Group task allocation method, device and equipment based on instant messaging tool and medium

Abstract

This application provides a method, apparatus, device, and medium for group task allocation based on instant messaging tools. The method includes: acquiring group theme information; acquiring group information within the current time window; wherein the group theme information represents the basic information of the group; processing the group information to obtain task semantic entities; processing the group theme information and task semantic entities to obtain a multi-dimensional task list; wherein the task semantic entities represent a set of task elements extracted from the group information and described in a structured form; the multi-dimensional task list represents a multi-dimensional task list containing task content, time constraints, execution entities, and priorities; and sending the multi-dimensional task list to users with corresponding responsibilities based on a preset responsibility library. This method aims to achieve accurate allocation of group tasks based on instant messaging tools, thereby improving team collaboration efficiency.

Description

Technical Field

[0001] This application relates to the field of information processing technology, and in particular to a method, apparatus, device and medium for group task allocation based on instant messaging tools. Background Technology

[0002] In enterprise-level collaboration, cross-departmental and multi-group communication has become the norm. Members can quickly create project group chats using instant messaging tools, and managers can directly issue tasks within the group.

[0003] In existing technologies, task assignment based on instant messaging tools mainly relies on keyword automatic triggering mechanisms. However, this method only mechanically matches preset words and cannot understand the contextual semantics of the conversation. This results in coarse task extraction, difficulty in refining complete task items from chat content, and problems such as task omission and inaccurate sending.

[0004] Therefore, there is an urgent need for a solution for precise allocation of group tasks based on instant messaging tools to improve team collaboration efficiency. Summary of the Invention

[0005] This application provides a method, apparatus, device, and medium for group task allocation based on instant messaging tools, in order to achieve the effect of accurate allocation of group tasks based on instant messaging tools, thereby improving team collaboration efficiency.

[0006] In a first aspect, embodiments of this application provide a group task allocation method based on an instant messaging tool, including:

[0007] Retrieve group theme information; and retrieve group information within the current time window; where group theme information represents the group's basic information;

[0008] The group information is processed to obtain the task semantic entity; the group theme information and the task semantic entity are processed to obtain the multidimensional task list; the task semantic entity represents the set of task elements extracted from the group information and described in a structured form; the multidimensional task list represents a multidimensional task list containing task content, time constraints, execution subject and priority.

[0009] Based on a pre-defined responsibility database, a multi-dimensional task list is sent to the user with the corresponding responsibility.

[0010] In one possible implementation, the group information is processed to obtain a task semantic entity, including:

[0011] The group information is encoded and fused to obtain a task information set; wherein the task information set includes at least one task information; the task information represents data records extracted from the group information that reflect the execution elements of a single task;

[0012] The task information set is aggregated to obtain the task semantic entity.

[0013] In one possible implementation, group information includes group messages, group instructions, and group document content; the group information is encoded and fused to obtain a task information set, including:

[0014] Semantic analysis is performed on group messages to determine group message features; keyword recognition processing is performed on group instructions to determine group instruction features; and group document content is processed to determine group document content features. Among these, group message features represent the semantic information related to task intent in group messages; group instruction features represent the task keyword attributes identified in group instructions; and group document content features represent the task context features extracted from group documents.

[0015] The characteristics of group messages, group instructions, and group document content are fused to obtain a set of task information.

[0016] In one possible implementation, the group document content is processed to determine group document content characteristics, including:

[0017] Based on the summarization algorithm, the content of group documents is processed to generate a summary information; wherein, the summary information represents the summary text that summarizes the content of group documents and contains task instruction information.

[0018] Contextual feature extraction is performed on the summary information to determine the content features of the group documents.

[0019] In one possible implementation, the task information includes the task execution subject, task execution behavior, and task constraints; the task information set is aggregated to obtain a task semantic entity, including:

[0020] For each task information in the task information set, perform similarity processing on the task execution subject to obtain the subject similarity vector;

[0021] Perform similarity processing on the task execution behavior of each task in the task information set to obtain a behavior similarity vector;

[0022] The similarity of the task constraints for each task in the task information set is processed to obtain a constraint similarity vector.

