration task data set as the target configuration task data;

[0063] The target configuration task data and the target search content are segmented into words according to the first word segmentation method, resulting in the first word set of the target configuration task data under the first word segmentation method and the second word set of the target search content under the first word segmentation method.

[0064] The target configuration task data and the target search content are segmented separately according to the second segmentation method to obtain the third word set of the target configuration task data under the second segmentation method and the fourth word set of the target search content under the second segmentation method; the number of unit characters indicated by the second segmentation method is more than the number of unit characters indicated by the first segmentation method. The number of unit characters refers to the number of characters contained in a unit word.

[0065] Count the number of first words in the first intersection set, the number of second words in the second intersection set, the number of third words in the first word set, and the number of fourth words in the third word set;

[0066] The matching degree calculation function is used to calculate the matching degree of the first word count, the second word count, the third word count, and the fourth word count to obtain the word matching degree between the target configuration task data and the target search content.

[0067] In one embodiment, the specific implementation method of the step execution module calculating the semantic matching degree between each configuration task data and the target search content and the task data of historical search tasks according to semantic matching rules includes:

[0068] Select any one of the configuration task data from the configuration task data set as the target configuration task data;

[0069] The target search content, the task data of historical search tasks, and the target configuration task data are transformed into vectors to obtain the first transformation vector corresponding to the target search content, the second transformation vector corresponding to the task data of historical search tasks, and the third transformation vector corresponding to the target configuration task data.

[0070] The first transformation vector and the second transformation vector are fused to obtain the fused vector;

[0071] Calculate the vector similarity between the fusion vector and the third transformation vector to obtain the semantic matching degree between the target configuration task data, the target search content, and the task data of historical search tasks.

[0072] In one embodiment, both the target search content and the task data for historical search tasks are text data;

[0073] The task display module filters at least one candidate task based on historical search task data to obtain a filter set, including:

[0074] The task data of historical search tasks is segmented to obtain the word set corresponding to the task data of historical search tasks.

[0075] Extract the core words corresponding to the task data of the historical search task from the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task refer to the words that can reflect the textual semantics of the task data of the historical search task.

[0076] Aggregate the data of at least one candidate task to obtain a candidate set;

[0077] The candidate task data that does not contain the core words corresponding to the task data of historical search tasks are deleted from the candidate set to obtain the filter set.

[0078] One embodiment of this application provides a computer device, including: a processor and a memory;

[0079] The memory stores a computer program, which, when executed by a processor, causes the processor to perform the methods described in the embodiments of this application.

[0080] One aspect of this application provides a computer-readable storage medium storing a computer program, which includes program instructions. When executed by a processor, the program instructions perform the methods described in this application.

[0081] One aspect of this application provides a computer program product comprising a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium and executes the computer program, causing the computer device to perform the method provided in one aspect of the embodiments of this application.

[0082] In this embodiment, for a specific search application, when an object uses the search application, a search interface can be displayed within the application for the object to search. The object can input the content they want to search for in the search interface. After the object inputs the search content, this application can receive the target search content entered by the object in the search interface. Then, this application can display content-related data and target task data in the search interface. The target task data in this application differs from the content-related data. The content-related data only contains the target search content, but the target task data indicates the task requirements of the search task corresponding to the target search content. That is, after an object inputs search content, this application can analyze and process the search / task requirements of the search content and display task data reflecting those requirements. This presentation of task data not only enriches the relevant information presented in the search application interface, but also guides the object to trigger the task data to execute a search task that meets their search needs. This not only enriches the interaction between the object and the search application but also helps to achieve efficient searching for the object and obtain accurate answers. In summary, this application can provide richer interface presentation forms for search applications in search scenarios and enrich the interaction methods between objects and search applications. Attached Figure Description

[0083] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0084] Figure 1 This is a network architecture diagram of a search task processing system provided in an embodiment of this application;

[0085] Figure 2 This is a schematic diagram of a traditional search and answer scenario provided in an embodiment of this application;

[0086] Figure 3 This is a flowchart illustrating a method for processing a search task provided in an exemplary embodiment of this application;

[0087] Figure 4 This is a schematic diagram of a scenario for displaying task data and content-related data, provided in an embodiment of this application.

[0088] Figure 5 This is a schematic diagram of a process for displaying target task data and content association data in a search interface, provided by an embodiment of this application;

[0089] Figure 6 This is a schematic diagram of a logical architecture for determining target task data based on target search content, provided in an embodiment of this application.

[0090] Figure 7 This is a flowchart illustrating how to display the answer content corresponding to a search task, as provided in an embodiment of this application.

[0091] Figure 8 This is a schematic diagram illustrating a scenario where the answer is displayed after the target task data is triggered, as provided in an embodiment of this application.

[0092] Figure 9 This is a schematic diagram illustrating a scenario where the answer is displayed after triggering target subtask data, as provided in an embodiment of this application.

[0093] Figure 10 This is a schematic diagram illustrating a scenario of sliding to view subtask data, provided in an embodiment of this application.

[0094] Figure 11 This is a schematic diagram of a process for acquiring and displaying N subtask data associated with target task data, provided in an embodiment of this application;

[0095] Figure 12 This is a schematic diagram of the structure of a search task processing device provided in an embodiment of this application;

[0096] Figure 13 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0097] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0098] Please see Figure 1 , Figure 1 This is a network architecture diagram of a search task processing system provided in an embodiment of this application. For example... Figure 1As shown, this network architecture may include server 1000 and a terminal device cluster. The terminal device cluster may include one or more terminal devices; the number of terminal devices is not limited here. Figure 1 As shown, multiple terminal devices may include terminal device 100a, terminal device 100b, terminal device 100c, ..., terminal device 100n; as Figure 1 As shown, terminal devices 100a, 100b, 100c, ..., 100n can each connect to server 1000 via a network, enabling data interaction between each terminal device and server 1000. Furthermore, any terminal device in the terminal device cluster can refer to a smart device running an operating system; however, this embodiment does not specifically limit the operating system of the terminal devices.

[0099] like Figure 1 The terminal devices in the data processing system shown can be smartphones, tablets, laptops, PDAs, desktop computers, mobile internet devices (MIDs), POS (Point of Sales) machines, smart speakers, smart TVs, smartwatches, smart in-vehicle terminals, virtual reality (VR) devices, augmented reality (AR) devices, etc., but are not limited to these. Terminal devices are often equipped with display devices, which can be monitors, displays, touchscreens, etc., and touchscreens can be touchscreens, touch panels, etc.

[0100] like Figure 1 The data shown in the system indicates that the server can be a single physical server, a server cluster consisting of multiple physical servers, or a distributed system. It can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. Terminal devices and servers can be connected directly or indirectly via wired or wireless communication; this application does not impose any restrictions on this.

[0101] This application embodiment can select one terminal device from multiple terminal devices as the target terminal device. This terminal device may include, but is not limited to, smart terminals such as smartphones, tablets, laptops, desktop computers, smart TVs, smart speakers, desktop computers, smartwatches, smart in-vehicle devices, smart home devices, and smart voice interaction devices. For example, this application embodiment can... Figure 1The terminal device 100a shown serves as the target terminal device. This target terminal device may integrate a target application, allowing it to interact with the server 1000 via the target application. The server 1000 in this application can obtain business data based on these applications. For example, the server 1000 can obtain business data through a bound account of an object. The bound account can refer to the account bound to the object in the application; the object can log in to the application, upload data, retrieve data, etc., through its corresponding bound account, and the server can also obtain the object's login status, uploaded data, send data to the object, etc., through its bound account.

[0102] It is understandable that, such as Figure 1 Each terminal device shown can have the target application installed. When the target application runs on each terminal device, it can interact with... Figure 1 The servers 1000 shown interact with each other, enabling each server 1000 to receive business data from each terminal device. The target application can include applications capable of displaying text, images, audio, and video data, such as instant messaging applications, browsers, video applications, entertainment applications (e.g., games), etc., which will not be listed here. The application can be a standalone application or an embedded sub-application integrated into an application (e.g., educational applications, multimedia applications, etc.), without limitation. Taking a browser as an example, server 1000 can be a collection of multiple servers, including a browser-specific backend server and a data processing server. Therefore, each terminal device can transmit data with server 1000 through the browser. For example, each terminal device can upload its local images, videos, etc., to server 1000 through the browser. Server 1000 can then process the received images, videos, etc., for example, by distributing the data to other terminal devices or transmitting it to a cloud server.

[0103] In this embodiment, the target application provides a search function, allowing users to input the content they wish to search for (the input content can be called the search content) to obtain corresponding answers from the target application. For example, taking a browser as the target application, the browser provides a search function. Users can input the content they wish to search for (the input content can be called the search content) and perform a search operation on that content. After performing the search operation, it can be considered that the user has issued a search task in the browser, and the browser can output and display answers based on the search content. In practical applications, the search content input by the user may not accurately express their search needs, thus failing to obtain accurate answers. Therefore, users may spend considerable time carefully considering their wording and repeatedly editing their search content to obtain more accurate answers.

[0104] To address the aforementioned issues and improve the efficiency of users identifying search terms, traditional technologies typically employ an associative mechanism. After a user inputs their search terms, the system can supplement the input with keywords. Based on existing search terms in the corpus, it automatically adds keywords to form a complete associative search query. This query is then displayed to the user. If the user clicks on an associative query, it indicates that the user has completed inputting the complete search query and a search task for that query has been initiated. For example, after a user enters the search term "spicy crayfish" into a browser, the browser can traverse the corpus to find search terms containing "spicy crayfish." Taking "home-style cooking method for spicy crayfish" as an example, the browser can append the keyword "home-style cooking method" to the search terms, thus obtaining a complete search query, "home-style cooking method for spicy crayfish" (in fact, this complete search query is the search term "home-style cooking method for spicy crayfish" in the corpus). The browser can then display this query, and the user can initiate a search task by triggering the associated search query, "home-style cooking method for spicy crayfish." For easier understanding, please refer to [further details omitted]. Figure 2 , Figure 2 This is a schematic diagram illustrating a traditional search and answering scenario provided in an embodiment of this application. For example... Figure 2 As shown, taking a browser as an example, a user can launch a browser on a terminal device 100a. First, the browser can display a search interface 2001 to user a. The search interface 2001 may contain a search box M, and user a can enter the content they want to search for in the search box M.

[0105] likeFigure 2 As shown, assuming user a enters the text data "teacher's speech at parent-teacher conference" in the search box M, the browser's server can traverse the various search terms in the corpus based on user a's input. This allows the search term to be found that contains "teacher's speech at parent-teacher conference" and begins with that phrase. Here, it is assumed that the search terms found by the server include: search term 1 "concise and insightful teacher's speech at parent-teacher conference", search term 2 "excellent opening of teacher's speech at parent-teacher conference", search term 3 "what should a teacher say at a parent-teacher conference", and search term 4 "teacher's speech draft at parent-teacher conference". After finding these search terms, the server can return them to terminal device 100a, which can then display these four search terms on the search interface 2001. The user can then trigger any one of these search terms to perform a search operation. It should be understood that since these search terms all contain the search content entered by user A and begin with the search content entered by user A, these search terms may be able to match user A's search needs / intent.

[0106] Here is as Figure 2 As shown, assuming user A triggers a search operation on search term 3 "What should teachers say at a parent-teacher meeting?", terminal device 100a can determine that user A has performed a search operation on this search term, issuing a search task. Terminal device 100a can then respond to this trigger operation, retrieve the answer content corresponding to search term 3 from the server, display the answer interface 2002, and show the answer content corresponding to this search task in the answer interface 2002. For example, as... Figure 2 As shown, terminal device 100a can display the content of the teacher's speech at the parent-teacher meeting on the answer interface 2002. If the amount of data in the answer content is large and cannot be fully displayed on the answer interface 2002, terminal device 100a can provide a full-text view control on the answer interface 2002, which user a can trigger to view the complete answer content.

[0107] However, this associative search mechanism, which supplements the user's search query with related information, may not provide the exact content the user is looking for. Therefore, through mechanisms like... Figure 2 The illustrated embodiments demonstrate that, in conventional technologies, even if related search queries can be displayed based on the search content, the displayed related search queries are very limited, consisting mainly of keyword supplements to the search content. This fails to meet the user's search needs, and the user still needs to spend time repeatedly editing the description of the search content, which reduces search efficiency.

