Task processing method and device, electronic equipment and storage medium

Abstract

This disclosure proposes a task processing method, apparatus, electronic device, and storage medium, relating to the field of artificial intelligence technology, specifically NLP, computer vision, and intelligent search technology. The solution is as follows: In response to detecting a task processing operation triggered by a logged-in user, a task processing session interface is initialized based on the task context associated with the operation; in response to detecting a knowledge reference addition operation triggered on the task processing session interface, a knowledge association overlay is displayed; in response to a selection operation of at least one knowledge item from the set of knowledge items in the knowledge association overlay, a task processing instruction is sent to the server based on the selected knowledge item and the task context, and the task processing result returned by the server is received. This improves knowledge reference efficiency and content generation quality, effectively shortens the creation path for tasks such as high-value document writing or dialogue generation, and enhances the practicality of AI-assisted creation and the user's operational autonomy.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, specifically to the fields of natural language processing (NLP), computer vision, and intelligent search technology, and particularly to task processing methods, apparatus, electronic devices, and storage media. Background Technology

[0002] With the widespread application of intelligent office solutions and AI-assisted creative writing, users are increasingly relying on large models to complete complex tasks such as document writing, report generation, and requirements gathering. These tasks are usually not performed in isolation, but are closely linked to the user's business context, historical work results, and multi-source knowledge assets within the organization, including meeting minutes, project documents, enterprise knowledge base content, and key information from collaboration platforms.

[0003] To enhance the relevance and usability of generated content, the task processing has gradually evolved into a collaborative creation model that integrates external knowledge. In this model, knowledge citations are no longer merely supplementary explanations of the results, but rather an essential component of the task input, participating in the entire processing flow. Therefore, how to conduct task processing is crucial. Summary of the Invention

[0004] This disclosure provides a task processing method, apparatus, electronic device, and storage medium.

[0005] According to one aspect of this disclosure, a task processing method is provided, the method comprising: in response to detecting a task processing operation triggered by a logged-in user, initializing a task processing session interface based on a task context associated with the operation; in response to detecting a knowledge reference addition operation triggered on the task processing session interface, displaying a knowledge association overlay; wherein the knowledge association overlay displays a set of knowledge entries including at least recommended knowledge entries associated with the task context; in response to a selection operation on at least one knowledge entry in the set of knowledge entries, sending a task processing instruction to a server based on the selected knowledge entry and the task context, and receiving a task processing result returned by the server.

[0006] According to another aspect of this disclosure, a task processing apparatus is provided, the apparatus comprising: a detection module, configured to initialize a task processing session interface based on a task context associated with the operation in response to detecting a task processing operation triggered by a logged-in user; a display module, configured to display a knowledge association overlay in response to detecting a knowledge reference addition operation triggered on the task processing session interface; wherein the knowledge association overlay displays a set of knowledge entries including at least recommended knowledge entries associated with the task context; and a selection module, configured to send a task processing instruction to a server based on the selected knowledge entry and the task context in response to a selection operation on at least one knowledge entry in the set of knowledge entries, and receive a task processing result returned by the server.

[0007] According to another aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the task processing method described in the first aspect embodiment of this disclosure.

[0008] According to another aspect of this disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to perform the task processing method described in the first aspect embodiment of this disclosure.

[0009] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the task processing method described in the first aspect of this disclosure.

[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0011] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0012] Figure 1 A flowchart illustrating the process of generating documents using AI in related technologies; Figure 2 This is a flowchart illustrating the task processing method provided in Embodiment 1 of this disclosure; Figure 3 A schematic diagram of the task initiation interface provided in an embodiment of this disclosure; Figure 4 A schematic diagram of the interface for adding the trigger knowledge reference provided in the embodiments of this disclosure; Figure 5 This is a flowchart illustrating the task processing method provided in Embodiment 2 of this disclosure; Figure 6 This is a schematic diagram of the interface for recommending knowledge entries in the knowledge association overlay provided in an embodiment of this disclosure; Figure 7 This is a schematic diagram of the interface for platform knowledge entries in the knowledge association overlay provided in an embodiment of this disclosure; Figure 8 This is a schematic diagram of the interface for searching knowledge entries in the knowledge association overlay provided in an embodiment of this disclosure; Figure 9 This is a flowchart illustrating the task processing method provided in Embodiment 3 of this disclosure; Figure 10 This is a flowchart illustrating a task processing method provided in an embodiment of the present disclosure; Figure 11 A schematic diagram of the overall architecture of the intelligent knowledge recommendation and citation system provided in the embodiments of this disclosure; Figure 12 This is a schematic diagram of the structure of the task processing device provided in Embodiment 4 of this disclosure; Figure 13 This is a block diagram of an electronic device used to implement the request processing method of the embodiments of this disclosure. Detailed Implementation

[0013] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0014] With the rise of Artificial Intelligence (AI), automated document creation is becoming increasingly common. However, employee knowledge creation often requires integrating outputs from other stages of the workflow. For example, sales / user research / operations might combine backend data to generate data analysis reports; ordinary employees might break down their own OKRs based on their boss's / partner's Objectives and Key Results (OKRs); or product managers (PMs) might generate requirement documents based on meeting minutes from multiple requirement discussions. Therefore, such as... Figure 1As shown, current document AI has two main problems during generation: 1. It only supports adding general materials such as web page links and local files / images, and cannot quickly obtain relevant data / information in the workflow; 2. In knowledge base scenarios, when users add / process documents in batches, they cannot quickly select multiple documents across or within the knowledge base, and can only manually copy the source links of relevant knowledge one by one.

[0015] Therefore, in response to any of the above problems, this disclosure provides a task processing method, apparatus, electronic device, and storage medium.

[0016] The task processing method, apparatus, electronic device, and storage medium of this disclosure are described below with reference to the accompanying drawings.

[0017] Figure 2 This is a flowchart illustrating the task processing method provided in Embodiment 1 of this disclosure.

[0018] This disclosure describes an embodiment where a task processing method is configured in a task processing device, which can be deployed on a client so that the client can perform task processing functions.

[0019] like Figure 2 As shown, the task processing method may include the following steps: Step 201: In response to detecting a task processing operation triggered by a logged-in user, initialize the task processing session interface based on the task context associated with the operation.