[0023] Based on the subject similarity vector, behavior similarity vector, constraint similarity vector, and preset threshold, the task information in the task information set is aggregated to obtain the task semantic entity.

[0024] In one possible implementation, the group theme information and task semantic entities are processed to obtain a multi-dimensional task list, including:

[0025] The group theme information and task semantic entities are semantically fused to obtain a task list; the task list represents the initial task set generated after fusing the group theme context.

[0026] Feature extraction is performed on the task list to obtain a multidimensional task list.

[0027] In one possible implementation, based on a preset responsibility library, a multi-dimensional task list is sent to the user corresponding to the responsibility, including:

[0028] Based on a multidimensional task list and a preset responsibility library, a responsibility view is generated; the responsibility view indicates the mapping relationship between each task and the direct responsible person, supervisor, and collaborator.

[0029] Based on the responsibility view, a multi-dimensional task list is sent to the user with the corresponding responsibility.

[0030] In one possible implementation, the responsibility view indicates the directly responsible user, the indirectly responsible user, and the collaborating responsible user; based on the responsibility view, a multi-dimensional task list is sent to the user with the corresponding responsibility, including:

[0031] Send the multi-dimensional task list to the directly responsible user;

[0032] Tag the task summaries in the multidimensional task list with a supervisory responsibility label, and send the task summaries with the supervisory responsibility label to the indirectly responsible users;

[0033] Tag the task summaries in the multidimensional task list with collaborative task tags, and send the task summaries with collaborative task tags to the collaborative responsible users.

[0034] In one possible implementation, obtaining group theme information includes:

[0035] Retrieve the initial message history of a group chat on an instant messaging tool;

[0036] Feature extraction is performed on the initial historical messages to obtain feature vectors; where the feature vectors represent the semantic feature vectors in the initial historical messages that reflect the group chat objectives, group rules and member roles;

[0037] Contextual semantic analysis is performed on the feature vectors to obtain the group theme information.

[0038] Secondly, embodiments of this application provide a group task allocation device based on an instant messaging tool, comprising:

[0039] The acquisition module is used to acquire group theme information and group information within the current time window; where group theme information represents the basic information of the group.

[0040] The processing module is used to process group information to obtain task semantic entities; and to process group theme information and task semantic entities to obtain a multidimensional task list. Among them, the task semantic entity represents a set of task elements extracted from group information and described in a structured form; the multidimensional task list represents a multidimensional task list containing task content, time constraints, execution subject and priority.

[0041] The sending module is used to send a multi-dimensional task list to the user with the corresponding responsibility based on a preset responsibility library.

[0042] In one possible implementation, the processing module includes:

[0043] The first processing module is used to encode and fuse the group information to obtain a task information set; wherein the task information set includes at least one task information; the task information represents data records extracted from the group information that reflect the execution elements of a single task;

[0044] The second processing module is used to aggregate the task information set to obtain task semantic entities.

[0045] In one possible implementation, group information includes group messages, group commands, and group document content; the first processing module includes:

[0046] Semantic analysis is performed on group messages to determine group message features; keyword recognition processing is performed on group instructions to determine group instruction features; and group document content is processed to determine group document content features. Among these, group message features represent the semantic information related to task intent in group messages; group instruction features represent the task keyword attributes identified in group instructions; and group document content features represent the task context features extracted from group documents.

[0047] The characteristics of group messages, group instructions, and group document content are fused to obtain a set of task information.

[0048] In one possible implementation, the group document content is processed to determine group document content characteristics, including:

[0049] Based on the summarization algorithm, the content of group documents is processed to generate a summary information; wherein, the summary information represents the summary text that summarizes the content of group documents and contains task instruction information.

[0050] Contextual feature extraction is performed on the summary information to determine the content features of the group documents.

[0051] In one possible implementation, the task information includes the task execution subject, task execution behavior, and task constraints; the second processing module includes:

[0052] For each task information in the task information set, perform similarity processing on the task execution subject to obtain the subject similarity vector;

[0053] Perform similarity processing on the task execution behavior of each task in the task information set to obtain a behavior similarity vector;

[0054] The similarity of the task constraints for each task in the task information set is processed to obtain a constraint similarity vector.