[0108] Based on this, this application provides a search task processing scheme (actually a task data display scheme related to search tasks), adding a function to mine the search intent of the search content, and can display task data reflecting the user's search task requirements, enabling the user to quickly complete the search through this task data during the search process, thus improving search efficiency. The search task processing scheme designed in this application can include at least three consecutive steps: 1. When a user wants to search for something, they can launch a search application on the terminal device (i.e., the target application providing the search function, such as a browser). Then, the terminal device can display the search interface corresponding to the search application, where the user can enter the content they want to search for; 2. After the user enters the content they want to search for in the search interface, the content entered by the user can be called the target search content, and the terminal device can receive the target search content entered; 3. After receiving the target search content entered, as in the traditional method, the terminal device can... Text matching is performed in the corpus to find search terms that contain the target search content and begin with it. These search terms can serve as content-related data for the target search content. In addition, the terminal device can send the target search content to the corresponding server. The server can then analyze the user's search intent / demand based on the target search content and find the target task data reflecting the search intent / demand from pre-configured task data. The server can then return the target task data to the terminal device, which can display the target task data along with the aforementioned content-related data in the search interface. In this context, after a user inputs their target search content, it can be understood that the user is about to initiate a search task based on that content. By analyzing the search intent / needs based on the target search content, target task data reflecting this intent / needs can be identified. This target task data can be text data. Since it reflects the search intent / needs, it can be considered data used to reflect the task requirements of this search. Unlike content-related data, this target task data may not contain the target search content, but it accurately reflects the user's search intent / needs. Thus, the user can directly trigger the target task data to initiate a search task and accurately obtain the answers to their search queries without needing to search based on the target search content. In this application, the configuration task data refers to the text data configured by the search application's operating entity to reflect its search needs. Each configuration search task can refer to a pre-configured search task by the relevant operating entity (e.g., the entity developing the search application in the search scenario).

[0109] It should be understood that this application, by conducting in-depth analysis of the user's input search content to uncover deeper search intent / needs, can obtain task data that more accurately reflects the task requirements of the user's upcoming search task. Displaying this task data within the search interface allows users to perform quick and accurate searches without repeatedly editing the wording of their search content, significantly improving search efficiency. Compared to traditional search methods, this approach, which mines and analyzes the task requirements of the user's upcoming search task based on the search content and displays task data reflecting these requirements, more accurately reveals the user's search intent / needs. This saves users time in determining the search content, thus improving search efficiency. Therefore, this application's search display method not only enriches the presentation of search content association data but also enhances search efficiency.

[0110] It is understood that the methods provided in this application embodiment can be executed by computer devices, including but not limited to terminal devices and / or servers. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0111] The terminal devices and servers can be connected directly or indirectly through wired or wireless communication, and this application does not impose any restrictions on this.

[0112] Optionally, and understandably, the aforementioned computer devices (such as server 1000, terminal device 100a, terminal device 100b, etc.) can be nodes in a distributed system. This distributed system can be a blockchain system, formed by connecting multiple nodes through network communication. In this distributed system, any type of computer device, such as servers, terminal devices, or other electronic devices, can become a node in the blockchain system by joining the peer-to-peer network. For ease of understanding, the concept of blockchain is explained below: Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. It is mainly used to organize data in chronological order and encrypt it into a ledger, making it tamper-proof and forgery-proof. It also allows for data verification, storage, and updating. When a computer device is a blockchain node, due to the immutability and anti-forgery characteristics of blockchain, the data in this application (such as user search content, etc.) possesses authenticity and security, thus making the results obtained after data processing based on this data more reliable.

[0113] It should be noted that, in the specific embodiments of this application, user-related data such as user information and user data (e.g., user-input search content) are only acquired and processed after obtaining permission granted by the user. In other words, when the embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant regions.

[0114] Based on the above-described scheme, this application proposes a more detailed method for processing search tasks. The following will describe in detail the method for processing search tasks proposed in this application with reference to the accompanying drawings.

[0115] Please see Figure 3 , Figure 3 This is a flowchart illustrating a search task processing method provided in an exemplary embodiment of this application. The search task processing method can be executed by a computer device in the aforementioned system, and this computer device can be... Figure 1 In the corresponding embodiment, any terminal device (such as terminal device 100a) or server 1000 in the terminal device cluster can be used. The computer equipment can also be composed of terminal devices and a business server. For ease of understanding, this embodiment uses the method executed by the aforementioned terminal device as an example for explanation. The processing method of this search task can include at least the following steps S301-S303:

[0116] Step S301: Display the search interface.

[0117] In this application, the search interface can refer to the interface provided by a search application for users to perform searches. When a user runs the search application on their terminal device, the terminal device can display this search interface, and the user can enter the content they want to search for in the search interface. The search application can refer to an application that provides search functionality (such as the aforementioned target application, such as a browser, video application, game application, etc.).

[0118] Step S302: Receive the target search content entered in the search interface.

[0119] In this application, users can enter the content they want to search for in the search interface. The content entered by the user can be called the search content, and the target search content can refer to the search content entered by the user in the search interface. The terminal device can receive the target search content entered by the user in real time.

[0120] In step S303, in response to the input of the target search content, the content association data and target task data are displayed in the search interface. The content association data contains the target search content, and the target task data is used to indicate the task requirements of the search task corresponding to the target search content.

[0121] In this application, after receiving the target search content input by the user, the terminal device can obtain various search terms in the corpus based on the target search content (each search term can refer to a term searched by different users in the past. For example, if a user has searched for "how to stew crucian carp and tofu soup to make it milky white", that is, the user has issued a search task based on the text data "how to stew crucian carp and tofu soup to make it milky white", then the text data "how to stew crucian carp and tofu soup to make it milky white" can be added to the corpus as a search term, and the search task issued by the user can be regarded as a historical search task). Then, the terminal device can perform text matching processing between each search term and the target search content to match the search terms containing the target search content from each search term. These search terms can be used as content association data of the target search content. It should be noted that when performing text matching between various search terms and the target search content, after finding search terms that contain the target search content, these search terms can be further filtered. Specifically, from the search terms that contain the target search content, search terms that begin with the target search content can be identified. Then, search terms that begin with the target search content can be identified as content-related data of the target search content.

[0122] In addition, based on the target search content, this application can perform in-depth analysis of the user's search intent / needs to find target task data that reflects the search intent / needs from the configured task data set. Each task data in this application can refer to a piece of text data. Different search tasks can be pre-configured to obtain a set of configured search tasks. Furthermore, each set of configured search tasks can be configured with a piece of task data reflecting its task requirements, ultimately allowing one configured search task to correspond to one piece of configured task data. For example, this application can configure a search task 1, where the configured task data for search task 1 is the text data "Learn Advanced Mathematics Quickly in Two Months." Therefore, the configured search task can refer to a search task issued based on the text data "Learn Advanced Mathematics Quickly in Two Months." Thus, after obtaining the target search content, semantic analysis can be performed based on the semantics of the target search content and the semantics of each configuration task data in the configuration task data set to find target task data that can reflect the search intent / need. It is worth noting that since the user's purpose after entering the target search content is to issue a search task to obtain accurate answers, target task data can also be understood as reflecting the task requirements of the search task that the user is about to issue but has not yet issued. In other words, the task requirements of a search task in this application are actually the user's search intent / need. For a detailed explanation of how to determine target task data based on the target search content, please refer to the following sections. Figure 5 The relevant descriptions in the corresponding embodiments.

[0123] After acquiring the target task data and related content data, the terminal device can display both in the search interface, allowing users to view both simultaneously. It should be understood that target task data more accurately reflects the search intent / need than related content data. Therefore, while displaying related content data in the search interface, the target task data can be highlighted. This allows users to initiate a search by triggering the highlighted target task data without needing to edit the search content further. The target task data can be displayed normally or highlighted; highlighting it provides greater visibility and more intuitive guidance for users to trigger the search. In the search interface, the display methods for highlighting target task data may include, but are not limited to, any of the following: 1. Displaying the target task data at a specified brightness, i.e., when displaying the target task data, the brightness can be adjusted to a specified level to highlight the target task data; 2. Displaying the target task data in a dynamically flashing manner, i.e., when displaying the target task data, the target task data can be dynamically flashed to highlight the target task data; 3. During the display of the target task data, a set identifier can be set for the target task data display. The set identifier can be set based on actual business needs, for example, it can be set as a star identifier, a firework identifier, etc.

[0124] It's important to note that when simultaneously displaying various content-related data and target task data, the display methods for each search term and the target task data need to differ. In practical scenarios, the target task data can be highlighted in addition to displaying each search term. This highlighting visually differentiates the target task data from the various content-related data. In other words, the display method for content-related data in the search interface differs from the display method for target task data. For clarity, please refer to [link to relevant documentation]. Figure 4 , Figure 4 This is a schematic diagram illustrating a scenario for displaying task data and content-related data, provided in an embodiment of this application. For example... Figure 4 As shown, taking a browser as the target application as an example, user a can start the browser on terminal device 100a. First, the browser can display a search interface 4001 to user a. The search interface 4001 may contain a search box M, and user a can enter the content they want to search for in the search box M.

[0125] like Figure 4As shown, assuming user a enters the text data "teacher's speech at parent-teacher conference" into the search box M, the browser's server can perform text matching based on the target search content entered by user a. Specifically, the server can traverse each search term in the corpus based on the target search content to find content-related data that contains the target search content "teacher's speech at parent-teacher conference" and begins with the target search content "teacher's speech at parent-teacher conference". Here, it is assumed that the content-related data packets found by the server include: content-related data 1 "concise and insightful teacher's speech at parent-teacher conference", content-related data 2 "excellent opening of teacher's speech at parent-teacher conference", and content-related data 3 "speech draft of teacher's speech at parent-teacher conference". After finding the above content-related data, the server can return each content-related data to the terminal device 100a. The terminal device 100a can display these three content-related data in the search interface 4001. The user can trigger any one of the content-related data to perform a search operation and issue a search task. Specifically, when displaying each piece of content-related data, terminal device 100a can highlight characters in the content-related data other than the target search content (e.g., such as...). Figure 4 As shown, when displaying various content-related data, terminal device 100a can bold all characters except for "Teacher's speech at the parent-teacher meeting".

[0126] In addition to finding and displaying related data, the server corresponding to the terminal device can also analyze the search intent / demand based on the target search content to determine the search demand behind user A's input of the target search content "teacher's speech at the parent-teacher meeting". The analyzed search intent can be considered a task requirement for the search task that user A will perform. Through the analysis of search intent, the server can match a target task data that reflects the task requirement of this search task from the configuration task data set corresponding to the configuration search task set. Here, it is assumed that the target task data is the text data "preparation of key aspects of the teacher's first parent-teacher meeting". That is to say, user A's purpose in inputting the target search content "teacher's speech at the parent-teacher meeting" is not to inquire about the content of the teacher's speech at the parent-teacher meeting, but to search for information on the preparations that should be made for the first parent-teacher meeting. After determining the target task data, the server can return the target task data to the terminal device 100a. The terminal device 100a can also display the target task data in the search interface 4001. When displaying the target task data, since the target task data reflects the user a's search intent better than the various related content data, the terminal device 100a can highlight the entire target task data. For example, the terminal device 100a can display the entire target task data in bold. In this way, by highlighting the entire target task data in bold, the importance of the target task data can be more intuitively shown, thereby guiding the user a to trigger the target task data to execute the current search task.

[0127] In this embodiment, after an object inputs search content, the search / task requirements can be analyzed and processed, and task data reflecting those requirements can be displayed. This presentation of task data not only enriches the relevant information presented on the interface of the search application, but also guides the object to trigger the task data to execute a search task that meets its search needs. This not only enriches the interaction between the object and the search application, but also helps to achieve efficient searching and obtain accurate answers. In summary, this application can provide a richer interface presentation for search applications in search scenarios and enrich the interaction between objects and search applications.

[0128] Furthermore, to more clearly understand the method for determining the target task data in this application, the process of displaying the target task data will be described below with reference to the accompanying drawings. This description includes the process of determining the target task data based on the target search content. For details, please refer to the accompanying drawings. Figure 5 , Figure 5This is a schematic diagram of a process for displaying target task data and content association data in a search interface, provided by an embodiment of this application. This process can correspond to the above. Figure 3 The corresponding embodiment describes the process of displaying target task data in the search interface. For example... Figure 5 As shown, the process may include at least the following steps S501-S505:

[0129] Step S501: Based on the target search content and the task data of historical search tasks, recall at least one candidate task data from the configuration task data set.

