Method, device, equipment and product for processing information

Abstract

The invention relates to a method, a device, equipment and a product for processing information. The method includes obtaining data from a plurality of data sources. The method further includes generating at least one task to be processed based on the data from the plurality of data sources. In addition, the method further comprises the step of displaying at least one to-be-processed task and a task processing mode corresponding to each to-be-processed task. And in response to the feedback of the user for the first task in the at least one to-be-processed task, generating the first task based on the task processing mode. Through the method, the working efficiency and the user experience of the user are improved.

Description

Technical Field

[0001] This disclosure generally relates to the field of computers, and more specifically to methods, apparatus, devices, and computer program products for processing information. Background Technology

[0002] Regarding the processing of information from information sources / platforms, most current technical solutions trigger the process through dialogue with users. For example, some products integrate multiple information sources, extracting and periodically updating information according to unified permissions. Users then extract usable information using structured indexes and semantic searches. Users configure rules to generate task recommendations based on the extracted usable information. Another example is products that support targeted adaptation and templated configuration for third-party applications. Summary of the Invention

[0003] In embodiments of this disclosure, a method, apparatus, electronic device, and computer program product for processing information are provided.

[0004] In a first aspect of this disclosure, a method for processing information is provided. The method includes acquiring data from multiple data sources. The method also includes generating at least one task to be processed based on the data from the multiple data sources. Furthermore, the method includes displaying at least one task to be processed, and a task processing mode corresponding to each task to be processed. The method further includes generating a first task based on the task processing mode in response to user feedback regarding a first task among the at least one task to be processed.

[0005] In a second aspect of this disclosure, an electronic device for processing information is provided. The device includes an acquisition module configured to acquire data from multiple data sources. The device also includes a first generation module configured to generate at least one task to be processed based on the data from the multiple data sources. Furthermore, the device includes a display module configured to display the at least one task to be processed and a task processing mode corresponding to each task. Finally, the device includes a second generation module configured to generate a first task based on the task processing mode in response to user feedback regarding a first task among the at least one task to be processed.

[0006] In a third aspect of this disclosure, an electronic device is provided. The electronic device includes a processor. The electronic device also includes a memory coupled to the processor, the memory having instructions stored therein, which, when executed by the processor, cause the electronic device to perform the method according to a first aspect of this disclosure.

[0007] In a fourth aspect of this disclosure, a computer program product is provided, comprising computer-executable instructions, wherein the computer-executable instructions are executed by a processor to implement the method according to a first aspect of this disclosure.

[0008] In a fifth aspect of this disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores computer-executable instructions, which are executed by a processor to implement the method according to the first aspect.

[0009] The summary section is intended to present the chosen concepts in a simplified form, which will be further described in the detailed description below. The summary section is not intended to identify key or principal features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter. Attached Figure Description

[0010] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0011] Figure 1 A schematic diagram is shown illustrating an example environment in which one or more embodiments of the present disclosure may be implemented;

[0012] Figure 2 Flowcharts of methods for processing information according to some embodiments of this disclosure are shown;

[0013] Figure 3 Detailed block diagrams of information processing modules according to some embodiments of the present disclosure are shown;

[0014] Figure 4 A block diagram of an apparatus for processing information from a data source according to some embodiments of the present disclosure is shown;

[0015] Figure 5 A block diagram of an apparatus for analyzing information from a data source according to some embodiments of the present disclosure is shown;

[0016] Figure 6 A flowchart is shown illustrating a method for converting detected near-deadline (DDL) and high-priority to-do items into standardized suggested task cards;

[0017] Figure 7 A block diagram of an apparatus for a recommendation task according to some embodiments of the present disclosure is shown;

[0018] Figure 8 A block diagram illustrating a method for scoring, deduplicating, or sorting tasks according to some embodiments of the present disclosure is shown;

[0019] Figure 9 A block diagram of an apparatus for displaying or interacting with a recommendation task according to some embodiments of the present disclosure is shown;

[0020] Figure 10 A block diagram of an apparatus for safeguarding information security and access rights according to some embodiments of the present disclosure is shown;

[0021] Figure 11 Block diagrams of electronic devices for processing information according to some embodiments of the present disclosure are shown; and

[0022] Figure 12 A schematic block diagram of an electronic device according to some embodiments of the present disclosure is shown.

[0023] In all the accompanying figures, the same or similar reference numerals denote the same or similar elements. Detailed Implementation

[0024] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0025] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0026] For example, upon receiving a user's proactive request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message.

[0027] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0028] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0029] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0030] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects unless explicitly stated. Other explicit and implicit definitions may also be included below.

[0031] As mentioned above, current technologies for processing information from information sources / platforms mostly rely on user interaction to trigger responses. However, in multi-source work platforms with multiple fragmented information sources, users struggle to manually and efficiently identify deadlines, to-dos, or critical events requiring immediate response (such as code review requests or online alerts). This results in critical signals being buried in massive amounts of information, leading to task delays and decreased collaboration efficiency.

[0032] Therefore, one or more embodiments of this disclosure provide a proactive task suggestion method or system through an alternative approach. This method transforms information silos scattered across multiple collaboration tools into a series of pre-emptive tasks that can be initiated with a single click on the user's workbench through unified event perception or task mining and prioritization. This proactively prepares the user for the next action before deadlines or critical events, achieving a leap from passive response to proactive pre-emption, thereby significantly improving the user experience.

[0033] Therefore, more specifically, one or more embodiments of this disclosure provide a scheme for user information processing. The scheme includes acquiring data from multiple data sources. The scheme also includes generating at least one task to be processed based on the data from the multiple data sources. Furthermore, the scheme includes displaying at least one task to be processed, and a task processing mode corresponding to each task to be processed. The scheme further includes generating a first task based on the task processing mode in response to user feedback regarding a first task among the at least one task to be processed.

[0034] In this way, one or more embodiments of this disclosure take a different approach, providing a proactive task suggestion method or system. One or more embodiments of this disclosure, through unified event awareness or task mining and prioritization, transform information silos scattered across multiple collaboration tools into a series of pre-emptive tasks that can be initiated with a single click on the user's workbench. This proactively prepares the user for the next step before deadlines or critical events, achieving a leap from passive response to proactive pre-emption, thereby significantly improving user experience and work efficiency.

[0035] Figure 1 A schematic diagram of an example environment 100 in which one or more embodiments of the present disclosure may be implemented is shown. Figure 1 As shown, environment 100 includes processing device 102. Processing device 102 can be any device with processing capabilities. For example, processing device 102 can include smartphones, tablets, laptops, desktop computers, workstations, quantum computers, or smart wearable devices. In environment 100, processing device 102 can run application 104, which can be a machine learning model application. User 108 can interact with application 104 through processing device 102. In some embodiments, application 104 can obtain various types of information from a data source. In some embodiments, the data source can include platforms 110, 112, or 114, etc. In application 104, users can use relevant information from platforms 110, 112, or 114, etc., to create tasks or complete related tasks, such as completing requirements assessment, product design, code writing, or project management, etc. For example, application 104 can also be a machine learning model application platform, etc. Platforms 110, 112, or 114, etc., can be at least one of a product requirements platform, a code repository platform, an instant messaging platform, a product development task platform, etc. Application 104 may include an information processing module 106 according to one or more embodiments of the present disclosure. The information processing module 106 can extract at least one task to be processed from information from various platforms 110, 112, or 114, etc., based on interaction with user 108, and sort the at least one task to be processed according to a task sorting strategy, thereby recommending the at least one task to be processed to user 108 based on the sorted tasks. Since the information processing module 106 can simultaneously extract at least one task to be processed and recommend the sorted task to user 108, user 108 can shorten the initial value experience time for new users, thereby greatly improving user experience and work efficiency, and enabling timely completion of related tasks. On the other hand, the information processing module 106 can also form an independent application or application platform, rather than simply being an information processing module of application 104.

[0036] It should be understood that, for the sake of brevity, Figure 1This document only shows a specific number of processing devices 102, applications 104, information processing modules 106, and platforms 110, 112, or 114, etc., but it is not intended to limit the number of processing devices 102, applications 104, information processing modules 106, and platforms 110, 112, or 114, etc., as applicable to this disclosure. In other implementations, any number of processing devices 102, applications 104, information processing modules 106, and platforms 110, 112, or 114, etc., applicable to this disclosure may be used. For example, processing device 102 may also be a cloud computing platform, etc.