[0020] As one possible implementation, when a logged-in user is detected to have initiated a task processing operation, for example, ... Figure 3 As shown, clicking on entries such as "New Document," "Start Writing," "Generate Report," or entering dialogue mode will cause the client to dynamically construct and load an interactive interface, called the task processing session interface, based on the specific context information of the operation, i.e., the task context. This interface serves as the basic carrier for subsequent task execution and knowledge interaction, providing users with a structured input and operation environment. The task context may include, for example, the currently opened document type, user identity, operation time, and associated projects.

[0021] Step 202: In response to the detection of a knowledge reference addition operation triggered on the task processing session interface, display the knowledge association overlay.

[0022] The knowledge association overlay displays a set of knowledge entries, including at least recommended knowledge entries associated with the task context.

[0023] As one possible implementation, if the user performs a knowledge reference addition operation while the task processing session interface is active, for example, ... Figure 4As shown, clicking the "Add Knowledge Reference" control on the interface will immediately pop up a knowledge association overlay displayed on the current screen. This overlay is used to centrally present a set of knowledge items for reference, which includes at least recommended knowledge items intelligently filtered by the system based on the current task context, making it easy for users to quickly obtain information resources related to their current task.

[0024] It should be noted that the knowledge association overlay displays a set of knowledge entries, including at least those related to the task context. These recommended knowledge entries can include those related to the logged-in user's role, historical behavior, current project, functional permissions, and the business scenario they are in. For example, if the user is a Quality Assurance (QA) user, test cases or defect analysis reports from the test management platform might be recommended; if the user is a salesperson, customer business tags related to the current task within their authorized access scope might be recommended.

[0025] Step 203: In response to the selection operation of at least one knowledge item in the knowledge item set, based on the selected knowledge item and the task context, send a task processing instruction to the server and receive the task processing result returned by the server.

[0026] Furthermore, when a user performs a selection operation on a set of knowledge items in the knowledge association overlay—such as checking one or more recommended knowledge items, manually adding knowledge items, or searching for knowledge items—the client captures this selection behavior, integrates the selected knowledge items with the current task context, constructs a structured task processing instruction, and sends it to the server. Upon receiving the instruction, the server executes the corresponding processing logic based on the knowledge content and context information it contains, such as calling a large model to generate text, updating document content, or providing a dialogue response, and returns the generated task processing result to the client. The client then receives the result and displays it in the task processing session interface for the user to view, edit, or perform further operations.

[0027] In summary, the task processing session interface is dynamically initialized based on the task context associated with the user's triggered action, thus gaining a full understanding of the user's intent and the surrounding scenario at the task's inception. Building upon this, when the user triggers a knowledge reference addition operation, the displayed knowledge association overlay accurately aggregates knowledge items highly relevant to the current task, effectively avoiding manual searching and interface switching between multiple work platforms. Furthermore, by selecting a set of knowledge items, the user actively injects referenced knowledge into the task processing flow, enabling the server to construct content based on the referenced knowledge items when generating task processing results. This improves knowledge reference efficiency and content generation quality, effectively shortens the creation path for tasks such as high-value document writing or dialogue generation, and enhances the practicality of AI-assisted creation and the user's operational autonomy.

[0028] To clearly illustrate how the above embodiments display a knowledge association overlay in response to the detection of a knowledge reference addition operation triggered on the task processing session interface, this disclosure proposes another task processing method.

[0029] Figure 5 This is a flowchart illustrating the task processing method provided in Embodiment 2 of this disclosure.

[0030] like Figure 5 As shown, the task processing method may include the following steps: Step 501: In response to detecting a task processing operation triggered by a logged-in user, initialize the task processing session interface based on the task context associated with the operation.

[0031] Step 502: In response to detecting a knowledge reference addition operation triggered on the task processing session interface, a recommended knowledge retrieval request is sent to the server, and recommended knowledge entries associated with the task context are returned by the server, and the recommended knowledge entries are displayed in the knowledge association overlay; and / or, in response to detecting an interaction operation of a knowledge source identifier triggered on the knowledge association overlay, a knowledge source query request is sent to the server, and platform knowledge entries matching the knowledge source identifier are returned by the server, and the platform knowledge entries are displayed in the knowledge association overlay; and / or, in response to detecting an input operation of a search keyword triggered on the knowledge association overlay, a search request is sent to the server, and search knowledge entries matching the search keyword are returned by the server, and the search knowledge entries are displayed in the knowledge association overlay.

[0032] In one implementation, within the task processing session interface, when a user triggers a knowledge reference addition operation (e.g., clicking the "Add Knowledge Reference" control), the client sends a recommended knowledge retrieval request to the server. This request carries the current task context information. The server intelligently generates and returns associated recommended knowledge entries based on this context. Figure 6As shown, the client then displays these recommended contents in the knowledge association overlay and preloads the interface frameworks for the "Manual Association" and "Search" tabs. Furthermore, the recommended knowledge entries can be preloaded when the task processing session interface is initialized, so that they can be presented immediately when the user first triggers the knowledge reference addition operation, reducing waiting time and improving the smoothness of interaction.

[0033] The recommended knowledge items are generated by the server using the following steps: (1) Based on the user identifier of the logged-in user in the task context, obtain the logged-in user's historical interaction behavior logs, schedule data and functional role tags; As one possible implementation, when a task processing session starts, the system extracts the unique user identifier of the logged-in user from the current task context and, based on this identifier, synchronously retrieves key information such as historical interaction logs, schedule data, and job role tags from multiple data sources within the enterprise. The historical interaction logs record the user's operational trajectory regarding knowledge content across various work platforms, including browsing, editing, referencing, or saving. The schedule data comes from the enterprise's unified calendar system, reflecting the meetings and work events the user has recently participated in or scheduled. The job role tags are maintained by the organizational structure management system and are used to identify the user's job type or business function. These three types of information together constitute a multi-dimensional contextual profile for individual users.