[0130] In this application, a historical search task can refer to a search task executed by the search object (i.e., the object that will issue the search task when the target search content is entered this time, such as a user) within a historical time period. This historical time period can be a time period preceding the current moment (i.e., the moment the target search content is received). Specifically, the historical search task can refer to the previous search task executed by the search object before the current input of the target search content. The task data of the historical search task can refer to the content searched by the search object when executing the previous search task. For example, if the search object searched for "how to steam turbot" in its previous search, then the search object's previous search task can be considered a historical search task, and the text data "how to steam turbot" can be used as the task data for that historical search task.

[0131] After obtaining the task data of the historical search task, at least one candidate task data can be recalled from the configured task data set based on the target search content and the task data of the historical search task. The specific implementation process can be as follows. First, since the target search content, the task data of the historical search task, and each configured task data are all text data, the recall can be performed according to the processing method of text data. Specifically, first, first, according to the keyword matching rule, calculate the word matching degree between the target search content and each configured task data in the configured task data set. Taking a certain configured task data as an example, its specific implementation process can at least include but is not limited to: First, for the convenience of distinction, this configured task data can be called the target configured task data. Then, the target configured task data and the target search content can be segmented respectively according to the first segmentation method. As a result, the first word set of the target configured task data under the first segmentation method and the second word set of the target search content under the first segmentation method can be obtained. The first segmentation method here can refer to the unit component segmentation method. The unit component segmentation method is a method of segmenting text data according to unit groups. Each segmentation obtained by the unit component segmentation method is a single character. For example, taking the text data "I ate a Hami melon today" as an example, after segmenting it by the unit component segmentation method, the obtained word set is {I, today, ate, a, Hami, melon}, and each segmentation (also called a word) in the word set is a single character. Based on this, after segmenting the target configured task data according to the first segmentation method, the obtained word set can be called the first word set. The first word set will contain different segmentations (words), and each segmentation is a single character. Similarly, after segmenting the target search content according to the first segmentation method, the obtained word set can be called the second word set. The second word set will contain different segmentations (words), and each segmentation is also a single character.

[0132] Further, the target configuration task data and the target search content can be separately segmented according to the second segmentation method, whereby a third word set of the target configuration task data under the second segmentation method and a fourth word set of the target search content under the second segmentation method can be obtained; wherein, the second segmentation method here may refer to a multi-component segmentation method (preferably a binary-component segmentation method in this application), and the multi-component segmentation method refers to a method of segmenting text data according to multi-components, and each segment obtained by the multi-component segmentation method consists of multiple characters (the number of characters specifically included in a segment corresponds to the multi-component; for example, in the binary-component segmentation method, a segment after segmentation specifically consists of two characters; in the ternary-component segmentation method, a segment after segmentation specifically consists of three characters). For example, taking the binary-component segmentation method as an example, assuming the text data is "I ate Hami melons today", after segmenting it by the binary-component segmentation method, the obtained word set is {I today, today ate, ate Hami, Hami melons}, and each segment (also called a word) in the word set consists of two adjacent characters in the text data. Based on this, after segmenting the target configuration task data according to the second segmentation method, the obtained word set can be called the third word set, and this third word set will contain different segments (words), and each segment consists of multiple characters; similarly, after segmenting the target search content according to the second segmentation method, the obtained word set can be called the fourth word set, and this fourth word set will contain different segments (words), and each segment also consists of multiple characters. Based on the above, the number of unit characters indicated by the second segmentation method is more than the number of unit characters indicated by the first segmentation method, and the number of unit characters refers to the number of characters included in a unit word, and a unit word refers to any segment obtained after segmentation processing.

[0133] After obtaining the first set of words, the second set of words, the third set of words, and the fourth set of words, the word matching degree between the target configuration task data and the target search content can be determined based on the first intersection between the first set of words and the second set of words, and the second intersection between the third set of words and the fourth set of words. Specifically, the number of first words contained in the first intersection, the number of second words contained in the second intersection, the number of third words contained in the first set of words, and the number of fourth words contained in the third set of words can be counted. After obtaining the number of words contained in each of the above sets, the matching degree calculation function can be used to calculate the matching degree of the first word count, the second word count, the third word count corresponding to the target configuration task data, and the fourth word count corresponding to the target configuration task data. In this way, the word matching degree between the target configuration task data and the target search content can be obtained. The specific implementation method of using the matching degree calculation function to calculate the matching degree to obtain the word matching degree can be as shown in formula (1):

[0134]

[0135] Wherein, as shown in formula (1), Score can refer to the word matching degree between the target configuration task data and the target search content; J1 num The number of first words that can be used to characterize the first intersection set mentioned above; J2 num R1 can be used to characterize the number of second words contained in the second intersection set mentioned above; num R2 can be used to characterize the number of third words in the first word set obtained after segmenting the target configuration task data according to the first segmentation method; num It can be used to characterize the number of fourth words in the third word set obtained after the target configuration task data is segmented according to the second word segmentation method.

[0136] After calculating the word matching degree between each configuration task data and the target search content, each configuration task data in the configuration task data set with a word matching degree greater than the first matching degree threshold can be identified as the first candidate task data.

[0137] Simultaneously, according to semantic matching rules, the semantic matching degree between each configuration task data and the target search content and the task data of historical search tasks can be calculated. Specifically, any configuration task data in the configuration task data set can be identified as the target configuration task data. Then, the target search content, the task data of historical search tasks, and the target configuration task data can be vector-transformed respectively, thereby obtaining the first transformation vector corresponding to the target search content, the second transformation vector corresponding to the task data of historical search tasks, and the third transformation vector corresponding to the target configuration task data. Further, the first transformation vector and the second transformation vector can be fused to obtain the fused vector, where vector fusion can refer to vector concatenation. After obtaining the fused vectors corresponding to the task data of the target search content and the historical search tasks, the vector similarity between the fused vector and the third transformation vector can be calculated. The method of calculating the vector similarity is not limited in this application. For example, the vector distance between the fused vector and the third transformation vector can be calculated, and this vector distance can be used as the vector similarity between the two. Alternatively, the cosine similarity between the fused vector and the third transformation vector can be calculated, and this cosine similarity can be used as the vector similarity between the two. After obtaining the vector similarity between the two, the vector similarity can be used as the semantic matching degree between the target configuration task data, the target search content, and the task data of historical search tasks.

[0138] Following the above method, the semantic matching degree corresponding to each configuration task data can be calculated. Each configuration task data in the configuration task data set with a semantic matching degree greater than the second matching degree threshold can be determined as the second candidate task data. Finally, each first candidate task data and each second candidate task data can be determined as candidate task data, thereby obtaining at least one candidate task data.

[0139] The thresholds in this application (such as the first matching threshold and the second matching threshold mentioned above) can be set based on actual business needs. In general, after obtaining the target search content and historical search task data, this application can combine the current target search content with the historical search task data, and use keyword matching rules and semantic matching rules to perform corresponding calculations to retrieve at least one candidate task data from the configured task data set. These candidate task data are matched with the current target search content, both in terms of content itself and semantics.

[0140] Step S502: Filter at least one candidate task data based on the task data of historical search tasks to obtain a filter set.

[0141] In this application, after obtaining each candidate task data, it can be de-duplicated and filtered to filter out candidate task data with low relevance to the current target search content or the task data of historical search tasks. In this way, the accuracy of the final obtained target task data can be improved. Specifically, for the implementation method of filtering at least one candidate task data based on the task data of historical search tasks to obtain a filtered set, it can at least include but is not limited to: First, the task data of historical search tasks can be segmented, and thus a word set corresponding to the task data of historical search tasks can be obtained; where the segmentation here is different from the above-mentioned segmentation according to the first segmentation method and the second segmentation method. The segmentation here can refer to segmenting the task data of historical search tasks according to a mixed segmentation method, which includes the above-mentioned first segmentation method and the second segmentation method. The mixed segmentation method will perform unit component segmentation, binary component segmentation, ternary component segmentation,..., (K-1) (K is the total number of characters included in the text data) tuple segmentation on the text data. The number of characters included in each segment obtained by segmenting according to the mixed segmentation method may be different. For example, taking the text data "I ate rice" as an example, after segmenting it according to the mixed segmentation method, the obtained word set is {I, ate, rice, I ate, ate rice}. Based on this, after segmenting the task data of historical search tasks, the obtained word set will include different types of segments, and the number of characters included in each segment may be different, but for the segment with the largest number of characters, the number of characters it includes will not exceed the total number of characters included in the task data of this historical search task.

[0142] After obtaining the word set corresponding to the task data of historical search tasks, the core words corresponding to the task data of historical search tasks can be obtained from this word set. These core words are words that reflect the textual semantics of the task data of historical search tasks. These core words can be identified using a model with core word recognition capabilities. For example, taking the text data "I went to City A today," the place I went to and the time I went are important and reflect the semantics of the text. Therefore, "today" and "City A" can be considered as the core words corresponding to this text data. After obtaining the core words corresponding to the task data of historical search tasks, at least one candidate task data can be aggregated (aggregation means combining them into a set), thus obtaining a candidate set. Then, each candidate task data in the candidate set can be traversed, and candidate task data that does not contain the core words corresponding to the task data of historical search tasks can be deleted. The candidate set after deletion can be called the filter set.

[0143] Step S503: Determine the estimated trigger rate corresponding to each candidate task data in the filter set.

[0144] In this application, after filtering the candidate task data to obtain a filter set, the estimated trigger rate corresponding to each candidate task data in the filter set can be determined. This estimated trigger rate is the probability that the search object might trigger that candidate task data. This probability can be achieved by calling a probability prediction model with classification capabilities (any artificial intelligence model with classification capabilities, which can be pre-trained to make the model output more accurate). It should be noted that if, after the above filtering, the obtained filter set does not contain any candidate task data, then the subsequent steps do not need to be performed, and it is not necessary to provide target task data for the search object.

[0145] Step S504: Obtain the maximum estimated trigger rate from the estimated trigger rate set corresponding to the filter set, and determine the candidate task data corresponding to the maximum estimated trigger rate as the target task data.

[0146] In this application, after obtaining the estimated trigger rate corresponding to each candidate task data in the filter set, an estimated trigger rate set can be obtained. Then, the maximum estimated trigger rate can be obtained from the estimated trigger rate set, and the candidate task data corresponding to the maximum estimated trigger rate can be determined as the target task data.

[0147] It is worth noting that in some scenarios, there are certain requirements regarding the quantity of target task data, which may require multiple data sets. In such cases, based on actual business needs, the estimated trigger rates in the estimated trigger rate set can be sorted in descending order, for example, to obtain an estimated trigger rate sequence. Then, the first W (W being the number of target task data required by the actual business needs) estimated trigger rates in the sequence can be used as target trigger rates. Subsequently, the candidate task data corresponding to each target trigger rate can be identified as a target task data set, thus obtaining W target task data sets.

[0148] Step S505: Obtain the content association data of the target search content, and display the content association data and target task data in the search interface.

[0149] In this application, after the target task data is determined, the terminal device can obtain the content association data of the target search content in the manner described above. Then, the terminal device can display the target task data and the content association data in the search interface.

[0150] It should be understood that the target task data in this application is not obtained based on superficial text matching, but rather retrieved from the configured task data set through keyword matching and semantic matching based on the target search content and historical search task data. Compared to content-related data obtained through simple text matching, target task data is closer to natural human understanding and better reflects the user's search intent / need. For example, a user's actual search process might be: 1. Searching for "expectations and requirements for students at parent-teacher meetings"; 2. Searching for "how to write simple expectations and requirements for students"; 3. Searching for "how to write feedback on student homework completion"; 4. Searching for "things to emphasize to parents at parent-teacher meetings"; 5. Searching for "what teachers should say at parent-teacher meetings". Based on this user's search process, it can be concluded that the user's true search intent is: to seek advice on how a new teacher should prepare for the key aspects of their first parent-teacher meeting. In essence, search has always addressed the common dimensions of needs underlying the expression of individual search content. However, a significant proportion of search content corresponds to latent search needs. These latent needs are transformed in the user's mind through thought, becoming a series of search terms. The current search content is merely one step in the user's expression of these latent needs. Based on this, this application combines the user's current target search content with task data from the user's historical search tasks to infer and analyze the user's true search intent. This yields target task data that accurately reflects the task requirements of the user's current search. Compared to the various search terms provided by traditional association mechanisms, this application's analysis of task requirements based on the context of the user's search results results in more accurate and reasonable target task data.