[0037] In this way, an environment 100 containing the information processing module 106 implemented in one or more embodiments of this disclosure provides a proactive task suggestion method or system. The environment 100 containing the information processing module 106 implemented in one or more embodiments of this disclosure, through unified event awareness or task mining and sorting, transforms information silos scattered across multiple collaboration tools into a series of pre-emptive preparation tasks that can be initiated with a single click on the user's workbench. This proactively prepares the user for the next step before deadlines or critical events occur, achieving a leap from passive response to proactive pre-emptive action, thereby greatly improving the user experience.

[0038] The following will combine Figures 2 to 12 The process according to embodiments of this disclosure is described in detail. For ease of understanding, the specific data mentioned in the following description are exemplary and not intended to limit the scope of this disclosure. It should be understood that the embodiments described below may also include additional actions not shown and / or actions shown may be omitted, and the scope of this disclosure is not limited in this respect.

[0039] Figure 2 A flowchart illustrating a method 200 for processing information according to some embodiments of the present disclosure is shown. Method 200 can be executed by a processing device. For example, method 200 can be performed by… Figure 1 The processing device 102 performs the operation. The processing device may include, for example, a smartphone, tablet computer, laptop computer, desktop computer, workstation, quantum computer, cloud computing platform, or smart wearable device.

[0040] In box 202, data from multiple data sources is acquired. In some embodiments, the data source may include one or more platforms. In some embodiments, the platform may be, for example, platform 110, 112, or 114. In some embodiments, the platform may be at least one of a product requirement platform, a code repository platform, an instant messaging platform, a product development task platform, etc. In some embodiments, acquiring data from multiple data sources includes at least one of the following: subscribing to event information from the data source through the event subscription interface of the data source, or determining whether an event has been processed based on the event identifier in the event information, and refusing to receive event information in response to the event being processed. In some embodiments, acquiring data from multiple data sources may also include periodically querying the data source through the information access interface of the data source to obtain change information missed due to event loss or subscription delay. In some embodiments, acquiring data from multiple data sources may also include at least one of the following: converting information from the data source into information conforming to the normalized event model based on a normalized event model, storing information from the data source in a database that meets high-concurrency read and write requirements, storing hot data from the information from the data source in a cache, archiving or cleaning cold data from the information from the data source, or controlling the access rate of information from the data source based on a polling cycle or batch processing mechanism. In some embodiments, method 200 may further include at least one of processing user personal data using a user-accessible token or processing public data authorized by an administrator using a tenant-accessible token.

[0041] In box 204, at least one task to be processed is generated based on data from multiple data sources. In some embodiments, generating at least one task to be processed based on data from multiple data sources may include extracting at least one task to be processed from information from the data sources based on keywords or sentence structure rules. In some embodiments, generating at least one task to be processed based on data from multiple data sources may further include parsing information from the data sources to determine a first field in the information from the data sources, and extracting at least one task to be processed from the first field in the information from the data sources. In some embodiments, generating at least one task to be processed based on data from multiple data sources may further include identifying information from the information from the data sources that indicates the need for action, and extracting at least one task to be processed based on the information indicating the need for action. In some embodiments, generating at least one task to be processed based on data from multiple data sources may further include, in response to information from the data sources including multiple time information, employing the time information associated with the information indicating the need for action from the multiple time information. In some embodiments, generating at least one task to be processed based on data from multiple data sources may further include identifying information from the information from the data sources that indicates the need for action through a large language model (LLM), and extracting at least one task to be processed based on the information indicating the need for action. In some embodiments, generating at least one task to be processed based on data from multiple data sources may further include, in response to information from the data sources including multiple time information, using the time information associated with the information indicating that action needs to be taken from the multiple time information.

[0042] In some embodiments, generating at least one pending task based on data from multiple data sources may further include triggering an LLM in response to a failure of other strategies for extracting information besides LLM to match and the information from the data source meeting a first threshold, or extracting at least one of the at least one pending task based on the information from the data source and other strategies for extracting information besides LLM in response to an unavailable LLM or an LLM response time exceeding a first time threshold.

[0043] In some embodiments, generating at least one task to be processed based on data from multiple data sources may further include determining the priority of a time window and extracting at least one task to be processed based on information from the data sources, the priority of the time window of the information from the data sources, and a strategy; or identifying time information in the unstructured information from the information from the data sources through natural language processing or named entity recognition; or determining at least one of the responsible persons for at least one task to be processed based on a strategy for determining task attribution.

[0044] Back Figure 2 In box 206, at least one pending task and a corresponding task processing mode are displayed. In some embodiments, method 200 may further include sorting the at least one pending task based on a first strategy for task sorting. In some embodiments, sorting the at least one pending task based on the first strategy may include scoring the at least one pending task based on the first strategy, and sorting the at least one pending task based on the scored at least one pending task. In some embodiments, scoring the at least one pending task based on the first strategy may include scoring the at least one pending task based on at least one of the following: task urgency, task impact, task complexity, task relevance, or task source credibility. In some embodiments, sorting the at least one pending task based on the first strategy may further include adjusting the sorting of the at least one pending task based on the interaction between a user (e.g., user 108) and the at least one pending task.

[0045] In some embodiments, displaying the at least one pending task may include aggregating and sorting the at least one pending task based on the topic key or text similarity of the at least one pending task, or merging key fields in the at least one pending task after sorting, or collapsing the at least one pending task that is ranked lower, or hiding at least one of the at least one pending task after sorting that has exceeded a first waiting time threshold and has not been triggered by new events. In some embodiments, displaying the at least one pending task may further include obtaining a task template associated with the at least one pending task, parsing information associated with the at least one pending task from information from a data source, and filling the associated fields in the task template with the parsed information associated with the at least one pending task from the data source. In some embodiments, displaying the at least one pending task may also include recording the generation basis of the recommended at least one pending task, the generation basis including the event identifier referenced by the at least one pending task or the rating of the at least one pending task, or recording at least one of the source, extraction method, or credibility of the parameters of the task template.

[0046] In block 208, in response to user feedback regarding a first task among at least one pending tasks, a first task is generated based on a task processing mode. In some embodiments, method 200 may include generating the first task based on a task template associated with the first task. In some embodiments, method 200 may further include generating the first task based on information associated with the first task from information from a data source and the task template associated with the first task. In some embodiments, method 200 may further include any of the following: in response to the existence of a template corresponding to the first task, generating the first task based on the template corresponding to the first task; in response to the absence of a template corresponding to the first task, generating the first task based on data from multiple data sources associated with the first task; or in response to the absence of a template corresponding to the first task, requesting a platform corresponding to another data source to process the first task.

[0047] In this way, one or more embodiments of Method 200 provide a proactive task suggestion method or system. One or more embodiments of Method 200, through unified event awareness or task mining and prioritization, transform information silos scattered across multiple collaboration tools into a series of pre-emptive tasks that can be initiated with a single click on the user's workbench. This proactively prepares the user for the next action before deadlines or critical events, achieving a leap from passive response to proactive pre-emption, thereby significantly improving the user experience. Furthermore, one or more embodiments of Method 200 connect cross-source incremental extraction, near-DDL detection and To-Do extraction, recommendation engine, and one-click template initiation into a closed loop, significantly reducing the user's cognitive and initiation costs, improving the consumption efficiency of personalized matching and team collaboration, and generating continuous and measurable beneficial technical effects.

[0048] Figure 3A detailed block diagram of an information processing module 300 according to some embodiments of the present disclosure is shown. The information processing module 300 includes an information processing main interface 301. The information processing main interface 301 includes a text input box 303 for receiving task prompts from the user. The information processing main interface 301 also includes a work summary control 304, a recommended task control 305, and a task template control 307. In some embodiments, the recommended task control 305 can obtain information from the platform and, based on the information from the platform and a strategy for extracting information, extract multiple tasks, wherein the strategy includes multi-level strategies for extracting information. It can also display the at least one task to be processed and, in response to user feedback on a first task among the at least one recommended task to be processed, generate the first task. In some embodiments, the work summary control 304 included in the information processing module 300 can define a unified information source access and processing layer, designed to standardize heterogeneous information from different products and platforms as data input for associated tasks. All information entry strictly follows the principle of permissions following the currently logged-in user, and supports users setting the permissions below to globally public information during configuration, allowing space members to share information sources. The work summary control 304 can be used to summarize a user's work content and key points over a period of time. In some embodiments, the user can set the interval between work summaries. In some embodiments, the information processing module 300 only retrieves and processes resources that the currently logged-in user has access to, such as personal documents, calendar events, chat logs, etc. This is the basic permission constraint for all data reading. In some embodiments, users with administrator privileges can also centrally bind team-shared resources, such as code repositories and project management platform data, in the resource configuration module (not shown) of the project space. This configuration only serves as an index entry point for resources; the user's personal permissions are still verified during actual access.