[0034] (2) Based on historical interaction behavior logs, recall at least one knowledge item that the logged-in user performed a specified interaction behavior within a set time period, and sort at least one knowledge item according to the relevance score and timeliness weight of at least one knowledge item to obtain the first recommended knowledge item. Next, knowledge items with high intent signals that users have historically performed within a preset time frame are selected from historical interaction behavior logs. These include items viewed multiple times, actively referenced, or marked as important. Then, a comprehensive ranking score is calculated for each candidate item. This score consists of two parts: a relevance score, which measures the semantic or thematic match between the item and the current task context; and a timeliness weight, which attenuates earlier actions to ensure more recent operations have higher priority. Finally, the candidate items are ranked based on the comprehensive score, and the top-ranked items are selected as the first recommended knowledge item.

[0035] (3) Based on the schedule data, obtain the meeting minutes and related documents of the logged-in user within the set time window, and generate a second recommended knowledge item based on the meeting minutes and related documents; One possible implementation involves parsing users' schedule data to identify meetings that logged-in users have participated in or are about to participate in within a recent time window, and automatically retrieving corresponding meeting minutes and related documents shared or mentioned during the meetings from the meeting collaboration platform. These meeting minutes and / or related documents are then structured and encapsulated into directly referenceable knowledge units, forming the second set of recommended knowledge items.

[0036] (4) Based on the functional role tags of the triggering user, query the work platforms associated with the functional role tags, and generate third recommended knowledge items according to the platform entry of the work platform; One possible implementation is to match pre-configured role-to-work platform mappings based on the user's job function tags to identify frequently used professional systems in the user's daily work. For example, quality assurance personnel might correspond to a test management platform, while sales personnel might correspond to a customer relationship management or data analysis platform. Subsequently, the system generates guiding entries pointing to typical functions or content entry points on these platforms, such as viewing test cases for a specific module or accessing a certain type of customer analysis report, and these entries are designated as third-party recommended knowledge items.

[0037] (5) Determine the recommended knowledge items based on at least one of the first recommended knowledge items, the second recommended knowledge items, and the third recommended knowledge items.

[0038] One possible approach is to integrate recommendation results from user behavior preferences, schedule context, and job roles. Integration strategies may include deduplication and merging, dynamic weight allocation, or priority ranking, ultimately forming a unified set of recommended knowledge items. Even if a certain path cannot generate results due to missing data, other paths can still independently provide effective recommendations, ensuring the continuous availability of the recommendation service.

[0039] Therefore, by leveraging the user's login identifier, the system can simultaneously obtain their historical interaction logs, schedule data, and job role tags, thus constructing a comprehensive user profile encompassing behavioral preferences, temporal context, and job responsibilities. Based on this, it generates first, second, and third recommended knowledge items from three dimensions: personal usage habits, recent collaborative activities, and role-specific professional needs. At least one of these is then integrated to form the final recommendation result, ensuring that the recommended content not only fits the current task context but also deeply integrates into the user's actual workflow and organizational role positioning. This improves the relevance and timeliness of knowledge discovery, avoiding the information generalization or lag issues common in general recommendations. Simultaneously, the multi-path recall strategy ensures effective references even when some data is missing, enhancing the system's robustness and applicability. Ultimately, users can instantly obtain personalized, structured, and actionable knowledge support within the task interface without switching between multiple systems or relying on memory recall, improving the efficiency and quality of content creation.

[0040] To enhance the understandability, operability, and user decision-making efficiency of knowledge recommendations, the system presents recommended knowledge items from different sources in a structured manner within a knowledge association overlay. When the system determines that the final recommended knowledge item includes a first recommended knowledge item generated from the user's historical interaction logs, it displays the first recommended knowledge item in the first sub-section of the knowledge association overlay. If the recommendation result includes a second recommended knowledge item generated based on schedule data, it displays the second recommended knowledge item in the second sub-section. If it includes a third recommended knowledge item generated based on job / role tags, it displays the third recommended knowledge item in the third sub-section. Each sub-section is independent in its interface layout and semantically clear; for example, they correspond to three knowledge scenarios: "Personally Frequently Used," "Recent Collaborations," and "Role-Specific." This partitioning design not only preserves the traceability of the sources of each recommendation but also enables users to quickly identify the contextual basis of each type of knowledge, thereby more efficiently judging its applicability. Simultaneously, the partitioned display supports on-demand browsing and selection, avoiding cognitive interference caused by the mixing of different types of knowledge, and improving the user's information processing efficiency and operational experience with the knowledge association overlay.

[0041] Furthermore, each sub-section can be configured with an independent title, icon, and interactive guidance. For example, the first sub-section could be labeled "Content You Frequently Reference," the second sub-section "From Recent Meetings," and the third sub-section "Relevant to Your Role." In addition, the sub-section structure reserves interfaces for future expansion. For instance, when adding recommended sources such as "Team Sharing" or "Project Documents," only the corresponding sub-section needs to be added; there is no need to refactor the overall interactive framework, thereby improving the system's maintainability and adaptability.

[0042] In another implementation, if a logged-in user enters or selects a knowledge source identifier in the knowledge association overlay, such as manually entering or pasting the URL of the work platform, or selecting platform icons or labels such as "iTest" or "Data Backend" from the preset platform list, the client sends a corresponding knowledge source query request to the server; the server then retrieves matching platform knowledge entries from the specified platform and returns them, such as... Figure 7 As shown, the client displays the returned entries in the knowledge association overlay.

[0043] The platform knowledge entries are generated by the server using the following steps: (1) Determine the target knowledge platform corresponding to the knowledge source identifier; When a user selects or enters a knowledge source identifier (such as a platform name, URL, or preset tag) in the knowledge association overlay, the system first parses the semantic or structural features of the identifier and matches it with a pre-configured knowledge platform registry. This registry maintains the mapping relationship between each knowledge source identifier and its corresponding target knowledge platform, including metadata such as platform type, access protocol, authentication method, and interface specifications. Through this matching process, the system accurately identifies the specific target knowledge platform the user intends to access, such as an internal enterprise test management platform, customer data dashboard, or document collaboration system.

[0044] (2) Call the adapter that interfaces with the target knowledge platform and pull the list of content associated with the knowledge source identifier through the content interface; Then, the system calls a dedicated adapter pre-developed and deployed for the platform. This adapter encapsulates the authentication logic, request format, pagination strategy, and error handling mechanism required for communication with the target platform. It can call open content interfaces in a standardized manner. Through these interfaces, the adapter sends query requests to the target platform to obtain a list of content directly associated with the knowledge source identifier selected by the user. For example, a set of test cases under a project, a list of analysis reports for a certain customer dimension, or the latest policy documents in a document library.