[0151] It is worth noting that in a feasible scenario, if the amount of target task data is large, the search interface may not display content-related data, but only the data of each target task.

[0152] As described above, the determination of target task data in this application relies heavily on the current target search content and the task data of historical search tasks. If it only relies on the current target search content, the accuracy of the obtained target task data may not be high enough. Therefore, to save computing resources, after obtaining the target search content and the task data of historical search tasks, this application can first calculate the correlation between the target search content and the task data of historical search tasks. If the correlation is small, it can be considered that the two search tasks executed before and after the search object are unrelated. In this case, the mechanism of pushing target task data to the search object can be turned off (i.e., the target task data is no longer determined or displayed; only the content-related data is displayed). In other words, for the execution of the step of displaying target task data in the search interface, this application can set corresponding conditions. Only when the conditions are met can the step of displaying target task data in the search interface be executed. Specifically, after receiving the target search content, task data of historical search tasks performed by the search object within a historical time period can be obtained; where the search object refers to the object into which the target search content is input. Then, based on the target search content and the task data of historical search tasks, recommendation attribute analysis can be performed on the search object, thereby obtaining the corresponding recommendation attributes of the search object. Specifically, word segmentation processing can be performed on the task data of historical search tasks, thereby obtaining the word set corresponding to the task data of historical search tasks. Here, word segmentation processing can refer to word segmentation processing of the task data of historical search tasks using a hybrid word segmentation method. After obtaining the word set corresponding to the task data of historical search tasks, the historical search task... The core keywords corresponding to the task data of the historical search task refer to words that can reflect the semantics of the task data of the historical search task. Further, based on the core keywords corresponding to the task data of the historical search task, the target search content can be traversed. If the target search content contains the core keywords corresponding to the task data of the historical search task, it can be determined that there is a correlation between the target search content and the task data of the historical search task, and the recommended attribute corresponding to the search object can be determined as a recommendable attribute. If the target search content does not contain the core keywords corresponding to the task data of the historical search task, it can be determined that there is no correlation between the target search content and the task data of the historical search task, and the recommended attribute corresponding to the search object can be determined as a rejectable attribute. When the recommended attribute corresponding to the search object is determined to be a recommendable attribute, the step of displaying the content-related data and the target task data in the search interface can be executed (that is, executing steps S501-S505 above); when the recommended attribute corresponding to the search object is determined to be a rejectable attribute, the step of displaying only the content-related data in the search interface can be executed.

[0153] To facilitate understanding of the method for determining target task data provided in this application, please also refer to... Figure 6 , Figure 6 This is a schematic diagram of a logical architecture for determining target task data based on target search content, provided in an embodiment of this application. For example... Figure 6 As shown, this logical architecture can include at least the following components: content receiving component, historical data acquisition component, data retrieval component, trigger rate estimation component, and data determination component. For ease of understanding, the functions implemented by each component will be briefly described below:

[0154] Content receiving component: The content receiving component can be used to receive the target search content entered by the search object in the search interface.

[0155] Historical data acquisition component: This component can be used to obtain the historical search tasks initiated by the search object within a historical time period, as well as the task data of those historical search tasks. The historical search task can refer to the most recent historical search task initiated by the search object (that is, the search task initiated before the current input of the target search content).

[0156] Data Retrieval Component: The data retrieval component can combine the currently received target search content with the acquired historical search task data, and simultaneously request keyword matching online services and semantic matching online services to retrieve at least one candidate task data from the database. For the retrieved candidate task data, the data retrieval component can also filter it to obtain a filter set.

[0157] Trigger Rate Prediction Component: The trigger rate prediction component can call the probability prediction model to predict the trigger rate of each candidate task data in the filter set, so as to obtain the predicted trigger rate corresponding to each candidate task data.

[0158] Data Determination Component: The data determination component can be used to determine the final target task data based on the estimated trigger rates corresponding to each candidate task data. For example, the data determination component can select the candidate task data corresponding to the highest estimated trigger rate as the target task data.

[0159] For details on the specific implementation of each of the above components, please refer to the preceding text. Figure 5 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0160] It should be noted that while the determination of target task data can be based on the currently input target search content and historical search task data, as described above, this is not the only way to determine target task data. For example, it can also rely on the currently input target search content and the object attribute tags of the search object (tags used to indicate the attributes of the search object, which may refer to the search attributes of the search object). For instance, if the search object frequently searches for content such as fashion tips, then its object attribute tags will include the tag "fashion expert." If the target search content is related to fashion, then the target task data can be determined based on this "fashion expert" tag and the target search content entered by the search object. For example, if the target search content entered by the search object is "colorful scarf," since the object attribute tags of the search object include the tag "fashion expert," then it can be determined that the search object's search intent is very likely to be about matching techniques for colorful scarves. Based on this, the target task data can be determined as "how to match colorful scarves." In summary, this application can infer potential search intent based on the currently input target search content in order to find target task data that reflects the user's true search intent. This application does not impose any restrictions on how to determine the target task data based on the target search content.

[0161] In this embodiment, a new task requirement derivation based on user-input search content is added. This derivation process involves more than just surface text matching; it focuses on semantic analysis. By combining the user's current target search content with task data from the user's historical search tasks, the user's true search intent is deduced and analyzed. This yields target task data that accurately reflects the task requirements the user intends to perform. Compared to the various search terms provided by traditional association mechanisms, this application uses the context of the user's search to perform task requirement derivation analysis, resulting in more accurate and reasonable target task data.

[0162] It should be understood that after configuring the corresponding configuration task data for each configuration search task, this application can further decompose each configuration search task to obtain the decomposition results (which can be referred to as configuration decomposition results) for each configuration search task. The configuration decomposition results for each configuration search task will include at least one configuration decomposition execution step of the configuration search task and subtask data corresponding to each configuration decomposition execution step. Here, a configuration decomposition execution step refers to a step that needs to be executed when executing the configuration search task (which can be understood as a subtask that needs to be executed), and the subtask data refers to text data used to reflect the search intent / requirement of the corresponding configuration decomposition execution step. For example, suppose the configuration task data corresponding to a configuration search task is "Guide to Handling Fever in Children". After breaking down the configuration search task, the configuration breakdown result may include configuration breakdown execution steps 1-3. Configuration breakdown execution step 1 is "Measure the child's temperature with a thermometer", configuration breakdown execution step 2 is "Select medication according to the temperature", and configuration breakdown execution step 3 is "Take the child to the hospital for treatment". The configuration task breakdown result may also include sub-task data corresponding to each configuration breakdown execution step 1-3. For example, the sub-task data for configuration breakdown execution step 1 is "How to correctly measure the child's temperature". This sub-task data can be used to instruct the configuration breakdown execution. Step 1 corresponds to a search task. Users can search for "how to correctly measure a child's temperature" to find the correct way to measure a child's temperature. It's clear that each configuration breakdown execution step has a corresponding execution method / process. However, users may not be aware of the specific execution method / process for each configuration breakdown execution step. Therefore, this application can configure corresponding search data for the execution method / process of each configuration breakdown execution step. This search data is the sub-task data for that step. Users can initiate a search task based on this sub-task data to obtain the specific execution method / process for that step. In other words, the sub-task data for each configuration breakdown execution step can be understood as the text data that needs to be searched to complete that step. For example, for configuration breakdown execution step 2, the specific execution process requires selecting the appropriate medication based on body temperature. Therefore, the sub-task data for configuration breakdown execution step 2 could be "medications suitable for different body temperatures".

[0163] Based on the above, this application can configure different configuration search tasks and configure corresponding configuration task data for each configuration search task. Furthermore, this application can decompose each configuration search task to obtain configuration decomposition results for each configuration search task. After obtaining the configuration task decomposition results for each configuration search task, a mapping relationship can be established between the configuration task data corresponding to that configuration search task and its corresponding configuration decomposition results. Then, the configuration task data corresponding to each configuration search task and its corresponding configuration decomposition results can be stored in the configuration library. In subsequent search operations, after matching the target task data reflecting the search intent / requirement from the configuration task data based on the target search content, the target task data can be displayed. The search object can trigger the target task data. Once the search object triggers the target task data, it can be considered that the search object has performed a search operation and issued the current search task. The terminal device can respond to this trigger and display the answer content of the target task data to the search object based on the configuration decomposition results corresponding to the target task data.

[0164] To better understand the process after the target task data is triggered and a search task is issued, please refer to [link / reference needed]. Figure 7 , Figure 7 This is a schematic flowchart illustrating how to display the answer content corresponding to a search task, as provided in an embodiment of this application. This process can correspond to the above... Figure 3 In the corresponding embodiment, the process after displaying the target task data in the search interface. For example... Figure 7 As shown, the process may include at least the following steps S701-S702:

[0165] Step S701: In response to the target task data being triggered, the result details page is displayed; the result details page includes a rough answer area and a task breakdown area.

[0166] In this application, after the target task data is displayed in the search interface, the search object can trigger the target task data to issue a search task. If the search object triggers the target task data, the terminal device can respond to the trigger and display a result details page for the current search task. This result details page may include a rough answer area and a task breakdown area. The rough answer area can be used to display a rough answer to the current search task. This rough answer may refer to a general answer, which may include the execution steps of each configuration breakdown result corresponding to the target task data (but does not include the specific execution method / process of each configuration breakdown execution step). This rough answer can be used to indicate the execution logic flow of the search object for the current search task, but the specific execution process of a certain step is not included in the rough answer. The task breakdown area can be used to display the sub-task data corresponding to each configuration breakdown execution step in the configuration breakdown result corresponding to the target task data. Each sub-task data can be used as a knowledge point of the current search task. The search object can obtain the specific execution method / process of the corresponding step by triggering any sub-task data.

[0167] Step S702: Display the rough solution content of the target task data in the rough solution area, and display N sub-task data associated with the target task data in the task decomposition area; N is a positive integer, and a sub-task data is used to indicate a search task corresponding to a decomposition execution step. The decomposition execution step refers to the steps that need to be performed when executing the search task indicated by the target task data.

[0168] In this application, based on the above, after displaying the results details page, a rough answer to the target task data can be displayed in the rough answer area, and N sub-task data associated with the target task data can be displayed in the task breakdown area. These N sub-task data are the sub-task data corresponding to the N configuration breakdown execution steps in the configuration task breakdown result corresponding to the target task data in the configuration library. That is to say, the configuration task breakdown result of the target task data will contain multiple configuration breakdown execution steps associated with the target task data. Each configuration breakdown execution step of the target task data can be referred to as the breakdown execution step of the target task data in this application. The configuration breakdown result will also contain sub-task data corresponding to each configuration breakdown execution step. Therefore, there will be multiple sub-task data in the task breakdown result, and these multiple sub-task data can be understood as the sub-task data associated with the target task data. In the task decomposition area, this application can display the data of each subtask associated with the target task data. However, the area of ​​the task decomposition area is limited, and it may not be possible to display all the subtask data associated with the target task data completely. Therefore, some subtask data (i.e., N subtask data) can be displayed in the task decomposition area. It can be seen that N in this application can be less than or equal to the total number of all subtask data associated with the target task data (that is, the total number of configuration decomposition execution steps included in the configuration task decomposition result of the target task data).

[0169] For ease of understanding, please refer to the following: Figure 8 , Figure 8 This is a schematic diagram illustrating a scenario where the answer is displayed after triggering target task data, as provided in an embodiment of this application. Figure 8 As shown, combined with Figure 4 In the scenario shown in the corresponding embodiment, after the terminal device 100a displays the target task data "Preparation of Key Aspects of the Teacher's First Parent-Teacher Meeting" and other search terms (content-related data) in the search interface 4001, user a can trigger any of the displayed content-related data or target task data to perform a search operation, thereby issuing a search task for the triggered content. Of course, user a can also trigger the search control in the search box to issue a search operation for the search content, thereby issuing a search task for the search content. Here, it is assumed that user a... Figure 8The displayed target task data triggered an operation, indicating that user A has executed a search operation and issued a search task. The terminal device can respond to this search operation by generating a result query request for the task breakdown of the target task data and sending this request to server 1000 (which can be the backend server corresponding to the search application). As described above, server 1000 can pre-configure a set of configuration task data, with each configuration task data set associated with a configuration breakdown result. Based on this, server 1000 can retrieve the configuration breakdown result associated with the target task data from the configuration library and return it to terminal device 100a. After receiving the configuration breakdown result, terminal device 100a can display it to user A according to the actual business requirements of the scenario.