[0049] In some embodiments, the task recommendation control 305 can recommend task suggestions 309, 313, or 315 to the user. Task suggestions 309, 313, or 315 may include a task description and related information about creating the task (in some embodiments, task suggestions 309, 313, or 315 may also be figuratively called suggested task cards or task cards). In some embodiments, task suggestions 309, 313, or 315 may include a task creation control 311. If a user wishes to create the task, they can click the task creation control 311, and the information processing module 300, in response to the user's feedback on the task creation suggestion, generates a relevant task based on the information associated with the task and the task template. In some embodiments, task suggestions 309, 313, or 315 are recommended in a sorted manner based on a first strategy for task sorting. For example, based on the first strategy for task sorting, if task suggestion 309 is more important than task suggestions 313 or 315, then task suggestion 309 is ranked first, thus receiving more attention and importance from the user.

[0050] In some embodiments, the task template control 307 can be used to generate or manage related task templates. In some embodiments, the task templates may be designed or created by a service provider. In some embodiments, the task templates may also be designed or created by a user. In some embodiments, related task templates can be associated with related tasks. In some embodiments, related task templates can be associated with related platforms or related events on related platforms. Users can configure related settings through a settings module (not shown).

[0051] Figure 4 A block diagram of an apparatus 400 for processing information from a platform, according to some embodiments of the present disclosure, is shown. Figure 4 The technical solution of one or more embodiments of this disclosure is based on building a data pipeline capable of efficiently and near real-time extracting incremental information from multiple heterogeneous information sources 401 (such as instant messaging (IM) chat 403, meeting minutes 405, cloud documents 407, tasks 409, or information from third-party applications 411), and converting it into a unified and standardized event model 419 (Canonical Event Model). This model is the basis for all subsequent analysis and suggestion generation, and its uniqueness lies in upgrading simple information synchronization into a structured, decision-making event stream.

[0052] In some embodiments, device 400 can provide an incremental information retrieval mechanism. In some embodiments, to ensure data freshness and minimize system overhead, device 400 employs a hybrid strategy at the event acquisition layer 413, prioritizing event subscription 415 and supplementing it with periodic polling 417 for incremental data retrieval. In some embodiments, event subscription 415 can utilize the event subscription capabilities of a real-time communication platform (long-connection WebSocket is recommended) to receive change notifications from various data sources in near real-time. This approach ensures low information latency. For example, for cloud document 407, events such as drive.file.edited_v1 (file editing) and drive.file.created_v1 (file creation) can be subscribed to. For chat logs 403, events such as im.message.receive_v1 (receiving new messages) can be subscribed to. And for task 409, events such as task.task.updated_v1 (task update) and task.task.created_v1 (task creation) can be subscribed to. Additionally, for meeting minutes 405, the meeting end signal can be captured as an indirect trigger by subscribing to events such as calendar.event.changed_v4 (calendar.event.change_v4).

[0053] In some embodiments, the device 400 can perform idempotent processing on the information. For example, each received event contains a unique event_id (event identifier). Before processing an event, the device 400 can check whether the identifier (ID) has already been consumed, ensuring that the same event is not processed repeatedly and guaranteeing data consistency. In some embodiments, for the periodic polling 417, as a supplement to event subscription and a fault tolerance mechanism, the device 400 can periodically and proactively query each data source to capture changes that may be missed due to event loss or subscription delays. In some embodiments, the polling strategies for each data source are different. For example, for cloud documents 407, the device 400 can periodically call interfaces such as GET / open-apis / drive / v1 / files / list, use commands such as order_by = EditedTime and direction = DESC for sorting, and record the latest modified_time returned by each query. In the next poll, this timestamp is used as the starting point for the query, thereby obtaining all subsequent changes. For chat 403 records, for high-value sessions (such as project groups), you can periodically call interfaces such as GET / open-apis / im / v1 / messages, and use the start_time and end_time parameters to limit the query time window. The system will save the end time of the last poll as the start time of the next poll. In some embodiments, for meeting minutes 405, after receiving the meeting end event, you can delay for a preset time (e.g., 5 minutes), and then try to call the relevant interface through the meeting ID to obtain the minute_token (meeting minutes token) to check whether meeting minutes 405 has been generated. In some embodiments, for tasks 409, you can periodically call interfaces such as GET / open-apis / task / v2 / tasks, and use the task's updated_at (when it was updated) or completed_at (when it was completed) timestamp fields for incremental retrieval. In some embodiments, for pagination processing, all polling interfaces can follow the pagination specifications of the real-time communication platform, using parameters such as page_token and page_size for iteration until all updated data is retrieved. In some embodiments, device 400 also provides a data deduplication mechanism. For example, the data obtained through polling is compared with existing data using its unique identifier (e.g., message_id, file_token, task_guid) to prevent duplicate entries into the database.

[0054] In some embodiments, the device 400 also provides a normalized event model 419 to normalize information or data obtained through event subscription 415 or timed polling 417. All raw data is converted into a unified structured event, masking differences between source platforms. This model not only contains basic information but also includes metadata for subsequent decision-making. To mask multi-source heterogeneity, all raw data obtained from different sources is converted into a unified normalized event model 419 after preliminary processing (parsing, deduplication). This model is the basis for all subsequent analyses (such as DDL detection, To-Do extraction).

[0055] In some embodiments, the device 400 stores or caches the collected information or data at the data processing and caching layer 421. For example, normalized event data is stored in a database suitable for high-concurrency read / write operations (such as a NoSQL database). In some embodiments, hot data (such as data from the last 7 days) is stored in a cache to accelerate the loading of workbench components. In some embodiments, considering the time window requirements for DDL detection (typically 24 hours to 7 days), the device 400 sets a rolling data retention window (e.g., retaining event data from the last 30 days). Cold data exceeding the retention period is archived or cleaned up to control storage costs and query performance. In some embodiments, the device 400 may also provide a performance constraint mechanism. For example, all calls to the instant messaging platform application interface (API) strictly adhere to its rate limiting, and through reasonable polling cycles and batch processing mechanisms, the impact on external services is avoided, thereby ensuring system stability.

[0056] Figure 5 A block diagram of an apparatus 500 for analyzing information from a platform, according to some embodiments of the present disclosure, is shown. One or more embodiments of the present disclosure are proactive pre-emptive suggestion flows, which, instead of passively identifying DDLs and To-Dos, proactively generate pre-emptive tasks within a reasonable time window before these events occur. The following will be combined with... Figure 5 This section details how the system proactively discovers DDL (Detailed Description of Requirements) and To-Do (To-Do) items from multi-source heterogeneous information and transforms them into actionable suggestions for the user. Figure 5In this system, device 500 includes an analysis engine layer 501. The analysis engine layer 501 includes an engine 503 for DLL proximity detection and an engine 505 for To-Do extraction. In some embodiments, the engine 503 for DLL proximity detection can manage and implement DDL proximity detection windows. For example, to effectively remind users of time-sensitive tasks, the system introduces a configurable proximity window concept to define the time range of "proximity." For example, the system can use the following time window configuration, with three preset time windows covering scenarios of different urgency levels: Urgent: Items due within 24 hours; Important: Items due within 72 hours; Notice: Items due within 7 days. In some embodiments, the configurability of the time window management and coverage strategy allows project space administrators to customize and modify the time thresholds of each window according to the team's work rhythm and business needs. For example, for teams requiring rapid response, the "urgent" window can be adjusted to 12 hours. In some embodiments, for time priority coverage, for example, the detection logic follows the proximity principle. If a task meets multiple window conditions simultaneously, the system will prioritize the most urgent label. For example, a task due in 20 hours will be marked as "Urgent" rather than "Important" or "Attention." Furthermore, for global and personal configurations, a project-space-level global configuration can be used to ensure consistent DDL awareness within the team. In some embodiments, it can be extended to allow users to override global configurations to meet personalized time management needs.

[0057] In some embodiments, the DDL extraction logic involves the device 500 achieving robust DDL extraction by combining structured data parsing and unstructured text entity recognition. For example, for structured source extraction, for task-type data, the `due` field in the task object can be directly parsed. This field contains a clear due date and time (Unix timestamp) and is the highest priority DDL source. In some embodiments, for bitable data, when the user configures a bitable table as an information source in the workbench, the device 500 can guide the user to specify a date type field as the DDL source.

[0058] In some embodiments, in terms of text entity recognition (NER), for example, for unstructured text such as instant messaging chat logs, meeting minutes, cloud document titles and contents, the extraction method that the device 500 may employ includes utilizing a natural language processing (NLP) module and employing named entity recognition (NER) technology, specifically trained to recognize entities related to dates and times.