[0045] (3) Convert the format of the content list to obtain platform knowledge entries.

[0046] It's important to understand that because the field naming and organization methods of the content list returned from the target knowledge platform may differ, the system performs format conversion processing on the content list to ensure consistent rendering and interaction of the knowledge association overlay. This processing may include field mapping (such as mapping the platform-specific "doc_id" to the generic "knowledge_id"), content summary generation, metadata extraction (such as creation time, author, and permission scope), and desensitization or aggregation operations when necessary. The final output is structured data that conforms to the system's internal knowledge item standard model, i.e., platform knowledge items, which can be directly used for overlay display, user selection, and subsequent injection task processes.

[0047] In summary, the system accurately identifies the target knowledge platform based on the user-specified knowledge source identifier, ensuring the directionality and legitimacy of data retrieval. Then, by calling the platform's content interface through a pre-built adapter, it reliably obtains the content list within the authorized scope while shielding differences in underlying protocols, authentication mechanisms, and data formats. Finally, by standardizing the format of the original content, multi-source heterogeneous data is unified into platform knowledge entries that are compatible within the system. This avoids the need to repeatedly develop dedicated access logic for each knowledge platform, ensuring consistency and usability of knowledge entries during display, selection, and subsequent citation. Furthermore, this mechanism achieves on-demand aggregation and secure reuse of dispersed knowledge assets within the enterprise without altering the original architecture of each knowledge platform. It effectively supports seamless access and accurate citation of content from multiple systems within a single task interface, thereby significantly improving knowledge acquisition efficiency and the practicality of AI-assisted creation.

[0048] Furthermore, it's worth noting that within the knowledge association overlay, the system can dynamically create a dedicated tab for each identified target knowledge platform. This tab is identified by the platform's name, icon, or business semantics, such as "iTest Test Cases" or "Sales Data Dashboard," thus intuitively reflecting the content's source. Subsequently, the system loads and displays the platform's knowledge entries, after format conversion, within this tab, presenting them as lists, cards, or other adapted interface formats, allowing users to browse, search, or select entries for citation. Thus, by assigning an independent tab to each target knowledge platform, the system achieves isolated and modular display of content from multiple platforms. This avoids the mixing of knowledge entries from different platforms and allows users to switch between viewing content from specific platforms as needed, significantly improving the clarity of information organization and the directionality of operations. Simultaneously, the tab mechanism provides a flexible interface container for subsequent expansion with new knowledge platforms, supporting the integration of more external systems without refactoring the overall overlay structure.

[0049] In yet another implementation, such as Figure 8 As shown, if a user enters a search keyword in the overlay, the client will encapsulate the keyword into a search request and send it to the server. The server will return search knowledge entries that match the keyword, which will also be presented in the knowledge association overlay.

[0050] The search knowledge entries are generated by the server using the following steps: (1) Generate a set of search terms based on the search keywords; wherein, the set of search terms includes: the search keywords themselves, each segmentation unit obtained after segmenting the search keywords, and the synonyms obtained by synonym expansion of each segmentation unit; As one possible implementation, when a user enters a search keyword in the knowledge association overlay, the system first retains the keyword as a whole in the search term set to preserve its original semantic integrity. Then, the natural language processing module is invoked to segment the keyword, breaking it down into multiple semantic or syntactic basic units. For example, "customer satisfaction analysis" is split into three segmentation units: "customer," "satisfaction," and "analysis." Furthermore, the system utilizes a pre-built thesaurus or semantic embedding model to perform synonym expansion on each segmentation unit, generating words with similar meanings in the business or technical context. For example, "satisfaction" is expanded to "satisfaction level," "user feedback," or "NPS." Finally, the original keyword, all segmentation units, and their corresponding synonyms together constitute an enhanced search term set.

[0051] (2) Based on the set of search terms, perform full-text search concurrently in each of the knowledge sources that have been accessed, so as to obtain the search results returned by each knowledge source; Furthermore, the set of search terms is used as the query input, and full-text search requests are initiated in parallel to all connected knowledge sources (such as document libraries, test management platforms, customer data systems, meeting record libraries, etc.). Each knowledge source independently performs matching operations within its local index scope, supporting multiple search strategies such as Boolean logic, fuzzy matching, or vector similarity to identify content entries containing any search term. Due to the concurrent execution mechanism, the retrieval processes of each knowledge source are independent and synchronous, which can shorten the overall response time. After each knowledge source completes its retrieval, it returns the search results that meet the matching conditions. These results can then be uniformly aggregated, deduplicated, sorted, and displayed in the knowledge association overlay. In addition, this concurrent retrieval mechanism not only improves search efficiency but also ensures the breadth and real-time coverage of multi-source knowledge, enabling users to access relevant information scattered across different systems in a single operation.

[0052] (3) Generate search knowledge entries based on the search results returned by each knowledge source.

[0053] As one possible approach, the search results from different knowledge sources are standardized and parsed to extract key fields for each result, such as title, abstract, source platform, creation time, unique identifier, and access permission information. Then, results with inaccessible or empty content are removed according to preset filtering rules, and the remaining entries are deduplicated to prevent the same knowledge content from being displayed repeatedly due to synchronization across multiple platforms. Based on this, the system can further reorder the entries based on relevance, for example, by calculating a comprehensive score combining search term matching, content timeliness, and user role preferences. Finally, the processed results are encapsulated into structured objects that conform to the internal knowledge entry data model, i.e., the search knowledge entries.

[0054] In summary, by generating a set of search terms based on the original search keywords, including the keywords themselves, word segmentation units, and synonym extensions, the system effectively covers both explicit expressions and potential semantic variations of user queries, mitigating the problem of missed detections caused by different expression habits. Furthermore, this enhanced set of search terms is used to concurrently perform full-text searches across all connected knowledge sources, fully leveraging the complementarity of multi-source heterogeneous data and significantly reducing response latency through parallel mechanisms. Finally, the original results returned by each knowledge source are uniformly parsed, filtered, and structured into standardized search knowledge entries, ensuring consistency and operability for subsequent display and citation. As a result, users can access relevant knowledge scattered across various platforms such as document libraries, task systems, meeting minutes, and code repositories with a single search action. This avoids cognitive blind spots caused by information silos and provides rich, accurate, and traceable factual evidence for AI generation, significantly improving the authenticity, efficiency, and business relevance of content creation.