[0170] For example, such as Figure 8 As shown, terminal device 100a can first display a result details page 4002. In this result details page 4002, terminal device 100a can display a rough answer area P1 and a task breakdown area P2. In the rough answer area P1, terminal device 100a can display a rough answer for this search task (relatively general and vague, possibly composed of N sub-task data from the configuration task breakdown results of the target task data). If the amount of data in the rough answer is large and the rough answer area P1 cannot display the complete answer, then terminal device 100a can provide a full-text view control in the rough answer area. User A can trigger this full-text view control to view the complete rough answer. In the task breakdown area P2, terminal device 100a can display the sub-task data corresponding to each configuration breakdown execution step of the target task data. Each sub-task data can serve as a knowledge point for this search task, and user A can trigger a specific sub-task data to view the detailed answer for that breakdown execution step. If the number of configuration decomposition and execution steps for the target task data is large, the corresponding number of subtask data will also be large. In this case, the entire amount of subtask data may not be displayed in the task decomposition area P2. Therefore, the subtask data corresponding to the more critical decomposition and execution steps can be displayed first, or the subtask data corresponding to the decomposition and execution steps that are executed earlier in the order can be displayed. For example... Figure 8As shown, the configuration breakdown execution steps of the target task data include: Step 1: Answering the questions asked at the parent-teacher meeting in detail; Step 2: Expressing expectations and requirements to the students; Step 3: Based on the expectations for the students, making requests to the parents to help them achieve those expectations. The sub-task data for Step 1 is Sub-task Data 1: "What questions will parents ask at the parent-teacher meeting?"; the sub-task data for Step 2 is Sub-task Data 2: "How to express expectations and requirements to the students at the parent-teacher meeting?"; and the sub-task data for Step 3 is Sub-task Data 3: "How to accurately express requirements to parents at the parent-teacher meeting?". Then, in the task breakdown area P2, terminal device 100a can display Sub-task Data 1, Sub-task Data 2, and Sub-task Data 3. User a can trigger any one of the sub-task data to obtain the detailed answer content for the corresponding step. This detailed answer content is a specific execution method / process for that step. For example, after user a triggers Sub-task Data 3, terminal device 100a can display a detailed answer page, showing the detailed execution process of Step 3.

[0171] As mentioned above, the amount of data in the rough answer to the target task may be quite large. The rough answer area displayed on the results details page may not be able to fully display the rough answer. Therefore, the terminal device can provide a "View Full Text" control in this rough answer area. This allows the search target (such as user a mentioned above) to trigger the "View Full Text" control to view the complete rough answer. In other words, even when the rough answer displayed in the rough answer area is not complete, after the rough answer to the target task is displayed in the rough answer area, the search target can perform a content viewing operation on the rough answer to view the complete rough answer. After the search target performs the content viewing operation, the terminal device can display the rough answer page, where the complete rough answer to the target task is displayed.

[0172] In this context, the terminal device can display a full-text view control in the rough answer area. The content viewing operation can be a trigger operation on the full-text view control. Of course, the content viewing operation can also refer to the operation of inputting a specified gesture, inputting a specified voice, etc. The specific form of the content viewing operation will not be restricted here.

[0173] It should be understood that, based on the above, after displaying N sub-task data associated with the target task data in the task decomposition area, the search object can trigger any one of the sub-task data to obtain the specific execution process of the corresponding step. The specific implementation process may include, but is not limited to, the following: First, after the search object triggers a certain sub-task data (which can be called the target sub-task data, i.e., any one of the sub-task data displayed in the task decomposition area, i.e., the target sub-task data refers to any one of the N sub-task data), the terminal device can respond to the triggering of the target sub-task data by displaying a decomposition solution interface for that target sub-task data; wherein, this decomposition solution interface is displayed independently. Above the results details page, the disassembly and solution interface can be scaled and transformed. By dynamically adjusting, the disassembly and solution interface can be adjusted to the same size as the results details page (i.e., completely covering the results details page), or it can be adjusted to a smaller area than the results details page (i.e., partially covering the results details page). Then, the terminal device can display a task display area and a detailed solution area in the disassembly and solution interface. The task display area can be used to display the data of each subtask associated with the target task data (that is, all the subtask data contained in the configuration task disassembly results corresponding to the target task data), while the detailed solution area can be used to display the specific execution process of the target subtask data (i.e., the detailed solution content).

[0174] In other words, the terminal device can display the subtask data corresponding to each disassembly execution step associated with the target task data (that is, each configuration disassembly execution step included in the configuration task disassembly result corresponding to the search task indicated by the target task data) in the task display area. Assuming the number of disassembly execution steps is M (a positive integer greater than or equal to N), the total number of subtask data associated with the target task data should also be M. The terminal device needs to display all M subtask data in the task display area. However, due to the limited display area, it may not be possible to display all subtask data completely. Therefore, the terminal device can prioritize displaying the target subtask data and display the other subtask data in the remaining area. The subtask data other than the target subtask data among the M subtask data can be called the remaining subtask data. That is, the terminal device can display the target subtask data and a portion of the remaining subtask data in the task display area. In the detailed solution area, the terminal device can display detailed solutions for the target subtask data. To clearly indicate that the detailed solutions displayed in this area are for the target subtask data, the terminal device can highlight the target subtask data while displaying it in the task display area (e.g., by making the entire target subtask data bold). In other words, the way the target subtask data is displayed in the task display area differs from how the remaining subtask data is displayed.

[0175] For ease of understanding, please refer to the following: Figure 9 , Figure 9 This is a schematic diagram illustrating a scenario where the answer is displayed after triggering target subtask data, as provided in an embodiment of this application. Figure 9 As shown, combined with Figure 8 In the scenario shown in the corresponding embodiment, after terminal device 100a displays the sub-task data associated with the target task data "Preparation of Key Aspects of the Teacher's First Parent-Teacher Meeting" (including sub-task data 1 "What Questions Should Parents Ask at the Parent-Teacher Meeting?", sub-task data 2 "How to Express Expectations and Requirements to Students at the Parent-Teacher Meeting?", and sub-task data 3 "How to Accurately Make Requests to Parents at the Parent-Teacher Meeting") in the task decomposition area P2, user a can trigger any sub-task data to view the specific execution process of the corresponding decomposition execution step. For example, as shown in the example... Figure 9As shown, after user a triggers subtask data 2 "How to express expectations and requirements for students at the parent-teacher meeting", terminal device 100a can respond to this operation and independently display a breakdown and answer interface 4003 on the result details page 4002. This breakdown and answer interface 4003 may include a task display area P3 and a detailed answer area P4. In the task display area P3, terminal device 100a can display all subtask data associated with the target task data. Here, it is assumed that the M subtask data associated with the target task data are subtask data 1 "What questions will parents ask at the parent-teacher meeting", subtask data 2 "How to express expectations and requirements for students at the parent-teacher meeting", and subtask data 3 "How to accurately make requests to parents at the parent-teacher meeting". In the task display area P3, terminal device 100a can prioritize and highlight subtask data 2 (e.g., ...). Figure 9 As shown, subtask data 2 is displayed in bold. Then, terminal device 100a can display the remaining subtask data. Since the display area for this task is limited, it's impossible to display all the remaining subtask data. Therefore, terminal device 100a can display the remaining subtask data from subtask data 1-3 sequentially, following the order of steps 1-3. Specifically, terminal device 100a can first display the remaining subtask data 1, then the remaining subtask data 3. For example... Figure 9 As shown, in the remaining display area of ​​task display area P3, terminal device 100a can display a portion of the content of subtask data 1 "What questions will parents ask at the parent-teacher meeting?" (specifically, it displays a portion of the text "Parent-Teacher Meeting" from subtask data 1 "What questions will parents ask at the parent-teacher meeting?"). Simultaneously, in the detailed answer area P4, terminal device 100a can display the detailed answer content corresponding to the target subtask data 2 (that is, the specific execution method / process of the above-mentioned decomposition and execution step 2). Through this detailed answer content, user a can clearly understand how to specifically decompose and execute step 2.

[0176] For example Figure 9 The disassembly and solution interface 4003 shown can be zoomed in or out by user A. When zoomed in, the task display area containing subtask data is hidden, while the detailed solution area is zoomed in. For example, as shown... Figure 9As shown, after user a performs a zoom-in operation on the disassembly and solution interface 4003, terminal device 100a can respond to this operation by hiding the task display area P3 and zooming in on the detailed solution area P4. As the detailed solution area P4 is zoomed in, more of the detailed solution content of the target subtask data 2 will be displayed. If the zoomed-in detailed solution area P4 still cannot display the complete detailed solution content, terminal device 100a can provide a full-text view control in the detailed solution area P4 for user a to trigger. In this way, after user a triggers the full-text view control, terminal device 100a can display the detailed content interface for the detailed solution content and display the complete detailed solution content corresponding to the target subtask data 2 in the detailed content interface.

[0177] Optionally, based on the above, the task display area may not display all the subtask data associated with the target task data. For example, ... Figure 9 As shown, the task display area P3 only highlights the target subtask data 2 and a portion of the remaining subtask data 1. Therefore, to allow users to view all subtask data, this application provides a function to view all subtask data; users can view each subtask data by swiping. For ease of understanding, please refer to [further details omitted]. Figure 10 , Figure 10 This is a schematic diagram illustrating a scenario for viewing subtask data by sliding, as provided in an embodiment of this application. For example... Figure 10 As shown, combined with Figure 9 In the illustrated embodiment, after displaying the target subtask data 2 and a portion of the remaining subtask data 1 in the task display area P3, user a can view the remaining subtask data by performing a swipe operation, for example, as shown. Figure 10 As shown, user a performed a swipe operation in the task display area P3 (the swipe direction is as shown). Figure 9 As indicated by the dotted arrow, terminal device 100a can respond to this operation by hiding the first half of the target subtask data 2 and displaying the remaining subtask data 1 completely. If user a follows the instructions... Figure 9 As indicated by the dotted arrow, by continuously performing the sliding operation, the terminal device 100a can continuously hide and display the already displayed subtask data while showing the remaining subtask data 3. In this way, user a can gradually view each subtask data. It is worth noting that in some feasible embodiments, when the terminal device displays each subtask data in the task display area, it can automatically scroll through all the subtask data according to the order of the execution steps. This way, the user can view all the subtask data without manually sliding.

[0178] It should be understood that, for the subtask data displayed in the task display area, the search object can select a specific subtask data to view its detailed solution content. Specifically, after the search object triggers the selection of a remaining subtask data, the terminal device can respond to the triggering of the remaining subtask data by switching the highlighted target subtask data in the task display area to highlight that remaining subtask data. At the same time, the terminal device can switch the display of the detailed solution content of the target subtask data in the detailed solution area to display the detailed solution content of the remaining subtask data.

[0179] It is worth noting that, based on the above, when displaying target task data in the search interface, this application can display the target task data by: during the display process, a set identifier is provided for the target task data. This identifier can be used to indicate that the target task data has decomposition and execution steps, as well as corresponding knowledge points (i.e., sub-task data corresponding to each decomposition and execution step). Therefore, when displaying N sub-task data in the decomposition task area, this application can also provide the same identifier for all N sub-task data, thus corresponding to the target task data in the search interface. For example, for a… Figure 8 In the scenario shown, when displaying subtask data 1, subtask data 2, and subtask data 3 in the task decomposition area P2, the title "Knowledge Point" and a set identifier (such as...) can be displayed at the top of the task decomposition area P2. Figure 8 The light bulb icon shown can also be displayed during the process of displaying target task data in the search interface. For example, in... Figure 4 In the embodiment shown, when the target task data "Preparation of key aspects of the teacher's first parent-teacher meeting" is displayed in the search interface 4001, the light bulb icon set can also be displayed in the area where the target task data is displayed.

[0180] In this embodiment, after an object inputs search content, the search / task requirements can be analyzed and processed, and task data reflecting these requirements can be displayed. This presentation of task data not only enriches the relevant information on the interface presented in the search application, but also guides the object to trigger the task data to execute a search task that meets its search needs. This not only enriches the interaction between the object and the search application, but also helps to achieve efficient searching and obtain accurate answers. Furthermore, for the task data reflecting the task requirements, this application configures a rough answer (pre-built), breakdown steps, and corresponding sub-task data. The rough answer provides a concise explanation of the search task corresponding to the task data, while the sub-task data corresponding to the breakdown steps provides the key knowledge points that the user needs to consider. The user can understand the overall process of solving the search task through the rough answer, and further understand the detailed process of executing each step through the sub-task data corresponding to the breakdown steps. This helps the user solve the search task more comprehensively and conveniently.