[0059] In some embodiments, regarding robustness assurance, the device 500 normalizes all parsed times according to context (such as the message sender's timezone setting) or default settings (such as the server's timezone) for timezone processing, converting them into standard UTC timestamps for storage and calculation, thereby avoiding cross-timezone confusion. In some embodiments, for natural language date parsing, it supports parsing multiple natural language expressions. For example, it includes relative dates: "tomorrow at 3 PM", "the day after tomorrow", "next Friday"; it can also include ambiguous dates: "end of the month", "before the weekend"; or standard formats: "2025-12-25", "12 / 25 / 2025 10:00 PM", etc. In some embodiments, ambiguity resolution can also be performed. For example, when multiple dates appear in the text, the device 500 will combine the context (such as keywords like "DDL is...", "deadline is...") to determine and prioritize the date strongly associated with the action item.

[0060] In some embodiments, the engine 505 used for To-Do extraction can employ a multi-layered To-Do extraction strategy. That is, the engine 505 can use a multi-layered strategy to accurately extract to-do items from different information sources. For example, it can employ rule-based heuristics, which may include high-priority rules. This method serves as a foundational and fast-response layer, identifying potential To-Do items by matching predefined keywords and sentence patterns.

[0061] In some embodiments, rule examples may include: explicit instructions, such as phrases like "@Zhang San, remember to complete...", "Please handle this...", "Needs follow-up..."; interrogative assignments, such as "Who is responsible for this plan?", "...the conclusion is..., the follow-up...", "...has the plan been decided?"; and implicit tasks, such as "Need to output by next Wednesday...", "The goal is...". The advantages of using these rules are low computational cost, fast response time, high accuracy, and suitability for processing text with relatively fixed formats.

[0062] In some embodiments, to-do items can be obtained directly from a highly reliable source. For example, the "Action Items" module in an instant messaging platform is a structured list of to-dos confirmed by the user or agent. Its extraction logic may include, when the engine 505 obtains the data, directly parsing its content and extracting each item in the "Action Items" list as an independent to-do. These to-dos contain clearly defined responsible parties and content, possessing extremely high reliability.

[0063] In some embodiments, device 500 may also employ LLM-assisted classification. To handle more complex or colloquial expressions that are difficult to cover by rules, device 500 may introduce LLM as a high-level classifier. Specifically, for text fragments not matched by rules (such as a chat message), device 500 invokes LLM for intent recognition. In some embodiments, LLM determines whether the text contains one or more action items and attempts to extract the executor, content, and deadline (if any) of the action items.

[0064] In some embodiments, device 500 may also provide a fallback mechanism. To balance cost and efficiency, LLM calls are not performed indiscriminately. LLM analysis is only triggered when heuristic rules fail to match and text length, source credibility (e.g., from the core project group), etc., reach certain thresholds. In some embodiments, if the LLM service is unavailable or the response times out, the system will smoothly fall back, relying only on rules and structured data for extraction, ensuring the stable operation of core functions.

[0065] In some embodiments, device 500 can also provide task attribution rules. To ensure that extracted DDLs and To-Dos accurately reach the relevant responsible parties, device 500 has designed explicit attribution rules. For example, by default, device 500 attributes all items scanned within its personal permission scope (such as personal task lists and private chat messages) to the currently logged-in user. Additionally, in group chats, document comments, or meeting minutes, if a To-Do or DDL explicitly points to one or more users using the @ symbol, device 500 can precisely attribute the item to the mentioned user. In some embodiments, the handling of tasks not explicitly assigned in group chats can be resolved through contextual attribution. For example, if a To-Do is proposed in a group chat but not explicitly @ anyone, device 500 can attempt to determine the attribution based on the context. For instance, if the message is a direct reply to someone's previous statement, the task may be assigned to that person.

[0066] In some embodiments, device 500 can provide a group pending task mechanism. For example, if the context does not clearly point to a single responsible party, the task will be marked as a "group pending task" and made visible to all group members in the "Following Events" section of the project space workbench, encouraging relevant personnel to proactively claim it. In some embodiments, device 500 can provide a responsible party priority mechanism. For example, if there is a clearly defined project leader or module leader in the group chat, device 500 may tend to assign unassigned tasks to them, or at least send them as important notifications.

[0067] Figure 6A flowchart of a method 600 for converting detected DDL and To-Do statements into standardized suggested task cards is shown. Figure 6 In the process, information from various platforms arrives at point 601 via event triggering, thus collecting relevant information from each platform. At point 603, the relevant information from each platform is parsed and deduplicated. Then, based on a normalized event model, at point 605, all relevant information or raw data from each platform is converted into a unified structured event, thereby masking the differences between source platforms. At point 605, the time zone of the data is normalized to form a standard and universal time zone. At point 609, each DDL or To-Do is identified through proximity window classification. At point 611, an independent card (not shown) can be generated for each identified DDL or To-Do. In some embodiments, the card content may include a task summary, a source link (allowing one-click jump to the original message, document, or task), a deadline (if specified), and suggested follow-up action buttons. In some embodiments, method 600 may also pre-fill task information at point 613 to generate relevant tasks. And at point 615, the relevant tasks are displayed on the corresponding workbench.

[0068] In some embodiments, for example, when a user clicks the "Create Task" control on a card, method 600 can automatically create a task window and pre-fill the extracted task content and deadline; the user only needs to confirm. In some embodiments, method 600 can also match existing task templates based on the content of the To-Do statement. For example, if the content contains keywords such as "code review" or "weekly report," the card will directly display controls for "Initiate Code Review with One Click" or "Generate Weekly Report with One Click." In some embodiments, method 600 can also automatically fill in template parameters. For example, when a user clicks a control linked to a template, method 600 can not only initiate the template but also attempt to extract more information from the original context to automatically pre-fill the parameters required for the template. For example, in a code review scenario, method 600 can also automatically find and fill in links such as merge request links from the context. In some embodiments, when no matching template recommendation is identified, method 600 will generate a query based on the identified To-Do content using a large language model and system experience / knowledge and pre-fill it into the main input box. Through the above methods, one or more embodiments of method 600 achieve the automated transformation from information scattered across various work collaboration tools to structured, actionable, and clearly defined task suggestions, greatly shortening the path for users from obtaining information to executing tasks.

[0069] Figure 7 A block diagram of an apparatus 700 for a recommendation task according to some embodiments of the present disclosure is shown. Figure 8This illustrates a block diagram of a method 800 for scoring, deduplicating, or sorting tasks according to some embodiments of the present disclosure. The following will be combined with... Figure 7 and Figure 8 This document describes how one or more embodiments of the present disclosure generate task recommendation cards for cross-source signals, and perform task scoring, repetition suppression, and sorting. It also illustrates the display linkage with workbench components (e.g., insight reports, events of interest) and the one-click initiation of templates.

[0070] like Figure 7 As shown, the device 700 includes a recommendation engine 701 for recommending tasks. The recommendation engine 701 may include a scoring module 703 for rating multiple tasks or events. In some embodiments, the recommendation engine 701 may further include a deduplication and aggregation module 705 for deduplicating or aggregating similar tasks or events among multiple tasks or events. In some embodiments, the recommendation engine 701 may further include a sorting module 707 for ranking multiple tasks or events. How one or more embodiments of this disclosure perform rating, duplication suppression, and ranking of multiple tasks or events based on the recommendation engine 701 will be described below in conjunction with... Figure 8 Please provide a detailed explanation.

[0071] exist Figure 8In the context of the recommendation engine 701, at box 801, the input can include unified event fields for the event and context. For example, fields derived from the normalized event model include source_type, event_time, content.normalized_text, content.entities, metadata.url, metadata.due_date, and actor. The input information or data needs to undergo task cue recognition (e.g., DDL detection, To_Do extraction), rating dimension calculation at boxes 803, 805, and 817 (e.g., urgency 807, impact 809, difficulty (effort required) 811, attribution 813, source credibility 815), topic aggregation and repetition suppression at box 819, and sorting and binning (or regional) display at box 821 (e.g., pinned 823, currently displayed 825, or later displayed 827). In some embodiments, the recommendation engine 701 can output structured fields for task recommendation cards. These include, for example, title, summary, source_type, topic_key, score, ranking_bucket, due_date, owner, source_cred, merged_sources, and actions, and may include an initiation template entry and pre-filled parameters. To help those skilled in the art understand the specific details of each step of method 800, scoring, deduplication, and sorting will be explained in more detail below.