[0055] To achieve clear organization and efficient browsing of multi-source search results, after obtaining search knowledge entries, the system categorizes them based on the knowledge source identifiers carried by each entry. All entries belonging to the same knowledge source are aggregated into a single category result, forming a grouped set corresponding one-to-one with each accessed knowledge source. Furthermore, for each category result, the system dynamically creates a dedicated tab within the knowledge association overlay, displaying all knowledge entries contained in that category on the corresponding tab. The tab title directly uses the platform name corresponding to the knowledge source, such as "iTest" or "Data Dashboard," to intuitively reflect the content source. This allows users to clearly identify the information boundaries returned by different platforms and switch to view search results from specific systems as needed, effectively avoiding information interference caused by heterogeneous content and significantly improving the understandability, operability, and overall interaction efficiency of cross-platform search results.

[0056] To enhance users' ability to quickly identify relevant content in search results, as a possible approach, the system performs refined processing on the knowledge entries in each category result. For text fragments in the entry's body or summary that match any search term in the search term set, the system automatically applies highlighting marks, such as through background coloring, bolding, or underlining. It should be noted that this highlighting processing is based on precise or fuzzy matching strategies, which can cover different forms of semantic expression, ensuring that even if the keywords entered by the user are not completely consistent with the wording of the original text, their core semantic connections can still be effectively identified.

[0057] Based on any or more of the above implementation methods, a set of knowledge entries consisting of recommended knowledge entries, platform knowledge entries, and / or searchable knowledge entries can be generated and displayed in the knowledge association overlay. This improves the comprehensiveness and accuracy of knowledge discovery and avoids the operational burden of users repeatedly switching between multiple heterogeneous systems. At the same time, since all knowledge entries are displayed according to their source and retain the original source and citation mark, users can quickly assess the relevance and credibility of the content, enhance their trust in the AI-assisted generation process, and thus improve the overall creation efficiency.

[0058] Step 503: In response to the selection operation of at least one knowledge item in the knowledge item set, based on the selected knowledge item and the task context, send a task processing instruction to the server and receive the task processing result returned by the server.

[0059] It should be noted that the execution process of steps 501 and 503 can be implemented in any of the embodiments of this disclosure. This disclosure does not limit this and will not elaborate further.

[0060] In summary, the task context-driven recommendation mechanism provides highly relevant recommended knowledge items when users trigger the knowledge citation addition operation, improving the accuracy and timeliness of knowledge discovery and avoiding users blindly sifting through massive amounts of information. Simultaneously, by supporting interactive operations on knowledge source identifiers, users can access specific work platforms (such as iTest, data backends, etc.) on demand, retrieving structured knowledge content within their authorized scope, achieving precise cross-platform knowledge retrieval. Furthermore, combined with user-inputted search keywords, the system can dynamically respond to personalized search needs, supplementing recommendations and information not covered by the platform, enabling the knowledge association overlay to intelligently push information while retaining sufficient user control. Thus, users can complete the entire knowledge discovery and citation process—from passive recommendations and targeted platform access to active search—without leaving the current task interface, effectively avoiding problems such as switching between multiple systems, repeated queries, and context fragmentation. This significantly shortens the creation path of high-value content and enhances the practicality, flexibility, and reliability of the AI-assisted system.

[0061] To clearly illustrate how the above embodiments perform task processing based on the selected knowledge items and task context, send task processing instructions to the server, and receive task processing results returned by the server, this disclosure proposes another task processing method.

[0062] Figure 9 This is a flowchart illustrating the task processing method provided in Embodiment 3 of this disclosure.

[0063] like Figure 9 As shown, the task processing method may include the following steps: Step 901: In response to detecting a task processing operation triggered by a logged-in user, initialize the task processing session interface based on the task context associated with the operation.

[0064] Step 902: In response to the detection of a knowledge reference addition operation triggered on the task processing session interface, display the knowledge association overlay.

[0065] The knowledge association overlay displays a set of knowledge entries, including at least recommended knowledge entries associated with the task context.

[0066] Step 903: Construct input prompts based on the selected knowledge items and task context.

[0067] One possible implementation is to semantically fuse the selected knowledge items with the current task context to generate structured input prompts. It's important to note that these prompts are organized in natural language instructions or templated formats to guide the large model on how to utilize the selected knowledge to complete a specific task. Examples include: "Refer to the following test cases and requirements documents to generate an acceptance criteria description" or "Combining customer profiles and meeting minutes, draft a follow-up email." During this process, the system can perform summary extraction, key information extraction, or format standardization on the knowledge items to ensure that the prompts are concise, focused, and meet the model's input requirements.

[0068] Step 904: Send a task processing instruction containing input prompts to the server and receive the task processing result returned by the server.

[0069] The constructed input prompts are then encapsulated into task processing instructions and sent to the server's large-scale model inference engine. Upon receiving the instructions, the server invokes the corresponding AI model for inference processing, generating structured or natural language task processing results (such as text drafts, code snippets, and analysis conclusions) based on the task objectives and knowledge base in the prompts, and returns these results to the client. The client can then directly insert the results into the current task interface, display them in the preview area, or allow the user to further edit and confirm them. It's worth noting that the task processing results may include generated text content (such as requirement descriptions, email drafts, test steps, or analysis conclusions) and embedded knowledge citation identifiers corresponding to the referenced knowledge entries, such as unique knowledge IDs, source platform names, original document titles, and access links. These citation identifiers can be automatically rendered by the system as clickable annotations, footnotes, or floating tips, ensuring that every key piece of information in the generated content can be traced back to its original knowledge source.