[0181] As described above, this application will display N subtask data associated with the target task data in the task breakdown area of ​​the results details page. This N subtask data needs to be obtained from the configuration breakdown results of the target task data in the configuration library. For ease of understanding, the specific process of obtaining and displaying the N subtask data of the target task data will be explained below with reference to the accompanying drawings. Please refer to the attached figures. Figure 11 , Figure 11 This is a schematic diagram illustrating a process for acquiring and displaying N subtask data associated with target task data, provided in an embodiment of this application. For example... Figure 11 As shown, the process may include at least the following steps S1101-S1103:

[0182] Step S1101: Obtain the configuration decomposition result corresponding to the target task data from the configuration library. The configuration library contains the correspondence between the configuration task data set and the configuration decomposition result set. There is a correspondence between a configuration task data in the configuration task data set and a configuration decomposition result in the configuration decomposition result set. A configuration task data refers to the task data corresponding to a configuration search task. The configuration decomposition result corresponding to a configuration task data is obtained by decomposing the corresponding configuration search task through the optimized optimization task decomposition model. The configuration decomposition result corresponding to each configuration search task includes at least one configuration decomposition execution step corresponding to the configuration search task, and the subtask data corresponding to each configuration decomposition execution step.

[0183] Specifically, the configuration library of this application may contain a correspondence between a set of configuration task data and a set of configuration decomposition results. Each configuration task data in the configuration task data set may be task data (text data, i.e., the text data to be searched to complete the configuration search task) configured for a configuration search task. Each configuration search task may refer to a pre-configured search task of a relevant operational object (e.g., the operational object of a search application). Each configuration decomposition result in the configuration decomposition result set may refer to the task decomposition result obtained after decomposing the configuration search task. After obtaining the configuration decomposition result of each configuration search task, a correspondence can be established between the configuration decomposition result of a configuration search task and the configuration task data of that configuration search task, and the corresponding data can be stored in the configuration library. Thus, the configuration library can contain the correspondence between each configuration task data and each configuration decomposition result.

[0184] Each configuration search task's configuration decomposition result includes at least one configuration decomposition execution step and corresponding subtask data for each step. The task decomposition of each configuration search task can be implemented by calling an optimized task decomposition model. In this application, the optimized task decomposition model refers to the model obtained after training and optimizing the task decomposition model based on training samples and online experimental data. This task decomposition model can be a large language model with language understanding capabilities. For ease of understanding, the optimization process of the optimized task decomposition model will be described below: First, this application can obtain training samples; then, the task decomposition model can be pre-trained based on these training samples to enable it to perform task decomposition, resulting in a pre-trained task decomposition model. For the pre-trained task decomposition model, this application can further optimize it using the Direct Preference Optimization (DPO) algorithm. Specifically, this application can obtain online experimental data, which is a certain online search task (not offline data), referred to as the first search task. Based on the first search task, the pre-trained model can be tested online. For example, the task data of the first search task (i.e., text data reflecting the task requirements of the first search task) can be obtained. Based on the task data of the first search task, the pre-trained task decomposition model is called to decompose the first search task, thereby obtaining the task decomposition result of the first search task. The task decomposition result of the first search task includes at least one decomposition execution step obtained by decomposing the first search task and the subtask data corresponding to each decomposition execution step. The training samples are generated based on at least one second search task executed by the sample object in the historical time period. The generation method of the training samples can be found in the following related description.

[0185] Furthermore, based on the task decomposition results of the first search task, task correction results for the first search task can be obtained; wherein, the task correction results of the first search task are obtained after correcting the task decomposition results of the first search task according to the task decomposition criteria, and the task correction results of the first search task include at least one labeled execution step obtained by the first search task after correction and the subtask data corresponding to each labeled execution step; it should be understood that in order to improve the task decomposition performance of the task decomposition model and make its decomposed task decomposition results conform to the task decomposition criteria, this application can further optimize the pre-trained task decomposition model through the DPO algorithm. Specifically, this application can modify the first task decomposition result to obtain a task-modified result that better conforms to the task decomposition standard (which can be referred to as the first task-modified result). For example, this application can push the task decomposition result of the first search task to a result modification object (the result modification object can refer to an object used to label each search task). The result modification object can modify at least one decomposition execution step in the first task decomposition result according to the task decomposition standard to obtain at least one labeled execution step. When modifying at least one decomposition execution step in the first task decomposition result, the result modification object can perform operations such as adding, modifying, or deleting at least one decomposition execution step. For example, the result modification object can modify each decomposition... The description of execution steps, the deletion of a specific execution step, or the addition of a new step can all be used to modify the execution steps, creating a new set of execution steps, which can be called labeled execution steps. Then, for at least one labeled execution step, the result modification object can sort these steps according to the task decomposition criteria, thus determining the execution order of each step. Simultaneously, the result modification object can also modify the subtask data of each execution step. For example, it can modify the subtask data based on the labeled execution steps, add subtask data for a newly added labeled execution step, or delete subtask data for a deleted execution step. Ultimately, this process yields at least one labeled execution step, the execution order of these steps, and the subtask data for each labeled execution step. Furthermore, the terminal device can receive at least one annotation execution step, the execution order of at least one annotation execution step, and the subtask data corresponding to each annotation execution step returned by the result correction object. Then, the terminal device can assemble the at least one annotation execution step, the execution order of at least one annotation execution step, and the subtask data corresponding to each annotation execution step (e.g., splicing or aggregating to obtain a set) to obtain the task correction result of the first search task.

[0186] It is worth noting that the number of result modification objects in this application may be one or more, and the task decomposition criteria in this application can be determined through negotiation based on one or more result modification objects. In this way, the task decomposition criteria can better align with the perceptions of different result modification objects. The task decomposition criteria can include guidelines for different items, such as guidelines for the way steps are described, guidelines for the way subtask data is described, guidelines for the logical structure of the decomposition, and guidelines for the order in which different steps are arranged. In general, the task decomposition criteria can be formulated based on the task decomposition preferences of the result modification objects, and the task decomposition criteria can be modified and changed accordingly based on changes in the task decomposition preferences of the result modification objects. That is to say, the task decomposition criteria can be a standard formulated based on the task decomposition logic of the result modification objects, and it conforms to the task decomposition preferences of the result modification objects.

[0187] After obtaining the task decomposition results and task correction results of the first search task, the pre-trained task decomposition model can be optimized based on the differences between them to obtain an optimized task decomposition model. Specifically, the task correction result of the first search task can be designated as a "good" result and used as the training label, while the task decomposition result of the first search task can be designated as a "bad" result and used as the data to be optimized. The pre-trained task decomposition model can then be trained using DPO (Directly Optimized Learning) with both "good" and "bad" results. The core idea of ​​DPO is to directly optimize the model (LM) to conform to human preferences, rather than first fitting a reward model and then using reinforcement learning (RL) for optimization. DPO is essentially a binary classification task; it compares good and bad responses and then adjusts the model to increase the probability of good responses. Its goal is to maximize the probability of generating "good" results while minimizing the probability of generating "bad" results. Based on this, this application employs a binary cross-entropy loss function strategy to optimize the pre-trained task decomposition model.

[0188] Specifically, this application updates model parameters by calculating the difference between the model output (i.e., the task decomposition result of the first search task) and the true label (i.e., the task correction result of the first search task). This process is similar to backpropagation and gradient descent techniques in supervised learning. In each post-training cycle, the corrected data is used to adjust the model parameters. The specific implementation of optimizing the pre-trained task decomposition model based on the difference between the task decomposition result and the task correction result of the first search task can include, but is not limited to: First, using a binary cross-entropy loss function to calculate a first error loss between at least one decomposition execution step in the task decomposition result of the first search task and at least one labeled execution step in the task correction result of the first search task; second, using a binary cross-entropy loss function to calculate a second error loss between the sub-task data corresponding to each decomposition execution step in the task decomposition result of the first search task and the sub-task data corresponding to each labeled execution step in the task correction result of the first search task; based on the first and second error losses, a total loss can be calculated. The model parameters of the pre-trained task decomposition model can be optimized and adjusted according to this total loss, ultimately obtaining the optimized task decomposition model.

[0189] It should be understood that pre-training a task decomposition model using training samples enables it to decompose tasks. However, the decomposition results output by this pre-trained model may have problems (e.g., disordered logical order of execution steps, incomplete descriptions), and may not conform to human perception. To reduce these discrepancies and enhance the logic and standardization of the decomposition results, this solution introduces direct preference optimization techniques. By employing human preference learning, the pre-trained task decomposition model is further optimized. Specifically, the task decomposition results output by the pre-trained model are revised again based on task decomposition standards (determined by human decomposition preferences). The order, logic, and description of the decomposition steps are adjusted to obtain a revised task result. Then, this solution can directly perform preference optimization based on this revised task result, making the decomposition results of the task decomposition model more accurate and reasonable, with a clear sequence and logical consistency.

[0190] It is understandable that the training samples in this application can be determined based on the search tasks performed by the sample objects within a historical time period. Specifically, users who have performed search tasks within a historical time period can be selected as sample objects. The search tasks performed by the sample objects within the historical time period can be identified as historical search tasks or second search tasks. At least one second search task performed by the sample objects within the historical time period, as well as the task data for each second search task, can be obtained. For example, if a user has searched for "how to mix apples and bananas", "home-style recipe for spicy crayfish", and "how to make scallion oil noodles" within a historical time period, then it can be determined that the user has performed three search tasks within the historical time period. Each search task can be identified as a second search task, and the text data "how to mix apples and bananas", "home-style recipe for spicy crayfish", and "how to make scallion oil noodles" can be used as the task data for the corresponding second search tasks.

[0191] Based on at least one second search task, a language understanding model with task decomposition capabilities (e.g., the GPT4 model) can be invoked to decompose the task, thereby obtaining the task decomposition results for each second search task. The task decomposition result for each second search task includes at least one initial decomposition execution step of the second search task output by the language understanding model, as well as the subtask data corresponding to each initial decomposition step. Then, similarly, for each task decomposition result of the second search task, it can be pushed to a result correction object, which can correct the task decomposition result of the second search task according to the task decomposition criteria to obtain the task correction results for each second search task. The task correction results for each second search task can then be used as training labels for the second search task. Based on the second search tasks with training labels, training samples can be formed for training the task decomposition model.

[0192] It is worth noting that, in one feasible embodiment, after obtaining at least one second search task, enrichment analysis processing can be performed on it (enrichment analysis processing in this application can refer to extracting valuable information or features from the task data of at least one second search task through a specific algorithm to facilitate subsequent data processing; here, enrichment analysis processing is to analyze and process the task data of each second search task to obtain the core search task that reflects the overall core search requirements) to obtain the core search task corresponding to at least one second search task. Then, the task core words that reflect the semantics of the core search task can be obtained. Based on the core search task and the task core words of the core search task, training samples can be formed to train the task decomposition model.

[0193] Step S1102: From at least one subtask data contained in the configuration decomposition result corresponding to the target task data, obtain N subtask data as N subtask data associated with the target task data.

[0194] Specifically, after obtaining the configuration breakdown result corresponding to the target task data, N subtask data can be extracted from at least one subtask data contained in the configuration breakdown result. For example, the N subtask data corresponding to the configuration breakdown execution steps with the earliest execution order can be selected sequentially according to the execution order of each configuration breakdown execution step in the configuration breakdown result, and used as the N subtask data associated with the target task data.

[0195] Step S1103: Display the N sub-task data associated with the target task data in the task decomposition area.

[0196] Specifically, after obtaining the data of N subtasks, the data of the N subtasks associated with the target task data can be displayed in the task decomposition area.

[0197] In this embodiment of the application, displaying a rough answer can concisely answer the search task corresponding to the task data. Displaying the sub-task data corresponding to the breakdown steps can provide the knowledge points that the user needs to focus on. The user can understand the overall process of solving the search task through the rough answer, and can further understand the detailed process of executing each step through the sub-task data corresponding to the breakdown steps. In this way, it can help the user solve the search task more comprehensively and conveniently. The task decomposition results can be obtained based on an optimized task decomposition model. This application pre-trains the task decomposition model using training samples, enabling it to decompose tasks. However, the decomposition results output by this pre-trained model may have problems (e.g., disordered logical order between decomposition steps, incomplete descriptions), and do not conform to human perception. To reduce this difference and enhance the logic and standardization of the decomposition results, this solution introduces direct preference optimization technology. By employing human preference learning, the pre-trained task decomposition model is further optimized. Specifically, for the task decomposition results output by the pre-trained model, this solution will revise them again based on task decomposition standards (determined by human decomposition preferences). The order, logic, and description of the decomposition steps will be adjusted to obtain a revised task result. Then, this solution can directly perform preference optimization based on this revised task result, making the decomposition results of the task decomposition model more accurate and reasonable, with a clear order and logical consistency.