[0072] In some embodiments, the multi-dimensional scoring and dynamic weights are configurable or learnable. To prioritize valuable tasks for users, one or more embodiments of this disclosure provide a unique ranking mechanism. Specifically, one or more embodiments of this disclosure first provide a unified scoring model and a unified scale. As an example, the unified scoring model uses a weighted linear combination as shown in equation (1): (1) In the "score" task, each dimension is mapped to [0.00, 1.00]. `w_u`, `w_i`, `w_e`, `w_o`, and `w_s` are the weights for each dimension, and the default weights satisfy `w_u + w_i + w_e + w_o + w_s = 1.00`. Furthermore, `urgency` represents urgency, `impact` represents impact, `effort` represents difficulty (i.e., the degree of effort required), `owner_match` represents task ownership, and `source_cred` represents source credibility. Table 1 lists the default baseline weights as follows. Table 1

[0073] In some embodiments, urgency can be determined by both the proximity to the deadline and the use of urgency-related words in the text. For example, Table 2 lists some examples of relevant scenarios and their corresponding urgency mapping values. Table 2 In some embodiments, if the text contains phrases such as "urgent / immediately / right away," the mapping value can be additionally increased by +0.10 (up to a maximum of 1.00). In some embodiments, the time window threshold and the increase can be configured by the administrator in the background.

[0074] In some embodiments, the base value of impact can be calculated as shown in equation (2): impact_base = min(1.00, log(1 + audience_size) / log(1 + 50)) (2) Where `impact_base` represents the base value of the impact score, and `audience_size` represents the audience size. In some embodiments, audience size can be combined with role weights. For example, for sources from administrators / responsible persons, a role bonus of +0.10 can be applied; for sources from core collaboration groups (such as project groups), a role bonus of +0.05 can be applied. The maximum role bonus is 1.00. In some embodiments, platform type bonuses can also be applied. For example, for sources from structured platforms, a bonus of +0.05 can be applied.

[0075] In some embodiments, Table 3 shows examples of complexity mapping based on complexity hierarchy for complexity effort. Complexity effort is a negative term; a larger value indicates a higher workload. Table 3 In some embodiments, the mapping value can be adjusted by estimating the source of any one of the following: textual clues (verb phrases, length), task type, or historical statistics. In some embodiments, the mapping value can be calibrated by an administrator.

[0076] In some embodiments, for task owner_match, Table 4 shows examples indicating the strength of the association between a task and the current user. Table 4

[0077] In some embodiments, Table 5 shows an example of the mapping between source type and source credibility, for source credibility source_cred. Table 5 In some embodiments, the source credibility (source_cred) can be weighted by user role; for example, a source from an administrator or responsible person can receive a weighting of +0.10 (up to a maximum of 1.00). Additionally, in some embodiments, a learning mechanism or feedback learning mechanism can be introduced to further adjust the weights of each dimension, making the weights dynamically change to better meet user needs and improve user experience.

[0078] In some embodiments, to improve work efficiency and user experience and reduce unnecessary time spent by users, multiple tasks can be subject to duplication suppression and topic aggregation. In some embodiments, a topic key can be used for cross-source aggregation of the same topic. For example, the main entity can be identified, and task attribution and deadline elements can be determined, as well as text normalization fingerprints can be determined (e.g., using stop word filtering, number normalization, etc.). For multiple tasks after processing, those with completely identical main entity identifiers can be directly deduplicated (e.g., only the latest event is retained). In some embodiments, for text similarity reaching a certain threshold (e.g., cosine similarity >= 0.85) and the same main entity, or for multiple reminders on the same topic within the same time window (e.g., within 2 hours), they can be aggregated into a single task or task card.

[0079] In some embodiments, for aggregated multiple tasks, further field fusion can be performed on the aggregated multiple tasks. For example, for multiple deadlines, the earliest deadline can be taken. Urgency can be recalculated based on the fused deadline. Impact can be recalculated based on the audience union and the highest role bonus (deduplicated counting). Source credibility can be aligned to more credible sources. In addition, a source list (e.g., containing fields such as source_type, url, event_time, etc.) is used to display sources from multiple sources for merging. In some embodiments, suppression strategies can also be adopted. For example, only one task card with the same topic key is retained, or "jitter suppression" (e.g., only one sort refresh within 10 minutes) is used when multiple events of the same topic arrive in a short period of time.

[0080] In some embodiments, this disclosure also provides overload control and obsolescence elimination mechanisms. For example, bucket-based or region-based display is used, and a display task limit is set. In some embodiments, the default display limit for each bucket (top display / current display / later display) can be, for example, 5 / 10 / 10, with any excess collapsed into "More". In some embodiments, this disclosure also provides a time-limited elimination mechanism. For example, cards that exceed the time window and are not triggered by new events are entered into a de-weighting queue, and are automatically hidden (not deleted, but their retrieval entry is retained) after reaching the TTL (Time-To-Live) threshold.

[0081] Figure 9 A block diagram of an apparatus 900 for displaying or interacting with a recommendation task according to some embodiments of the present disclosure is shown. Figure 9 In this device 900, an interactive display layer 901 is included. The interactive display layer 901 may include a workbench 903 for interacting with the user. In some embodiments, the workbench 903 may include an insight report module 905 for generating insight reports based on information or data from the platform. In some embodiments, the workbench 903 may include a follow-up event module 907 for generating follow-up events based on information or data from the platform. In some embodiments, the workbench 903 may also include suggested task cards 909 for displaying sorted recommended tasks. In some embodiments, the workbench 903 may also include a template linkage module 911 for providing task templates to the suggested task cards 909. In some embodiments, by interacting with the workbench 903, users can obtain relevant insight reports, follow-up event updates, and multiple recommended tasks. By interacting with the workbench 903, users can significantly improve their work efficiency, understand the specifics of each task, and complete tasks in a timely manner, greatly enhancing the user experience.

[0082] As described above, one or more embodiments of this disclosure also provide mechanisms regarding permissions, privacy protection, and compliance. The following will describe aspects such as authentication and authorization modes, data scope and permission tracking, audit trails and risk control, privacy and security assurance, and integration with workbench components, thereby ensuring that the workbench adheres to principles of minimal authorization, user visibility, compliance and transparency, and is auditable during cross-product information reading and writing and task linkage.

[0083] Figure 10 A block diagram of an apparatus 1000 for safeguarding information security and permissions according to some embodiments of the present disclosure is shown. Figure 10 In this system, device 1000 may include a security and permission layer 1001. In some embodiments, the security and permission layer 1001 may include a dual-token permission mirroring module 1003 for providing a dual-token mechanism and an audit log module 1005 for auditing workbench logs. In some embodiments, the dual-token permission mirroring module 1003 provides a dual-token mechanism. For example, from a user's perspective, using the user's user_access_token to request data in real time is strictly limited by the permission system of the relevant open platform; document titles that the user is not authorized to view are automatically filtered out. For example, from a project's (administrator's) perspective, the tenant_access_token is used to perform offline aggregation and summary calculations on data within the authorized scope; however, when ordinary members view the summary details, their personal permissions are verified again. Each time a recommendation is generated and parameters are pre-filled, the system or workbench records an audit log, which may include a source_id and an extract_method for post-event auditing and risk control.

[0084] In some embodiments, for sensitive user data (such as personal cloud documents, chat messages, and personal tasks), reading must be based on the user's actual platform permissions. Any data exceeding the user's permissions is invisible and unusable. In some embodiments, for aggregated summaries of project views (issued by the administrator), offline computation can be performed using the tenant_access_token to access the data set authorized by the administrator within the scope explicitly agreed and authorized by the administrator. The issued results follow the constraints of data minimization and the fact that member-side results are consumable but cannot be used to retrieve the original text without authorization. In some embodiments, all tokens are stored encrypted (key escrow and rotation), and access minimization is ensured, meaning that a user's authorization can be revoked at any time, and the authorization immediately becomes invalid and the cache is cleared after revocation. Table 6 shows examples of recommended token usage and boundaries for each information source. Table 6

[0085] In some embodiments, regarding authorization and consent (Consent / Authorization), for initial connections, a prompt can be displayed listing the capabilities, purposes, data types, and use cases requiring authorization, categorized by information source (e.g., for "Following Events" recommendation cards and template pre-filling). In some embodiments, for fine-grained authorization switches, users can individually enable / disable access to IM, Minutes, Drive, and Task, with a default minimum permission set (only reading necessary information, avoiding write permissions unless explicitly enabled by the user). In some embodiments, for authorization revocation and management, authorization for a specific source can be revoked with a single click in the "My Authorizations" panel, and the system simultaneously clears cached tokens and derived indexes. Additionally, administrators can revoke data source authorization for project progress aggregation on the space configuration page. In some embodiments, for notification and guidance regarding unauthorized access, for example, when a user clicks to initiate a task or view details but lacks sufficient platform permissions, a card is displayed in gray with a "no permission" prompt, providing guidance such as "Request permission / Contact administrator."