[0070] Furthermore, it should be noted that task processing instructions are categorized into three types based on the user's current operational intent and task scenario: document creation instructions, document update instructions, and dialogue interaction instructions, to support diverse AI-assisted tasks. Document creation instructions instruct the server to generate a completely new document based on the user's selected knowledge items; for example, automatically generating a test plan document based on selected test cases and requirements specifications. Document update instructions instruct the server to revise or supplement parts of an existing target document; for example, automatically inserting key points from the latest meeting minutes into the corresponding chapter of a project plan document to ensure the document content remains synchronized with the latest knowledge. Dialogue interaction instructions are suitable for conversational task scenarios, instructing the server to generate natural language responses that conform to the context, such as generating accurate and authoritative answers by referencing product knowledge base content in intelligent customer service or collaborative robot dialogues. Thus, by differentiating task processing instructions, different task types can be accurately adapted, deeply aligning AI generation behavior with the user's actual workflow, thereby improving the accuracy, flexibility, and practicality of automated processing.

[0071] In summary, by constructing input prompts based on the selected knowledge items and the current task context, and encapsulating these prompts within task processing instructions sent to the server, the server-side large model uses the user-selected credible knowledge items as the basis for generation during inference, significantly reducing the risk of illusions or factual biases. Simultaneously, the task processing instructions define the generation intent, ensuring that the model output highly meets the user's actual work needs. This not only improves the relevance, accuracy, and business compliance of the generated results but also ensures that each output is traceable to specific knowledge bases. Ultimately, users can obtain high-quality outputs that can be directly used or fine-tuned without switching interfaces or manually integrating information, significantly shortening the creation path of high-value content and enhancing the practicality, credibility, and organizational-level knowledge reuse efficiency of the AI-assisted system.

[0072] Based on any of the above embodiments, taking document AI as an example, such as Figure 10 As shown, the task processing method of this disclosure embodiment may include the following steps: 1. Open the document AI; Users can activate the document AI function in the current document editing interface to enter the intelligent assisted creation mode; 2. Click Add; When a user clicks the "Add" button in the document AI input box, a knowledge association pop-up appears. The system then reads the user's profile and, combined with their historical behavior data (such as recently viewed, cited, or saved knowledge items) and general workflow context (such as meetings and to-do tasks in the schedule), dynamically generates a personalized intelligent recommendation list. Users can directly view and select relevant knowledge items in the "Recommended" tab of the pop-up for efficient citation. 3. Add data from other work platforms / knowledge bases; If the recommended content does not fully meet the user's needs, the user can switch to the "Manual Association" tab and select the target platform from the list of connected enterprise knowledge platforms. After the system completes OAuth authorization via the MCP protocol or standard REST API, a secure connection is established. Subsequently, the user can enter keywords in the "Search" tab. The system will then perform a full-text search concurrently across all authorized knowledge sources based on those keywords and display the results categorized by knowledge source for easy and precise location of the required content. 4. Generate; After a user selects one or more knowledge items, the selected content will be automatically injected into the document AI's input box in the form of a structured reference block. Based on the knowledge context carried by the reference block and the current task context, the document AI calls the large model to generate text content (such as requirement descriptions, analysis reports, email drafts, etc.) that conforms to business logic, is factually accurate, and has a consistent style, while retaining the traceable identifier of the original knowledge source.

[0073] To achieve the above embodiments, such as Figure 11This disclosure proposes an overall architecture for a user-context-based intelligent knowledge recommendation and citation system, which consists of five layers: a user interaction layer, an intelligent recommendation engine, a knowledge access layer, a role profiling layer, and a data storage layer. In the user interaction layer, a knowledge association overlay is triggered through a document AI dialog box, providing functions such as add buttons, recommend / manually search tabs, and select and cite knowledge, enabling users to efficiently acquire and use knowledge. The intelligent recommendation engine constructs recommendation logic based on multi-dimensional data, including behavioral data analysis (such as recently viewed or selected content), general workflow recommendations (combining meeting minutes and calendar tasks), functional role matching (such as QA corresponding to iCafe / Test, and sales corresponding to backend data), and recommendation ranking and deduplication mechanisms (combining relevance scores and timeliness weights). The knowledge access layer is responsible for connecting with various knowledge sources, supporting technologies such as MCP protocol, REST API, OAuth authorization, content structured parsing, knowledge classification and display, and real-time search engines to ensure unified access to heterogeneous data. The user profile layer integrates user job level and role information (such as HR, PM, QA), organizational structure and permission relationships, historical usage preferences (frequently used platforms and knowledge types), and workflow stage tags (such as requirements, development, deployment), providing accurate profile support for personalized recommendations. The bottom layer of data storage contains multiple specific knowledge sources, such as knowledge bases, cards (requirements / bugs / tasks), meeting minutes (ASR minutes), code repositories (code / CR), goal management platforms (goals / AR), and third-party platforms (user-customized access), forming a complete knowledge asset support system. The entire architecture realizes a closed-loop process from user behavior to knowledge recommendation to content citation, supporting the intelligent aggregation and precise delivery of multi-source heterogeneous knowledge.

[0074] To implement the above embodiments, this disclosure proposes a task processing apparatus.

[0075] Figure 12 This is a schematic diagram of the task processing device provided in Embodiment 4 of this disclosure.

[0076] like Figure 12 As shown, the task processing device 1200 includes: a detection module 1210, a display module 1220, and a selection module 1230.

[0077] The detection module 1210 is used to initialize a task processing session interface based on the task context associated with the operation in response to the detection of a task processing operation triggered by a logged-in user; the display module 1220 is used to display a knowledge association overlay in response to the detection of a knowledge reference addition operation triggered on the task processing session interface; wherein the knowledge association overlay displays a set of knowledge items including at least recommended knowledge items associated with the task context; the selection module 1230 is used to send a task processing instruction to the server based on the selected knowledge item and the task context in response to a selection operation of at least one knowledge item in the set of knowledge items, and receive the task processing result returned by the server.

[0078] As one possible implementation of this disclosure, the display module 1220 is configured to: send a recommended knowledge acquisition request to the server, receive recommended knowledge entries associated with the task context returned by the server, and display the recommended knowledge entries in a knowledge association overlay; in response to detecting an interactive operation on a knowledge source identifier triggered by the knowledge association overlay, send a knowledge source query request to the server, receive platform knowledge entries matching the knowledge source identifier returned by the server, and display the platform knowledge entries in the knowledge association overlay; in response to detecting an input operation on a search keyword triggered by the knowledge association overlay, send a search request to the server, receive search knowledge entries matching the search keyword returned by the server, and display the search knowledge entries in the knowledge association overlay.