[0055] Based on the subject similarity vector, behavior similarity vector, constraint similarity vector, and preset threshold, the task information in the task information set is aggregated to obtain the task semantic entity.

[0056] In one possible implementation, the processing module includes:

[0057] The third processing module is used to perform semantic fusion processing on the group theme information and the task semantic entities to obtain a task list; wherein, the task list represents the initial task set generated after fusing the group theme context.

[0058] The fourth processing module is used to perform feature extraction on the task list to obtain a multi-dimensional task list.

[0059] In one possible implementation, the sending module includes:

[0060] Based on a multidimensional task list and a preset responsibility library, a responsibility view is generated; the responsibility view indicates the mapping relationship between each task and the direct responsible person, supervisor, and collaborator.

[0061] Based on the responsibility view, a multi-dimensional task list is sent to the user with the corresponding responsibility.

[0062] In one possible implementation, the responsibility view indicates the directly responsible user, the indirectly responsible user, and the collaborating responsible user; based on the responsibility view, a multi-dimensional task list is sent to the user with the corresponding responsibility, including:

[0063] Send the multi-dimensional task list to the directly responsible user;

[0064] Tag the task summaries in the multidimensional task list with a supervisory responsibility label, and send the task summaries with the supervisory responsibility label to the indirectly responsible users;

[0065] Tag the task summaries in the multidimensional task list with collaborative task tags, and send the task summaries with collaborative task tags to the collaborative responsible users.

[0066] In one possible implementation, the acquisition module includes:

[0067] Retrieve the initial message history of a group chat on an instant messaging tool;

[0068] Feature extraction is performed on the initial historical messages to obtain feature vectors; where the feature vectors represent the semantic feature vectors in the initial historical messages that reflect the group chat objectives, group rules and member roles;

[0069] Contextual semantic analysis is performed on the feature vectors to obtain the group theme information.

[0070] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0071] The memory stores the instructions that the computer executes;

[0072] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0073] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0074] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0075] This application provides a method, apparatus, device, and medium for group task allocation based on instant messaging tools. It achieves automated identification, structured generation, and accurate allocation of group tasks by first acquiring group theme information representing the group's basic attributes, simultaneously collecting group chat and announcement information within the current time window, then performing semantic analysis on the group information to extract task semantic entities, and finally automatically pushing tasks to users with corresponding responsibilities by associating and fusing the group theme information and task semantic entities. This method eliminates the need for manual sorting of scattered task information in group chats, automatically uncovers hidden task requirements within the group, and filters irrelevant content by combining group theme information to ensure that the extracted tasks highly match the group's core positioning. The generated multidimensional task list clearly presents the key dimensions of the tasks, facilitating users' quick understanding of task requirements. The automatic distribution mechanism based on the preset responsibility library effectively avoids the problems of incorrect or missed task distribution, significantly improving the efficiency and accuracy of task allocation in instant messaging groups and significantly reducing the cost of manual sorting and communication coordination. Attached Figure Description

[0076] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0077] Figure 1 A flowchart illustrating a group task allocation method based on an instant messaging tool provided in this application embodiment. Figure 1 ;

[0078] Figure 2 A flowchart illustrating a group task allocation method based on an instant messaging tool provided in this application embodiment. Figure 2 ;

[0079] Figure 3 A flowchart illustrating step S202 in a group task allocation method based on an instant messaging tool provided in an embodiment of this application;

[0080] Figure 4 A schematic diagram of a group task allocation device based on an instant messaging tool provided in this application embodiment. Figure 1 ;

[0081] Figure 5 A schematic diagram of a group task allocation device based on an instant messaging tool provided in this application embodiment. Figure 2 ;

[0082] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0083] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0084] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application.

[0085] Team members typically rely on instant messaging tools to create work groups, collaborating and communicating by continuously sending group messages, issuing group commands, uploading project documents, and sharing meeting minutes or requirements materials. In team collaboration or organizational management scenarios, instant messaging tools have become core tools for cross-departmental communication and task allocation. For example, in project management, team members use group chats to discuss requirements, assign sub-tasks, share documents, and provide progress updates; in cross-departmental collaboration, employees with different functions negotiate process details through group chats; and in emergency response scenarios, tasks need to be quickly generated within the group and assigned to appropriate personnel.