[0086] In some embodiments, for project progress, for example from an administrator's perspective, a public dataset authorized by the administrator is used to calculate a team-level progress summary and risk alerts, with the results displayed as a read-only summary to space members. In some embodiments, for personal progress, for example from a member's perspective, it is entirely based on the member's own user_access_token, consuming only data visible to the member, and allowing for one-click initiation of task templates.

[0087] Back Figure 10The audit log module 1005, used for auditing workbench logs, can be used for monitoring event trajectories, recommendation generation, parameter tracing, and risk control. In some embodiments, for event trajectory logs, the following fields can be monitored: event_id (event ID), source_type (source type), source_id (source ID), event_time (event time), action_type (action type), actor.user_id (actor.user ID), container_id (container ID), ingest_token_type (extraction token type) (user_access_token / tenant_access_token), permission_check_result (permission check result), and content_hash (content hash). The detailed original text of each field does not need to be stored; only the fingerprint content of each field is stored. These fields will be used for subsequent idempotency processing, deduplication, and post-event traceability.

[0088] In some embodiments, the authorization event log may record consent_id, scopes, granted_by (authorized by...) (user / administrator), grant_time, revoke_time, token_expire_at, refresh_action, and the reason for failure (e.g., missing scope). In some embodiments, one or more implementations of this disclosure may also provide authorization change audit reports and anomaly alerts (e.g., frequent authorization / revocation within a short period, scope expansion, etc.).

[0089] In some embodiments, one or more embodiments of this disclosure may also record recommendation generation records. For example, for each suggested task card, at least one of the following fields may be recorded: card_id (card ID), topic_key (topic key), merged_sources (merged sources), score (rating by dimension: urgency, impact, difficulty, attribution, source credibility), rank_bucket (ranking buckets) (top / current / later / low priority), origin_events (list of referenced event IDs). Recommendation generation records can serve as the data foundation for interpretability and subsequent feedback learning, and are also used for risk control auditing (identifying abnormally high scores or abnormal aggregations).

[0090] In some embodiments, one or more embodiments of this disclosure can also perform template parameter provenance. For example, for each parameter pre-filled for the template, at least one of the following is recorded: param_name, value, source_type (IM / Minutes / Drive / Task), source_id, extract_method (structured / text entity / rule / LLM), confidence (0–1), user_visibility (whether visible to the current user), etc. By retaining parameters and source chains in the template execution log, audit queries and reproducibility are supported.

[0091] In some embodiments, one or more of the present disclosures can also perform risk control. For example, unauthorized access can be blocked. When a token does not match the permissions of the target resource, the access is immediately blocked and flagged; repeated triggers will escalate the risk control level and trigger an alarm. In some embodiments, sensitive information detection can be performed. For example, personally identifiable information (PII) / sensitive words can be detected and anonymized (email, mobile phone number, ID card, customer data, etc.) of the text source. Additionally, administrators can set data classification and anonymization strategies for the project.

[0092] In some embodiments, one or more embodiments of this disclosure can also perform access rate and quota control. For example, following the rate limits of open platforms (referencing IM 50 times / second, Drive 20 times / second, etc.), rate limiting and backoff can be implemented at the source and user levels to prevent overload caused by malicious or accidental operations. In some embodiments, one or more embodiments of this disclosure can also provide tamper protection. For example, audit logs use chained hashing (fingerprint of the previous log entry + fingerprint of the current event) for evidence storage to improve non-repudiation and tamper detection capabilities.

[0093] In summary, one or more embodiments of this disclosure provide a multi-strategy to-do list extraction method based on rules, structured extraction, and LLM assistance. For example, to accurately extract to-do items from different information sources, one or more embodiments of this disclosure provide a multi-level extraction strategy. This strategy prioritizes high-reliability structured extraction and low-cost keyword / sentence rule matching. For complex text where rules are not matched, LLM is invoked for intent recognition and content extraction, and a mechanism is provided to smoothly fall back to the basic strategy when services are unavailable, balancing accuracy, cost, and system stability.

[0094] On the other hand, one or more embodiments of this disclosure also provide a dynamic weighted scoring and ranking model for task recommendation. For example, one or more embodiments of this disclosure provide a dynamic weighted scoring model for ranking task recommendation cards. This model comprehensively considers five core dimensions: urgency (determined by DDL proximity), influence (determined by audience size and role), difficulty (determined by estimated complexity), attribution (relevance to the current user), and source credibility. By setting configurable baseline weights for each dimension and reserving interfaces for the future introduction of online learning mechanisms based on user feedback, dynamic optimization and personalized ranking of recommendation results are achieved.

[0095] On the other hand, one or more embodiments of this disclosure also provide a hybrid incremental extraction mechanism combining event subscription and polling. For example, one or more embodiments of this disclosure also provide an incremental data acquisition strategy that prioritizes event subscription and supplements it with timed polling. This mechanism ensures low latency by subscribing to real-time events of the instant messaging open platform (e.g., im.message.receive_v1, drive.file.edited_v1), while using polling with timestamps or version numbers as fault tolerance and supplementation. Combined with idempotent processing based on event_id and deduplication logic based on the unique identifier of the source data, it ensures near real-time performance, completeness, and consistency of data acquisition.

[0096] On the other hand, one or more embodiments of this disclosure also provide a configurable due date proximity detection window. For example, one or more embodiments of this disclosure provide a DDL proximity detection method based on a multi-level configurable time window (default is 24 hours / 72 hours / 7 days). This method can not only directly extract DDL from structured data such as the due field of a task, but also identify relative dates ("tomorrow"), fuzzy dates ("end of the month"), etc. from unstructured text (such as chat messages, document content) through natural language processing technology, and perform cross-time zone normalization processing, thereby achieving accurate and proactive early warning for time-sensitive tasks.

[0097] On the other hand, one or more embodiments of this disclosure also provide a method for cross-source duplicate information suppression and aggregation based on topic keys. For example, to solve the information overload problem caused by multi-source information, one or more embodiments of this disclosure provide a duplicate suppression and topic aggregation mechanism based on topic_key. The topic_key consists of a main entity identifier, attribution / cutoff elements, and text content fingerprints, which can identify and merge duplicate signals from different channels but pointing to the same topic, aggregating multiple event sources into a recommendation card, while ensuring the stability of the user interface through a jitter suppression strategy.

[0098] On the other hand, one or more embodiments of this disclosure also provide a context-aware intelligent pre-filling technology for template parameters. For example, when a user initiates a task with a "one-click" recommendation card, one or more embodiments of this disclosure can not only recommend the most suitable task template based on the task content, but also automatically extract and fill in the key parameters required for the template from the context of the original information source. This technology minimizes manual input by mapping common fields (such as title, due_date, context_url) and platform-specific IDs (such as mr_id, task_guid).

[0099] On the other hand, one or more embodiments of this disclosure also provide a security isolation and access control method based on a dual-token model and permission mirroring. For example, to ensure data security and privacy, one or more embodiments of this disclosure provide a dual-token domain-based authorization mechanism, that is, using a user_access_token to process user personal data and using a tenant_access_token to process public data digests authorized by the administrator. Simultaneously, strict "permission mirroring" is implemented in the pre-filtering stage of data processing and the post-filtering stage of display, ensuring that users can only see the information they are authorized to access on the source platform, achieving zero unauthorized access.

[0100] On the other hand, one or more embodiments of this disclosure also support auditable parameter tracing and chained hash logs. For example, one or more embodiments of this disclosure can establish an auditable traceability chain for each recommendation and task initiation process. The system not only records the basis for the generation of recommendation cards (referenced event IDs, rating details), but also records the source, extraction method, and credibility of each parameter pre-filled in the template. All audit logs are stored using chained hash technology, ensuring the ability to trace operations after the fact.

[0101] Therefore, based on the aforementioned innovations of this disclosure, one or more embodiments of this disclosure, using a "proactive, two-way seamless" workbench model, connect cross-source incremental extraction, near-DDL detection and To-Do extraction, recommendation engine and one-click template initiation into a closed loop, significantly reducing the user's cognitive and initiation costs, and improving the consumption efficiency of personalized matching and team collaboration. It produces continuous and measurable beneficial effects on some key metrics, addressing issues such as users not remembering / using it effectively, insufficient personalization, and underutilization of collaboration. For example, in terms of initiation rate, through "recommended task cards + intelligent pre-filling of template parameters + one-click initiation," the initiation threshold is significantly lowered, replacing the passive link of "remembering first and then entering prompts." The workbench becomes one of the main initiation entry points for users, driving an overall improvement in the scale and quality of task output.