[0079] As one possible implementation of this disclosure, it is recommended that knowledge entries be generated using the following modules: an acquisition module, a recall module, and a generation module.

[0080] The system comprises the following modules: an acquisition module, which retrieves the logged-in user's historical interaction logs, schedule data, and job role tags based on the user identifier in the task context; a recall module, which retrieves at least one knowledge item from the logged-in user's specified interaction behavior within a set time period based on the historical interaction logs, and sorts the at least one knowledge item according to its relevance score and timeliness weight to obtain a first recommended knowledge item; a schedule data module, which retrieves the logged-in user's meeting minutes and related documents within a set time window, and generates a second recommended knowledge item based on the meeting minutes and related documents; a job role tag, which queries the work platform associated with the triggering user, and generates a third recommended knowledge item based on the work platform's entry point; and a generation module, which determines a recommended knowledge item based on at least one of the first, second, and third recommended knowledge items.

[0081] As one possible implementation of this disclosure, the display module 1220 is configured to display the first recommended knowledge item in a first sub-partition of the knowledge association overlay in response to the recommended knowledge item including a first recommended knowledge item; display the second recommended knowledge item in a second sub-partition of the knowledge association overlay in response to the recommended knowledge item including a second recommended knowledge item; and display the third recommended knowledge item in a third sub-partition of the knowledge association overlay in response to the recommended knowledge item including a third recommended knowledge item.

[0082] As one possible implementation of this disclosure, platform knowledge entries are generated using the following modules: a determination module, a calling module, and a conversion module.

[0083] The module consists of: a determination module for identifying the target knowledge platform corresponding to the knowledge source identifier; a calling module for calling the adapter that interfaces with the target knowledge platform to retrieve the content list associated with the knowledge source identifier through the content interface; and a conversion module for converting the content list to obtain the platform knowledge entries.

[0084] As one possible implementation of this disclosure, the display module 1220 is used to create a tab page corresponding to the target knowledge platform in the knowledge association overlay; and to display platform knowledge entries in the tab page.

[0085] As one possible implementation of this disclosure, the search knowledge entries are generated using the following module: a retrieval module.

[0086] The retrieval module is used to generate a set of search terms based on the search keywords. The set of search terms includes the search keywords themselves, the word segments obtained after segmenting the search keywords, and the synonyms obtained by expanding the word segments. Based on the set of search terms, full-text search is performed concurrently in the accessed knowledge sources to obtain the search results returned by each knowledge source. Based on the search results returned by each knowledge source, search knowledge entries are generated.

[0087] As one possible implementation of this disclosure, the display module 1220 is used to classify the searched knowledge entries according to the knowledge source identifiers of each knowledge source to obtain classification results with each knowledge source; for any classification result, a tab is created in the knowledge association overlay, and the knowledge entries in any classification result are displayed in the tab; wherein, the title of the tab is the platform name of the corresponding knowledge source.

[0088] As one possible implementation of this disclosure, the display module 1220 is also used to highlight and mark the text fragments in the knowledge entries of the classification results that match each search term in the search term set.

[0089] As one possible implementation of this disclosure, the selection module 1230 is used to construct input prompt words based on the selected knowledge items and task context; send a task processing instruction containing the input prompt words to the server; and receive the task processing result returned by the server.

[0090] As one possible implementation of this disclosure, the task processing instructions include any of the following: a document creation instruction, used to instruct the generation of a new document based on selected knowledge items; a document update instruction, used to instruct the updating of at least a portion of the content of a target document based on selected knowledge items; and a dialogue interaction instruction, used to instruct the generation of a dialogue response based on selected knowledge items.

[0091] The task processing device of this disclosure dynamically initializes the task processing session interface based on the task context associated with the user-triggered operation, thereby gaining full awareness of the user's intent and the current scenario at the task initiation stage. Furthermore, when the user triggers a knowledge reference addition operation, the displayed knowledge association overlay accurately aggregates knowledge items highly relevant to the current task, effectively avoiding manual searching and interface switching between multiple work platforms. Moreover, by selecting a set of knowledge items, the user actively injects referenced knowledge into the task processing flow, enabling the server to construct content based on the referenced knowledge items when generating task processing results. This improves knowledge reference efficiency and content generation quality, effectively shortens the creation path for tasks such as high-value document writing or dialogue generation, and enhances the practicality of AI-assisted creation and the user's operational autonomy.

[0092] It should be noted that the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in the technical solution disclosed herein are all carried out with the consent of the user, and all comply with the provisions of relevant laws and regulations, and do not violate public order and good morals.

[0093] To implement the above embodiments, this disclosure also provides an electronic device, which may include at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the task processing method proposed in any of the above embodiments of this disclosure.

[0094] To implement the above embodiments, this disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the task processing method proposed in any of the above embodiments of this disclosure.

[0095] To implement the above embodiments, this disclosure also provides a computer program product, which includes a computer program that, when executed by a processor, implements the task processing method proposed in any of the above embodiments of this disclosure.

[0096] Figure 13 A schematic block diagram of an example electronic device 1300 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0097] like Figure 13 As shown, device 1300 includes a computing unit 1301, which can perform various appropriate actions and processes according to a computer program stored in read-only memory (ROM) 1302 or a computer program loaded from storage unit 1308 into random access memory (RAM) 1303. The RAM 1303 may also store various programs and data required for the operation of device 1300. The computing unit 1301, ROM 1302, and RAM 1303 are interconnected via bus 1304. Input / output (I / O) interface 1305 is also connected to bus 1304.

[0098] Multiple components in device 1300 are connected to I / O interface 1305, including: input unit 1306, such as keyboard, mouse, etc.; output unit 1307, such as various types of monitors, speakers, etc.; storage unit 1308, such as disk, optical disk, etc.; and communication unit 1309, such as network card, modem, wireless transceiver, etc. Communication unit 1309 allows device 1300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0099] The computing unit 1301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1301 performs the various methods and processes described above, such as task processing methods. For example, in some embodiments, the task processing method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 1308. In some embodiments, part or all of the computer program may be loaded and / or installed on device 1300 via ROM 1302 and / or communication unit 1309. When the computer program is loaded into RAM 1303 and executed by the computing unit 1301, one or more steps of the request processing method described above may be performed. Alternatively, in other embodiments, the computing unit 1301 may be configured to perform task processing methods by any other suitable means (e.g., by means of firmware).