[0198] Further, please see Figure 12 , Figure 12This is a schematic diagram of a search task processing device provided in an embodiment of this application. The search task processing device can be a computer program (including program code) running on a computer device; for example, the search task processing device is an application software. The search task processing device can be used to execute... Figure 3 The method shown. (As illustrated) Figure 12 As shown, the processing device 1 for the search task may include: an interface display module 11, a content receiving module 12, and a task display module 13.

[0199] Interface display module 11 is used to display the search interface;

[0200] Content receiving module 12 is used to receive the target search content entered in the search interface;

[0201] The task display module 13 is used to respond to the input of target search content and display content association data and target task data in the search interface. The content association data contains the target search content, and the target task data is used to indicate the task requirements of the search task corresponding to the target search content.

[0202] The specific implementation methods of the interface display module 11, the content receiving module 12, and the task display module 13 can be found in the above description. Figure 3 The descriptions of steps S301-S303 in the corresponding embodiments will not be repeated here.

[0203] In one embodiment, after the task display module 13 displays the content association data and the target task data in the search interface, the task display module 13 is further specifically used for:

[0204] In response to the target task data being triggered, a results details page is displayed; the results details page includes a preliminary answer area and a task breakdown area.

[0205] The rough solution area displays a rough solution to the target task data, and the task breakdown area displays N subtask data associated with the target task data; N is a positive integer. A subtask data is used to indicate a search task corresponding to a breakdown execution step. The breakdown execution step refers to the steps that need to be performed when executing the search task indicated by the target task data.

[0206] In one embodiment, after the task display module 13 displays a rough solution to the target task data in the rough solution area, the task display module 13 is further specifically used for:

[0207] In response to the operation of viewing content for the rough answer to the target task data, the rough answer page is displayed;

[0208] The rough solution page displays the complete rough solution for the target task data.

[0209] In one embodiment, after the task display module 13 displays N sub-task data associated with the target task data in the task decomposition area, the task display module 13 is further specifically used for:

[0210] In response to the triggering of target subtask data, the breakdown and solution interface is displayed; target subtask data refers to any one of the N subtask data, and the breakdown and solution interface is displayed independently on the results details page;

[0211] The task display area and detailed solution area are shown in the disassembly and solution interface;

[0212] The task display area shows the target subtask data, and the detailed solution area shows the detailed solution content for the target subtask data.

[0213] In one embodiment, the number of disassembly and execution steps corresponding to the search task indicated by the target task data is M, and the total number of subtask data associated with the target task data is M, where M is a positive integer greater than or equal to N.

[0214] During the process of displaying the target subtask data in the task display area, the remaining subtask data is also displayed in the task display area. The display method of the target subtask data in the task display area is different from the display method of the remaining subtask data. The remaining subtask data refers to the subtask data other than the target subtask data among the M subtask data associated with the target task data.

[0215] In one embodiment, the target subtask data is displayed in the task display area using a highlighted mode, and the task display module 13 is further specifically used for:

[0216] In response to the triggering of remaining subtask data, the target subtask data highlighted in the task display area is switched to the remaining subtask data.

[0217] Switch the display of detailed solutions for the target subtask data in the detailed solution area to detailed solutions for the remaining subtask data.

[0218] In one embodiment, the specific implementation method of the task display module 13 displaying content-related data and target task data in the search interface includes:

[0219] While displaying related data in the search interface, the target task data is highlighted.

[0220] The display methods for highlighting target task data include at least one of the following:

[0221] Display the target task data at a specified brightness;

[0222] Display target task data in a dynamically flashing manner;

[0223] The identifier set for displaying target task data during the process of displaying target task data.

[0224] In one embodiment, the content-related data displayed in the search interface is obtained by text matching processing of the target search content; the way the target task data is displayed in the search interface is different from the way the content-related data is displayed in the search interface.

[0225] In one embodiment, the specific implementation of the task display module displaying N sub-task data associated with the target task data in the task decomposition area includes:

[0226] Retrieve the configuration decomposition results corresponding to the target task data from the configuration library. The configuration library contains the correspondence between the configuration task data set and the configuration decomposition result set. There is a correspondence between a configuration task data in the configuration task data set and a configuration decomposition result in the configuration decomposition result set. A configuration task data refers to the task data corresponding to a configuration search task. The configuration decomposition result corresponding to a configuration task data is obtained by decomposing the corresponding configuration search task through an optimized task decomposition model. The configuration decomposition result corresponding to each configuration search task includes at least one configuration decomposition execution step corresponding to the configuration search task, and the subtask data corresponding to each configuration decomposition execution step.

[0227] From at least one subtask data contained in the configuration decomposition result corresponding to the target task data, obtain N subtask data as N subtask data associated with the target task data;

[0228] The task breakdown area displays the data of N subtasks associated with the target task data.

[0229] In one embodiment, the optimization process for optimizing the task decomposition model includes:

[0230] Obtain the task data for the first search task;

[0231] Based on the task data of the first search task, a pre-trained task decomposition model is invoked to decompose the first search task, resulting in the task decomposition result of the first search task. The task decomposition result of the first search task includes at least one decomposition execution step obtained by decomposing the first search task and the sub-task data corresponding to each decomposition execution step. The pre-trained task decomposition model is obtained by pre-training the task decomposition model based on training samples. The training samples are generated based on at least one second search task executed by the sample object within a historical time period.

[0232] Based on the task decomposition results of the first search task, the task correction results for the first search task are obtained. The task correction results for the first search task are obtained after correcting the task decomposition results of the first search task according to the task decomposition criteria. The task correction results for the first search task include at least one labeled execution step obtained by the first search task after correction and the subtask data corresponding to each labeled execution step.

[0233] Based on the difference between the task decomposition results of the first search task and the task correction results of the first search task, the pre-trained task decomposition model is optimized to obtain the optimized task decomposition model.

[0234] In one embodiment, the apparatus 1 further includes a step execution module 14.

[0235] Step execution module 14 is used to obtain task data of historical search tasks executed by the search object within a historical time period; the search object refers to the object into which the target search content is input; both the target search content and the task data of historical search tasks are text data;

[0236] The step execution module 14 is also used to perform word segmentation on the task data of historical search tasks to obtain the word set corresponding to the task data of historical search tasks.

[0237] Step execution module 14 is also used to obtain the core words corresponding to the task data of the historical search task from the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task refer to words that can reflect the semantics of the task data of the historical search task.

[0238] The step execution module 14 is also used to execute the step of displaying content-related data and target task data in the search interface if the target search content contains core words corresponding to the task data of historical search tasks.

[0239] For details on the implementation of step execution module 14, please refer to the above. Figure 5 The relevant descriptions in step S505 of the corresponding embodiments will not be repeated here.

[0240] In one embodiment, the specific implementation method of the step execution module 14 displaying content-related data and target task data in the search interface includes:

[0241] Based on the target search content and historical search task data, recall at least one candidate task data from the configuration task data set;

[0242] Based on the task data of historical search tasks, filter at least one candidate task data to obtain a filter set;

[0243] Determine the estimated trigger rate for each candidate task data in the filter set;

[0244] Obtain the maximum estimated trigger rate from the estimated trigger rate set corresponding to the filter set, and determine the candidate task data corresponding to the maximum estimated trigger rate as the target task data;

[0245] Obtain content-related data for the target search content and display the content-related data and target task data in the search interface.

[0246] In one embodiment, the specific implementation of step execution module 14 recalling at least one candidate task data from the configuration task data set based on the target search content and historical search task task data includes:

[0247] Calculate the word matching degree between the target search content and each configuration task data in the configuration task data set according to the keyword matching rules;

[0248] Calculate the semantic matching degree between each configuration task data and the target search content and the task data of historical search tasks according to the semantic matching rules;

[0249] Configuration task data with word matching degree greater than the first matching degree threshold is determined as the first candidate task data, and configuration task data with semantic matching degree greater than the second matching degree threshold is determined as the second candidate task data.

[0250] Both the first candidate task data and the second candidate task data are determined as candidate task data, resulting in at least one candidate task data.

[0251] In one embodiment, each configuration task data in the configuration task data set is text data;

[0252] The specific implementation method of step execution module 14, which calculates the word matching degree between the target search content and each configuration task data in the configuration task data set according to the keyword matching rules, includes:

[0253] Select any one of the configuration task data from the configuration task data set as the target configuration task data;

[0254] The target configuration task data and the target search content are segmented into words according to the first word segmentation method, resulting in the first word set of the target configuration task data under the first word segmentation method and the second word set of the target search content under the first word segmentation method.

[0255] The target configuration task data and the target search content are segmented separately according to the second segmentation method to obtain the third word set of the target configuration task data under the second segmentation method and the fourth word set of the target search content under the second segmentation method; the number of unit characters indicated by the second segmentation method is more than the number of unit characters indicated by the first segmentation method. The number of unit characters refers to the number of characters contained in a unit word.

[0256] Count the number of first words in the first intersection set, the number of second words in the second intersection set, the number of third words in the first word set, and the number of fourth words in the third word set;

[0257] The matching degree calculation function is used to calculate the matching degree of the first word count, the second word count, the third word count, and the fourth word count to obtain the word matching degree between the target configuration task data and the target search content.

[0258] In one embodiment, the specific implementation method of step execution module 14 calculating the semantic matching degree between each configuration task data and the target search content and the task data of historical search tasks according to semantic matching rules includes:

[0259] Select any one of the configuration task data from the configuration task data set as the target configuration task data;

[0260] The target search content, the task data of historical search tasks, and the target configuration task data are transformed into vectors to obtain the first transformation vector corresponding to the target search content, the second transformation vector corresponding to the task data of historical search tasks, and the third transformation vector corresponding to the target configuration task data.

[0261] The first transformation vector and the second transformation vector are fused to obtain the fused vector;

[0262] Calculate the vector similarity between the fusion vector and the third transformation vector to obtain the semantic matching degree between the target configuration task data, the target search content, and the task data of historical search tasks.

[0263] In one embodiment, both the target search content and the task data for historical search tasks are text data;

[0264] Step execution module 14 filters at least one candidate task data based on historical search task data to obtain a filter set, including:

[0265] The task data of historical search tasks is segmented to obtain the word set corresponding to the task data of historical search tasks.

[0266] Extract the core words corresponding to the task data of the historical search task from the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task refer to the words that can reflect the textual semantics of the task data of the historical search task.

[0267] Aggregate the data of at least one candidate task to obtain a candidate set;

[0268] The candidate task data that does not contain the core words corresponding to the task data of historical search tasks are deleted from the candidate set to obtain the filter set.

[0269] In this embodiment, after an object inputs search content, the search / task requirements can be analyzed and processed, and task data reflecting these requirements can be displayed. This presentation of task data not only enriches the relevant information on the interface presented in the search application, but also guides the object to trigger the task data to execute a search task that meets its search needs. This not only enriches the interaction between the object and the search application, but also helps to achieve efficient searching and obtain accurate answers. Furthermore, for the task data reflecting the task requirements, this application configures a rough answer (pre-built), breakdown steps, and corresponding sub-task data. The rough answer provides a concise explanation of the search task corresponding to the task data, while the sub-task data corresponding to the breakdown steps provides the key knowledge points that the user needs to consider. The user can understand the overall process of solving the search task through the rough answer, and further understand the detailed process of executing each step through the sub-task data corresponding to the breakdown steps. This helps the user solve the search task more comprehensively and conveniently.

[0270] Further, please see Figure 13 , Figure 13 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 13As shown, the aforementioned computer device 8000 may include: a processor 8001, a network interface 8004, and a memory 8005. Furthermore, the computer device 8000 also includes: a user interface 8003, and at least one communication bus 8002. The communication bus 8002 is used to enable communication between these components. The user interface 8003 may include a display screen and a keyboard; optionally, the user interface 8003 may also include a standard wired interface or a wireless interface. The network interface 8004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 8005 may be high-speed RAM or non-volatile memory, such as at least one disk storage device. Optionally, the memory 8005 may also be at least one storage device located remotely from the aforementioned processor 8001. Figure 13 As shown, the memory 8005, which is a computer-readable storage medium, may include an operating system, a network communication module, a user interface module, and a device control application program.