[0086] However, existing tools have significant limitations in task allocation: users must manually extract task information from group chats, manually identify the task subject, behavior, and constraints, and manually match the task to a responsibility database, leading to inefficiency and a high risk of errors. Furthermore, group chat messages typically contain a large amount of redundant content, with key task information scattered across different messages, making automatic aggregation difficult. Existing systems cannot understand the task context by combining the semantic meaning of the group chat, resulting in a lack of targeted task allocation.

[0087] Therefore, this application proposes a group task allocation method based on instant messaging tools, which can solve the above problems.

[0088] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0089] Figure 1 A flowchart illustrating a group task allocation method based on an instant messaging tool provided in this application embodiment. Figure 1 ,like Figure 1 As shown, the method includes:

[0090] S101. Obtain the group theme information; and obtain the group information within the current time window; where the group theme information represents the basic information of the group.

[0091] For example, this step can be performed by a group task processing module deployed on the instant messaging platform server. The group task processing module establishes communication connections with the group chat client, message forwarding service, document storage service, and responsibility database. When the target group generates new message interaction content, or when the system detects that the preset polling period has arrived, the processing module first reads the group identifier, group creation time, group member list, group announcement, group tags, and historical session index corresponding to the target group, and retrieves the group's initial historical messages from the message storage module accordingly.

[0092] Group theme information is used to represent the basic information of the group and serves as the contextual basis for subsequent processing tasks, semantic entities, and the generation of multi-dimensional task lists. Specifically, it can be composed of group chat goals, group rules, common discussion objects, member role distribution, long-term matters scope, and business scenario characteristics.

[0093] In one possible implementation, the initial historical messages can be a set of messages within a preset time period after the group is created, or a set of the most recent preset number of historical messages marked as highly relevant. The server performs text cleaning, deduplication, sentence segmentation, stop word filtering, named entity recognition, and contextual analysis on these initial historical messages to extract features that characterize the group's basic semantics. These features include topic descriptions such as "product defect handling group," "version release coordination group," and "after-sales upgrade response group," as well as rule-based information such as deadlines for project leaders to post messages and confirmation of technical issues by R&D contact persons. For member roles, additional information can be provided based on positions, departments, reporting chains, and responsibility tags returned by the organizational structure system. This ensures that the group's main information includes not only topic terms but also constraints and role frameworks. Based on this processing, the group's main information can be organized into structured data objects, with fields including at least the group topic category, long-term business goals, business boundaries, member role mapping, and group-level urgency attributes.

[0094] The group information within the current time window is used to provide real-time group chat input for processing and is the raw data source for generating task semantic entities. This group information can include group messages, group instructions, and group document content. Group messages include text messages, speech-to-text messages, image-recognized text, and quoted replies. Group instructions include reminder instructions, assignment instructions, confirmation instructions, and alert instructions with platform semantic identifiers. Group document content includes the main text of uploaded documents, online document change summaries, meeting minutes attachments, and requirements specifications.

[0095] For example, the system determines the start and end range of the sliding time window based on the current system time, such as using the most recent 10 minutes, 30 minutes, or 1 hour as the current time window. Alternatively, an event-triggered adaptive time window can be used, expanding the time window range when an increase in message density, the addition of key members to the discussion, or the appearance of task keywords is detected, to cover the entire discussion chain. The processing module synchronously obtains all group information within the window through message queue subscription, interface retrieval, or database change monitoring, and retains metadata such as sender identifier, sending time, @mentioned object, reply relationship, message type, attachment link, and message sequence number. To ensure that the information used for subsequent processing is consistent with the current task allocation scenario, the system can also perform out-of-order correction and merge threads within the same session after collection, aggregating referenced replies, continuous supplementary explanations, and document-related messages into the same context fragment.