[0100] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0101] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0102] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0103] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0104] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), the Internet, and blockchain networks.

[0105] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can also be servers for distributed systems or servers incorporating blockchain technology.

[0106] It's important to note that artificial intelligence (AI) is the study of enabling computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It encompasses both hardware and software technologies. AI hardware technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, and big data processing. AI software technologies primarily include computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graph technologies.

[0107] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0108] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A task processing method, the method comprising: In response to detecting a task processing operation triggered by a logged-in user, the task processing session interface is initialized based on the task context associated with the operation; In response to detecting a knowledge reference addition operation triggered on the task processing session interface, a knowledge association overlay is displayed; wherein, the knowledge association overlay displays a set of knowledge entries that includes at least recommended knowledge entries associated with the task context; In response to the selection operation of at least one knowledge item in the set of knowledge items, a task processing instruction is sent to the server based on the selected knowledge item and the task context, and the task processing result returned by the server is received.

2. The method according to claim 1, wherein, The response to detecting a knowledge reference addition operation triggered on the task processing session interface, displaying a knowledge association overlay, includes at least one of the following: Send a request for obtaining recommended knowledge to the server, and receive recommended knowledge entries associated with the task context returned by the server, and display the recommended knowledge entries in the knowledge association overlay; In response to detecting an interactive operation on the knowledge source identifier triggered by the knowledge association overlay, a knowledge source query request is sent to the server, and the platform knowledge entry matching the knowledge source identifier is returned by the server. The platform knowledge entry is then displayed in the knowledge association overlay. In response to detecting an input operation of a search keyword triggered by the knowledge association overlay, a search request is sent to the server, and the search knowledge entries matching the search keyword are returned by the server, and the search knowledge entries are displayed in the knowledge association overlay.

3. The method according to claim 2, wherein, The recommended knowledge items are generated using the following steps: Based on the user identifier of the logged-in user in the task context, obtain the logged-in user's historical interaction behavior logs, schedule data, and functional role tags; Based on the historical interaction behavior logs, at least one knowledge item in which the logged-in user performed a specified interaction behavior within a set time period is recalled, and the at least one knowledge item is sorted according to the relevance score and timeliness weight of the at least one knowledge item to obtain the first recommended knowledge item. Based on the schedule data, obtain the meeting minutes and related documents of the logged-in user within a set time window, and generate a second recommended knowledge item based on the meeting minutes and related documents; Based on the functional role tags of the triggering user, query the work platforms associated with the functional role tags, and generate a third recommended knowledge item based on the platform entry of the work platform; Recommended knowledge entries are determined based on at least one of the first recommended knowledge entry, the second recommended knowledge entry, and the third recommended knowledge entry.

4. The method according to claim 3, wherein, The recommended knowledge entries are displayed in the knowledge association overlay, including: In response to the fact that the recommended knowledge entry includes the first recommended knowledge entry, the first recommended knowledge entry is displayed in the first sub-partition of the knowledge association overlay; In response to the fact that the recommended knowledge entry includes the second recommended knowledge entry, the second recommended knowledge entry is displayed in the second sub-partition of the knowledge association overlay; In response to the fact that the recommended knowledge entry includes the third recommended knowledge entry, the third recommended knowledge entry is displayed in the third sub-partition of the knowledge association overlay.

5. The method according to claim 2, wherein, The platform knowledge entries are generated using the following steps: Determine the target knowledge platform corresponding to the knowledge source identifier; The adapter that interfaces with the target knowledge platform is invoked to retrieve the content list associated with the knowledge source identifier through the content interface; The content list is formatted to obtain the platform knowledge entries.

6. The method according to claim 5, wherein, The platform knowledge entries are displayed in the knowledge association overlay, including: Create a tab page corresponding to the target knowledge platform in the knowledge association overlay; The platform knowledge entries are displayed in the tab.

7. The method according to claim 2, wherein, The search knowledge entries are generated using the following steps: A set of search terms is generated based on the search keywords; wherein, the set of search terms includes: the search keywords themselves, each segmentation unit obtained after segmenting the search keywords, and synonyms obtained by synonym expansion of each segmentation unit; Based on the set of search terms, full-text search is performed concurrently in each of the accessed knowledge sources to obtain the search results returned by each knowledge source; The search knowledge entries are generated based on the search results returned by each of the knowledge sources.

8. The method according to claim 7, wherein, The display of the search knowledge entries in the knowledge association overlay includes: The search knowledge entries are classified according to the knowledge source identifiers of each knowledge source to obtain the classification results with respect to each knowledge source. For any classification result, a tab is created in the knowledge association overlay, and the knowledge entries in any classification result are displayed in the tab; wherein, the title of the tab is the platform name of the corresponding knowledge source.

9. The method according to claim 8, wherein, The method further includes: Text segments in the knowledge entries of the classification results that match each search term in the search term set are highlighted.

10. The method according to claim 1, wherein, The process of processing tasks based on the selected knowledge items and the task context, sending task processing instructions to the server, and receiving task processing results returned by the server includes: Based on the selected knowledge items and the task context, construct input prompt words; Send a task processing instruction containing the input prompt to the server and receive the task processing result returned by the server.

11. The method according to any one of claims 1-10, wherein, The task processing instructions include any of the following: Document creation instructions are used to instruct the generation of a new document based on the selected knowledge items; Document update instructions are used to instruct that at least a portion of the content of a target document be updated based on selected knowledge items; Dialogue interaction instructions are used to instruct the generation of dialogue responses based on the selected knowledge items.

12. A task processing apparatus, the apparatus comprising: The detection module is used to initialize the task processing session interface based on the task context associated with the detected task processing operation triggered by the logged-in user. The display module is used to display a knowledge association overlay in response to detecting a knowledge reference addition operation triggered on the task processing session interface; wherein the knowledge association overlay displays a set of knowledge entries that includes at least recommended knowledge entries associated with the task context; The selection module is configured to respond to a selection operation on at least one knowledge item in the set of knowledge items, send a task processing instruction to the server based on the selected knowledge item and the task context, and receive the task processing result returned by the server.

13. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-11.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-11.

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

Images