[0102] For example, regarding activation rates, the insights and events featured on the workbench homepage serve as a "zero-learning-cost" starting point. New users can initiate their first task immediately after registration by clicking on the recommended tasks in the workbench, shortening the initial value experience time. Furthermore, regarding penetration rates, the proportion of template tasks and recommended tasks among user task initiation methods has significantly increased. Different functional sequences can improve usage coverage through differentiated card configurations and scenario guidance, reducing the phenomenon of "only using it in the Q&A area."

[0103] For example, in terms of retention, the workbench continuously aggregates and displays task results, providing actionable steps for the next stage, thus creating an iterative experience. Furthermore, users who utilize the workbench feature show more stable return visits and task initiation on the following day / week / month, resulting in a significant improvement in overall retention.

[0104] Furthermore, one or more embodiments of this disclosure provide a closed-loop consumption scenario. For example, this involves task recommendation, task initiation, and the consumption of results within a collaborative scenario. Specifically, in task recommendation, scenario-based "suggested task cards" are generated based on multi-source incremental events and rules (DDL proximity, To-Do extraction), including source links, deadline elements, and suggested actions. Regarding task initiation, users can "initiate" a template task with their current login permissions by clicking the card, and key parameters are automatically pre-filled. As for the consumption of results within a collaborative scenario, for example, the write-back and aggregation of task results within the "Insight Report / Focus Events" section of the project space workbench encourages the team to further consume the results (further analysis, assignment, and review).

[0105] Furthermore, one or more embodiments of this disclosure enable bidirectional communication between the workbench and other platforms. Results and progress status are visible in the platform's native view, and new changes re-enter the event stream, triggering the next round of recommendations and actions. The closed-loop advantage of this consumption model lies in the fact that result writing back and team consumption link "information → action → result → re-action," forming a continuously driven growth and efficiency loop that directly impacts and improves user initiation rate, penetration rate, and retention rate.

[0106] Figure 11 A block diagram of an electronic device 1100 for processing information according to some embodiments of the present disclosure is shown. Figure 11As shown, device 1100 includes an acquisition module 1110 configured to acquire data from multiple data sources. Device 1100 also includes a first generation module 1120 configured to generate at least one task to be processed based on the data from the multiple data sources. Furthermore, device 1100 includes a display module 1130 configured to display at least one task to be processed and a task processing mode corresponding to each task. Additionally, device 1100 includes a second generation module 1140 configured to generate a first task based on the task processing mode in response to user feedback regarding a first task among the at least one task to be processed.

[0107] It is understood that, in this way, one or more embodiments of device 1100 provide a proactive task suggestion method or system. One or more embodiments of device 1100, through unified event awareness or task mining and prioritization, transform information silos scattered across multiple collaboration tools into a series of pre-emptive tasks that can be initiated with a single click on the user's workbench. This proactively prepares the user for the next action before deadlines or critical events, achieving a leap from passive response to proactive pre-emption, thereby significantly improving the user experience. Furthermore, one or more embodiments of device 1100 significantly reduce the user's cognitive and initiation costs, improve the efficiency of personalized matching and team collaboration, and generate continuous and measurable beneficial technical effects.

[0108] Figure 12 A block diagram of a device 1200 capable of implementing various embodiments of the present disclosure is shown. For example... Figure 12 As shown, device 1200 includes a central processing unit (CPU) and / or a graphics processing unit (GPU) 1201, which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) 1202 or loaded from storage unit 1208 into random access memory (RAM) 1203. Various programs and data required for the operation of device 1200 can also be stored in RAM 1203. The CPU / GPU 1201, ROM 1202, and RAM 1203 are interconnected via bus 1204. Input / output (I / O) interface 1205 is also connected to bus 1204. Although not shown in... Figure 12 As shown, device 1200 may also include a coprocessor.

[0109] Multiple components in device 1200 are connected to I / O interface 1205, including: input unit 1206, such as keyboard, mouse, etc.; output unit 1207, such as various types of monitors, speakers, etc.; storage unit 1208, such as disk, optical disk, etc.; and communication unit 1209, such as network card, modem, wireless transceiver, etc. Communication unit 1209 allows device 1200 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0110] The various methods or processes described above can be executed by CPU / GPU 1201. For example, in some embodiments, the methods may be implemented as computer software programs tangibly contained in a machine-readable medium, such as storage unit 1208. In some embodiments, part or all of the computer program may be loaded and / or installed on device 1200 via ROM 1202 and / or communication unit 1209. When the computer program is loaded into RAM 1203 and executed by CPU / GPU 1201, one or more steps or actions in the methods or processes described above can be performed.

[0111] In some embodiments, the methods and processes described above can be implemented as a computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of this disclosure.

[0112] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example, but not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0113] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, a local area network (LAN), a wide area network (WAN), and / or a wireless network, to an external computer or external storage device. The network may include copper cables, fiber optic cables, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0114] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​and conventional procedural programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to execute the computer-readable program instructions, thereby implementing various aspects of this disclosure.

[0115] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0116] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0117] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0118] This disclosure also provides the following examples:

[0119] Example 1. A method for processing information, comprising:

[0120] Retrieve data from multiple data sources;

[0121] Based on data from multiple data sources, generate at least one task to be processed, wherein the strategy includes a multi-layered strategy for extracting information;

[0122] Multiple tasks are ranked based on a first strategy for task ranking, wherein the first strategy includes interaction patterns with the user.

[0123] Display at least one of the tasks to be processed; and

[0124] The first task is generated in response to user feedback on the first task among the recommended at least one task to be processed.

[0125] Example 2. According to the method described in Example 1, generating at least one task to be processed based on data from multiple data sources includes:

[0126] Based on keyword or sentence structure rules, extract multiple tasks from information from data sources.

[0127] Example 3. The method described in Examples 1-2, wherein generating at least one task to be processed based on data from multiple data sources further includes:

[0128] Parse the information from the data source to determine the first field in the information from the data source; and

[0129] Extract multiple tasks from the first field of the information from the data source.

[0130] Example 4. The method described in Examples 1-3, wherein generating at least one task to be processed based on data from multiple data sources further includes:

[0131] Identify information from data sources using large language models that indicates the need for action; and

[0132] Based on information indicating that action needs to be taken, multiple tasks are extracted.

[0133] Example 5. The method described in Examples 1-4, wherein generating at least one task to be processed based on data from multiple data sources further includes:

[0134] In response to information from the data source, which includes multiple time-related information, the time-related information associated with the information indicating that action needs to be taken is adopted from among the multiple time-related information.

[0135] Example 6. The method according to Examples 1-5, wherein generating at least one task to be processed based on data from multiple data sources further includes at least one of the following:

[0136] In response to a failure to match other information extraction strategies besides the large language model and the information from the data source meeting the first threshold, the large language model is triggered; or

[0137] In response to the unavailability of a large language model or the response time of a large language model exceeding a first time threshold, multiple tasks are extracted based on information from the data source and other strategies for extracting information besides the large language model.

[0138] Example 7. The method according to Examples 1-6, wherein generating at least one task to be processed based on data from multiple data sources further includes at least one of the following:

[0139] Determine the priority of the time window, and

[0140] Based on the information from the data source and the priority of the time window of the information from the data source, extract multiple tasks;

[0141] Based on unstructured information from data sources, time information within the unstructured information is identified through natural language processing or named entity recognition; or

[0142] Based on the strategy used to determine task attribution, the responsible parties for multiple tasks are identified.

[0143] Example 8. The method described in Example 1 further includes:

[0144] Multiple tasks are sorted based on a first strategy used for task sorting.

[0145] Example 9. The method described in Examples 1-8, wherein ranking multiple tasks based on a first strategy for task ranking includes:

[0146] Based on the first strategy, multiple tasks are scored; and

[0147] Based on the scored tasks, the tasks are ranked.

[0148] Example 10. According to the method described in Examples 1-9, scoring multiple tasks based on a first strategy includes:

[0149] Multiple tasks are scored based on at least one of the following criteria: task urgency, task impact, task complexity, task relevance, or the credibility of the task source.

[0150] Example 11. The method according to Examples 1-10, wherein ranking multiple tasks based on a first strategy for task ranking further includes:

[0151] Adjust the order of multiple tasks based on user interaction with them.

[0152] Example 12. The method according to Examples 1-11, wherein the at least one task to be processed includes at least one of the following:

[0153] Based on topic key values ​​or text similarity of multiple tasks, aggregate and sort multiple tasks;

[0154] Key fields in multiple tasks after merging and sorting;

[0155] The folded section contains multiple tasks that are listed later in the sequence; or

[0156] Hide multiple tasks that have exceeded the first waiting time threshold and have not been triggered by any new events after being sorted.