[0271] exist Figure 13 In the computer device 8000 shown, the network interface 8004 provides network communication functionality; the user interface 8003 is mainly used to provide an input interface for the user; and the processor 8001 can be used to call the device control application program stored in the memory 8005 to achieve:

[0272] Display the search interface;

[0273] Receive the target search content entered in the search interface;

[0274] In response to the input of target search content, the search interface displays content-related data and target task data. The content-related data contains the target search content, and the target task data indicates the task requirements of the search task corresponding to the target search content.

[0275] It should be understood that the computer device 8000 described in the embodiments of this application can execute the foregoing text. Figures 3 to 11 The description of the processing method for the search task in the corresponding embodiment can also be executed as described above. Figure 12 The description of the processing device 1 for the search task in the corresponding embodiment will not be repeated here. Furthermore, the beneficial effects of using the same method will also not be repeated here.

[0276] Furthermore, it should be noted that this application embodiment also provides a computer-readable storage medium, which stores a computer program executed by the aforementioned data processing computer device 8000. The computer program includes program instructions, and when the processor executes the program instructions, it can execute the aforementioned... Figures 3 to 11 The description of the processing method for the search task in the corresponding embodiments is already provided and will not be repeated here. Furthermore, the beneficial effects of using the same method will also not be repeated. For technical details not disclosed in the computer-readable storage medium embodiments related to this application, please refer to the description of the method embodiments of this application.

[0277] The aforementioned computer-readable storage medium can be the processing apparatus for the search task provided in any of the foregoing embodiments or the internal storage unit of the aforementioned computer device, such as the hard disk or memory of the computer device. The computer-readable storage medium can also be an external storage device of the computer device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the computer device. Furthermore, the computer-readable storage medium can include both internal storage units and external storage devices of the computer device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0278] One aspect of this application provides a computer program product comprising a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium and executes the computer program, causing the computer device to perform the method provided in one aspect of the embodiments of this application.

[0279] The terms "first," "second," etc., in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the term "comprising," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or modules, but may optionally include steps or modules not listed, or may optionally include other step units inherent to these processes, methods, apparatuses, products, or devices.

[0280] In this application embodiment, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0281] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0282] The methods and related apparatuses provided in this application are described with reference to the method flowcharts and / or structural diagrams provided in this application. Specifically, each block of the method flowchart and / or structural diagram, as well as combinations of blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to create a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the process. Figure 1 A schematic diagram of one or more processes and / or structures. Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 A schematic diagram of one or more processes and / or structures. Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 Figure 1 A process or multiple processes and / or structures illustrate the steps of the functions specified in one or more boxes.

[0283] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

Claims

1. A method for processing a search task, characterized in that, The method includes: Display the search interface; Receive the target search content entered in the search interface; In response to the input of the target search content, content association data and target task data are displayed in the search interface. The content association data contains the target search content, and the target task data is used to indicate the task requirements of the search task corresponding to the target search content.

2. The method according to claim 1, characterized in that, After displaying content-related data and target task data in the search interface, the method further includes: In response to the target task data being triggered, a result details page is displayed; the result details page includes a rough answer area and a task breakdown area. The rough solution area displays a rough solution to the target task data, and the task breakdown area displays N subtask data associated with the target task data; N is a positive integer, and each subtask data is used to indicate a search task corresponding to a breakdown execution step. The breakdown execution step refers to the step that needs to be performed when executing the search task indicated by the target task data.

3. The method according to claim 2, characterized in that, After displaying a rough solution to the target task data in the rough solution area, the method further includes: In response to the operation of viewing the content of the rough answer to the target task data, the rough answer page is displayed; The rough solution page displays the complete rough solution content for the target task data.

4. The method according to claim 2, characterized in that, After displaying the N subtask data associated with the target task data in the disassembly task area, the method further includes: In response to the triggering of target subtask data, a breakdown and solution interface is displayed; the target subtask data refers to any one of the N subtask data, and the breakdown and solution interface is displayed independently on the result details page; The disassembly and solution interface displays a task display area and a detailed solution area. The target subtask data is displayed in the task display area, and the detailed solution content for the target subtask data is displayed in the detailed solution area.

5. The method according to claim 4, characterized in that, The number of disassembly and execution steps corresponding to the search task indicated by the target task data is M, and the total number of subtask data associated with the target task data is M, where M is a positive integer greater than or equal to N. During the process of displaying the target subtask data in the task display area, the task display area also displays the remaining subtask data, and the display method of the target subtask data in the task display area is different from the display method of the remaining subtask data in the task display area; the remaining subtask data refers to the subtask data other than the target subtask data among the M subtask data associated with the target subtask data.

6. The method according to claim 5, characterized in that, The target subtask data is displayed in a highlighted manner in the task display area, and the method further includes: In response to the triggering of the remaining subtask data, the target subtask data highlighted in the task display area is switched to highlight the remaining subtask data; The detailed solution content of the target subtask data displayed in the detailed solution area will be switched to the detailed solution content of the remaining subtask data.

7. The method according to claim 1, characterized in that, The display of content-related data and target task data in the search interface includes: While displaying the content-related data in the search interface, the target task data is highlighted. The display method for highlighting the target task data includes at least one of the following: The target task data is displayed at a specified brightness. The target task data is displayed in a dynamically flashing pattern; During the display of the target task data, an identifier is set for the target task data display.

8. The method according to claim 1, characterized in that, The content-related data displayed in the search interface is obtained after text matching processing of the target search content; the display method of the target task data in the search interface is different from the display method of the content-related data in the search interface.

9. The method according to claim 2, characterized in that, The display of N sub-task data associated with the target task data in the disassembly task area includes: The configuration decomposition results corresponding to the target task data are obtained from the configuration library. The configuration library contains the correspondence between the configuration task data set and the configuration decomposition result set. There is a correspondence between a configuration task data in the configuration task data set and a configuration decomposition result in the configuration decomposition result set. A configuration task data refers to the task data corresponding to a configuration search task. The configuration decomposition result corresponding to a configuration task data is obtained by decomposing the corresponding configuration search task through an optimized optimization task decomposition model. The configuration decomposition result corresponding to each configuration search task includes at least one configuration decomposition execution step corresponding to the configuration search task, and subtask data corresponding to each configuration decomposition execution step. From at least one subtask data included in the configuration decomposition result corresponding to the target task data, N subtask data are obtained as N subtask data associated with the target task data; The disassembly task area displays N subtask data associated with the target task data.

10. The method according to claim 9, characterized in that, The optimization process of the optimization task decomposition model includes: Obtain the task data for the first search task; Based on the task data of the first search task, a pre-trained task decomposition model is invoked to decompose the first search task, thereby obtaining the task decomposition result of the first search task. The task decomposition result of the first search task includes at least one decomposition execution step obtained by decomposing the first search task and the sub-task data corresponding to each decomposition execution step. The pre-trained task decomposition model is obtained by pre-training the task decomposition model based on training samples. The training samples are generated based on at least one second search task executed by the sample object within a historical time period. Based on the task decomposition result of the first search task, a task correction result for the first search task is obtained; the task correction result of the first search task is obtained after correcting the task decomposition result of the first search task according to the task decomposition standard, and the task correction result of the first search task includes at least one labeled execution step obtained by the first search task after correction and subtask data corresponding to each labeled execution step. Based on the difference between the task decomposition results of the first search task and the task correction results of the first search task, the pre-trained task decomposition model is optimized to obtain an optimized task decomposition model.

11. The method according to claim 1, characterized in that, The method further includes: Obtain task data of historical search tasks performed on the search object within a historical time period; the search object refers to the object into which the target search content is input; both the target search content and the task data of the historical search tasks are text data. The task data of the historical search task is segmented into words to obtain the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task are obtained from the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task refer to words that can reflect the semantics of the task data of the historical search task. If the target search content contains the core words corresponding to the task data of the historical search task, then the step of displaying the content association data and target task data in the search interface is executed.

12. The method according to claim 11, characterized in that, The display of content-related data and target task data in the search interface includes: Based on the target search content and the task data of the historical search tasks, at least one candidate task data is recalled from the configuration task data set; Based on the task data of the historical search tasks, the at least one candidate task data is filtered to obtain a filter set; Determine the estimated trigger rate for each candidate task data in the filter set; Obtain the maximum estimated trigger rate from the estimated trigger rate set corresponding to the filter set, and determine the candidate task data corresponding to the maximum estimated trigger rate as the target task data; Obtain the content association data of the target search content, and display the content association data and the target task data in the search interface.

13. The method according to claim 12, characterized in that, The step of recalling at least one candidate task data from the configured task data set based on the target search content and the historical search task data includes: According to the keyword matching rules, the word matching degree between the target search content and each configuration task data in the configuration task data set is calculated; According to semantic matching rules, calculate the semantic matching degree between each of the configuration task data and the target search content and the task data of the historical search tasks; Configuration task data with word matching degree greater than the first matching degree threshold is determined as the first candidate task data, and configuration task data with semantic matching degree greater than the second matching degree threshold is determined as the second candidate task data. Both the first candidate task data and the second candidate task data are determined as candidate task data, resulting in at least one candidate task data.

14. The method according to claim 13, characterized in that, Each configuration task data in the configuration task data set is text data; The step of calculating the word matching degree between the target search content and each configuration task data in the configuration task data set according to the keyword matching rules includes: Select any one of the configuration task data from the configuration task data set as the target configuration task data; The target configuration task data and the target search content are segmented into words according to the first word segmentation method to obtain the first word set of the target configuration task data under the first word segmentation method and the second word set of the target search content under the first word segmentation method. The target configuration task data and the target search content are segmented into words according to the second word segmentation method to obtain the third word set of the target configuration task data under the second word segmentation method and the fourth word set of the target search content under the second word segmentation method; the number of unit characters indicated by the second word segmentation method is greater than the number of unit characters indicated by the first word segmentation method, and the number of unit characters refers to the number of characters contained in a unit word; Count the number of first words in the first intersection set, the number of second words in the second intersection set, the number of third words in the first word set, and the number of fourth words in the third word set; A matching degree calculation function is used to calculate the matching degree of the first word count, the second word count, the third word count, and the fourth word count to obtain the word matching degree between the target configuration task data and the target search content.

15. The method according to claim 13, characterized in that, The step of calculating the semantic matching degree between each configuration task data and the target search content and the task data of the historical search tasks according to semantic matching rules includes: Select any one of the configuration task data from the configuration task data set as the target configuration task data; The target search content, the task data of the historical search task, and the target configuration task data are respectively vectorized to obtain a first conversion vector corresponding to the target search content, a second conversion vector corresponding to the task data of the historical search task, and a third conversion vector corresponding to the target configuration task data. The first transformation vector and the second transformation vector are fused to obtain a fused vector. Calculate the vector similarity between the fusion vector and the third transformation vector to obtain the semantic matching degree between the target configuration task data, the target search content, and the task data of the historical search tasks.

16. The method according to claim 12, characterized in that, The filtering process, based on the task data of the historical search task, of the at least one candidate task data yields a filter set, including: The task data of the historical search task is segmented into words to obtain the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task are obtained from the word set corresponding to the task data of the historical search task. The core words corresponding to the task data of the historical search task refer to words that can reflect the textual semantics of the task data of the historical search task. The candidate task data are aggregated to obtain a candidate set; Candidate task data that do not contain the core words corresponding to the task data of the historical search task are deleted from the candidate set to obtain the filter set.

17. A processing apparatus for a search task, characterized in that, The device includes: The interface display module is used to display the search interface; A content receiving module is used to receive the target search content entered in the search interface; The task display module is used to respond to the input of the target search content by displaying content association data and target task data in the search interface. The content association data contains the target search content, and the target task data is used to indicate the task requirements of the search task corresponding to the target search content.

18. A computer device, characterized in that, include: Processor, memory, and network interface; The processor is connected to the memory and the network interface, wherein the network interface is used to provide network communication functions, the memory is used to store computer programs, and the processor is used to call the computer programs to cause the computer device to execute the method according to any one of claims 1-16.

19. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted to be loaded by a processor and to execute the method of any one of claims 1-16.

20. A computer program product, characterized in that, The computer program product includes a computer program stored in a computer-readable storage medium, the computer program being adapted to be read and executed by a processor to cause a computer device having the processor to perform the method of any one of claims 1-16.

Images