[0157] Example 13. The method according to Examples 1-12, wherein displaying the at least one task to be processed further includes:

[0158] Retrieve task templates associated with multiple tasks;

[0159] Parsing information from a data source that relates to multiple tasks; and

[0160] The parsed information from the data source, which is associated with multiple tasks, is used to populate the relevant fields in the task template.

[0161] Example 14. The method according to Examples 1-13, wherein displaying the at least one task to be processed further includes at least one of the following:

[0162] Record the basis for generating the recommended multi-tasks, including the event identifiers referenced by the multi-tasks or the scores of the multi-tasks.

[0163] Record the source, extraction method, or reliability of the parameters in the task template.

[0164] Example 15. The method according to Examples 1-14 further includes at least one of the following:

[0165] Processing user personal data using a user-accessible token; or

[0166] Use tenant-accessible tokens to process public data that has been authorized by the administrator.

[0167] Example 16. The method described in Examples 1-15, wherein obtaining data from multiple data sources includes at least one of the following:

[0168] Subscribe to event information from the data source through its event subscription interface; or

[0169] Based on the event identifier in the event information, determine whether the event has been processed, and

[0170] In response to the event being processed, refuse to receive event information.

[0171] Example 17. The method described in Examples 1-16, wherein obtaining data from multiple data sources further includes:

[0172] Regularly query the data source through its information access interface to retrieve change information missed due to event loss or subscription delays.

[0173] Example 18. The method described in Examples 1-17, wherein obtaining data from multiple data sources further includes at least one of the following:

[0174] Based on the normalized event model, information from the data source is transformed into information that conforms to the normalized event model;

[0175] Store information from the data source in a database that supports high-concurrency read and write operations;

[0176] Store frequently accessed information from the data source in a cache.

[0177] Archive or clean up cold data from information from the data source; or

[0178] Based on polling cycles or batch processing mechanisms, the access rate to information from the data source is controlled.

[0179] Example 19. The method according to Examples 1-18, wherein generating the first task based on the task processing mode includes any one of the following:

[0180] In response to the existence of a template corresponding to the first task, the first task is generated based on the template corresponding to the first task;

[0181] In response to the absence of a template corresponding to the first task, the first task is generated based on data from multiple data sources associated with it; or

[0182] In response to the absence of a template corresponding to the first task, a request is made to the platform corresponding to another data source to process the first task.

[0183] Example 20. An electronic device comprising:

[0184] The acquisition module is configured to acquire data from multiple data sources;

[0185] A first generation module is configured to generate at least one task to be processed based on the data from the multiple data sources.

[0186] A display module, configured to display the at least one task to be processed, and a task processing mode corresponding to each task to be processed; and

[0187] A second generation module is configured to generate the first task based on the task processing mode in response to user feedback on the first task among the at least one pending tasks.

[0188] Example 21. An electronic device comprising:

[0189] Processor; and

[0190] A memory coupled to a processor, the memory having instructions stored therein, which, when executed by the processor, cause the electronic device to perform a method according to any one of Examples 1 to 19.

[0191] Example 22. A computer program product comprising computer-executable instructions, wherein the computer-executable instructions are executed by a processor to implement a method according to any one of Examples 1 to 19.

[0192] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical applications, or technical improvements to the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A task generation method characterized by comprising: The method comprises: obtaining data from a plurality of data sources; generating at least one pending task based on the data from the plurality of data sources; presenting the at least one pending task and a task processing mode corresponding to each of the pending tasks; generating a first task of the at least one pending task based on the task processing mode in response to a user feedback on the first task.

2. The method of claim 1, wherein generating the first task based on the task processing mode comprises any one of: generating the first task based on a template corresponding to the first task in response to the existence of the template corresponding to the first task; generating the first task based on the data from the plurality of data sources associated with the first task in response to the non-existence of the template corresponding to the first task; or requesting a platform corresponding to another data source to process the first task in response to the non-existence of the template corresponding to the first task.

3. The method of claim 1, wherein generating at least one pending task based on the data from the plurality of data sources comprises: extracting the at least one pending task from the information from the data source based on a keyword or a sentence pattern rule.

4. The method of claim 1, wherein generating at least one pending task based on the data from the plurality of data sources further comprises: parsing the information from the data source to determine a first field in the information from the data source; and extracting the at least one pending task from the first field in the information from the data source.

5. The method of claim 1, wherein generating at least one pending task based on the data from the plurality of data sources further comprises: identifying information indicating a need for action in the information from the data source; and extracting the at least one pending task based on the information indicating the need for action.

6. The method of claim 5, wherein generating at least one pending task based on the data from the plurality of data sources further comprises: adopting a time information associated with the information indicating the need for action in the plurality of time information in response to the information from the data source comprising a plurality of time information.

7. The method of claim 5, wherein generating at least one pending task based on the data from the plurality of data sources further comprises at least one of: triggering the large language model in response to a failure of other strategies for extracting information other than the large language model and the information from the data source satisfying a first threshold; or extracting the at least one pending task based on the information from the data source and other strategies for extracting information other than the large language model in response to the unavailability of the large language model or the response time of the large language model exceeding a first time threshold.

8. The method of claim 1, wherein generating at least one pending task based on the data from the plurality of data sources further comprises at least one of: determining a priority level of a time window, and ​ ​ ​ extracting the at least one pending task based on the information from the data source, a priority level of a time window of the information from the data source; identifying time information in unstructured information in the information from the data source by natural language processing or named entity recognition based on the unstructured information in the information from the data source; or determining a responsible person of the at least one pending task based on a policy for determining task ownership.

9. The method of claim 1, further comprising: ordering the at least one pending task based on a first policy for task ordering.

10. The method of claim 9, wherein ordering the at least one pending task based on a first policy for task ordering comprises: scoring the at least one pending task based on the first policy; and ordering the at least one pending task based on the scored at least one pending task.

11. The method of claim 10, wherein scoring the at least one pending task based on the first policy comprises: scoring the at least one pending task based on at least one of a task urgency, a task impact, a task complexity, a task relevance, or a task source credibility of the at least one pending task.

12. The method of claim 9, wherein ordering the at least one pending task based on a first policy for task ordering further comprises: adjusting the ordering of the at least one pending task based on user interaction with the at least one pending task.

13. The method of claim 1, wherein presenting the at least one pending task comprises at least one of: aggregating the ordered at least one pending task based on a topic key value or a text similarity of the at least one pending task; merging key fields in the ordered at least one pending task; collapsing part of the ordered at least one pending task; or hiding the ordered at least one pending task that exceeds a first wait time threshold and has no new event trigger.

14. The method of claim 1, wherein presenting the at least one pending task further comprises: obtaining a task template associated with the at least one pending task; parsing information in the information from the data source associated with the at least one pending task; and populating the parsed information in the information from the data source associated with the at least one pending task into associated fields in the task template.

15. The method of claim 14, wherein presenting the at least one pending task further comprises at least one of: recording a generation basis of the recommended at least one pending task, the generation basis comprising an event identifier referenced by the at least one pending task or a score of the at least one pending task; recording a source, an extraction method, or a credibility of a parameter of the task template.

16. The method of claim 1, further comprising at least one of: ​ ​ processing user personal data using user-accessible tokens; or processing publicly available data authorized by an administrator using tenant-accessible tokens.

17. The method of claim 1, wherein obtaining data from a plurality of data sources comprises at least one of: subscribing to event information of the data source through an event subscription interface of the data source; or determining whether an event is processed based on an event identifier in the event information, and in response to the event being processed, rejecting the event information, wherein obtaining data from a plurality of data sources further comprises: periodically querying the data source through an information access interface of the data source to obtain change information missed due to event loss or subscription delay.

18. The method of claim 1, wherein obtaining data from a plurality of data sources further comprises at least one of: transforming the information from data sources into information conforming to a normalized event model based on the normalized event model; storing the information from data sources in a database that satisfies high concurrent read and write; storing hot data in the information from data sources in a cache; archiving or purging cold data in the information from data sources; or controlling access rate to the information from data sources based on a polling period or a batch processing mechanism.

19. An electronic device, comprising: an obtaining module configured to obtain data from a plurality of data sources; a first generating module configured to generate at least one pending task based on the data from a plurality of data sources; a presenting module configured to present the at least one pending task and a task processing mode corresponding to each of the pending tasks; and a second generating module configured to generate a first task of the at least one pending task based on the task processing mode in response to a user feedback to the first task.

20. An electronic device, comprising: a processor; and a memory coupled with the processor, the memory having instructions stored therein that, when executed by the processor, cause the electronic device to perform the method of any of claims 1-18.

21. A computer program product comprising computer executable instructions, wherein the computer executable instructions are executed by a processor to implement the method of any of claims 1-18. ​ ​

Images