[0096] In the specific implementation of this step, the acquisition action can be completed through various technical paths. For example, the group chat client can report the message event to the platform server after a user sends a message, and the platform server can then call the group task processing interface to trigger subsequent processes. Alternatively, the server can periodically scan highly active groups and automatically initiate the main topic information update and time window information collection process for groups that meet the processing conditions. Regarding the formation of group main topic information, in addition to semantic analysis of initial historical messages, a topic prior model can be established by combining the group name, group announcements, pinned messages, and administrator configuration items. This topic prior model can then be corrected using historical messages to avoid relying solely on short text tags, resulting in overly coarse basic semantics for the group. For group information within the current time window, the system can also call the document parsing service when receiving document attachments to extract the main text, recognize titles, and segment paragraphs for documents, tables, and online collaborative documents; call the optical character recognition module to extract text from images when receiving them; and call the speech recognition module to generate transcribed text when receiving voice messages. Through the above processing, the group information within the current time window is uniformly converted into a computable standard input format, and a set of messages to be analyzed associated with the group identifier is generated for subsequent semantic extraction tasks.

[0097] By simultaneously establishing the long-term semantic context of the group and the real-time input of the current discussion, subsequent processing is no longer limited to the surface-level literal content of a single message. Instead, it can identify task cues under the joint constraints of group goals, group rules, member roles, and the current discussion chain. This provides an input foundation for solving the technical problems of scattered task semantics, lack of context, and easy misidentification of execution objects in unstructured group chat information. Especially in scenarios with multiple members communicating concurrently and many expressions with separated context, modeling the group's main theme information and the group information within the current time window separately can reduce the probability of misidentifying discussion objects as execution objects and improve the completeness of subsequent task extraction.

[0098] S102. Process the group information to obtain the task semantic entity; process the group theme information and the task semantic entity to obtain a multidimensional task list.

[0099] Among them, the task semantic entity represents a set of task elements extracted from the group information and described in a structured form; the multidimensional task list represents a multidimensional task list that includes task content, time constraints, execution subject and priority.

[0100] For example, a task semantic entity is used to represent a set of task elements extracted from group information and described in a structured form, and is a standardized expression of task content in group information.

[0101] A multidimensional task list is used to represent a multidimensional list of tasks that include task content, time constraints, execution subject, and priority. It is the final task set used for responsibility allocation.

[0102] For example, after receiving group information, differential preprocessing is first performed according to message type.

[0103] For group messages, we perform word segmentation, syntactic analysis, semantic role labeling, intent recognition, and event extraction to identify action words, object words, time limits, responsibility expressions, and constraints.

[0104] For group instructions, keyword recognition, instruction template parsing, and object binding are performed to extract the task execution subject, confirmation status, and urgency intensity.

[0105] For the content of group documents, perform summary generation, heading level parsing, key paragraph extraction, and constraint identification to extract task background, delivery requirements, acceptance criteria, and time nodes.

[0106] After the above processing, a task information set can be generated, in which each task information fragment is associated with at least one message source, one timestamp, and a set of candidate task elements. Task elements include at least one or more of the following: task content, time constraints, executing entity, priority, task object, source message, and evidence fragments.

[0107] In one possible implementation, a natural language understanding model is used to identify whether a message contains a task intent. Messages with task intent are then grouped with their preceding and following related messages to form candidate semantic segments. For example, a message may only contain the phrase "Give me the results before I leave work today." This message, viewed independently, lacks task content. The system can trace back along referencing relationships, consecutive message relationships, or threads on the same topic to obtain the action object for investigating the reason for the new version's login failure, thereby completing the task.

[0108] For individuals mentioned (@), the system doesn't directly identify them as the executor. Instead, it considers verb dependencies, command tone, confirmation statements, and group message information. For example, when discussing a bug, if someone mentions "@testers, please reproduce this," and developers simultaneously check the logs, the system can identify the tester as reproducing the bug and the developer as checking the logs, thus avoiding incorrect assignments based solely on the @ relationship. Regarding time constraints, the system can standardize expressions like "today," "as soon as possible," "before Wednesday," and "before launch," converting them into specific timestamps, time ranges, or urgency levels.

[0109] Regarding priority, the system can calculate the priority by combining the urgency attribute in the group subject information, the mandatory words in the current message, the interval between the deadline and the current time, and the fault level words, forming a three-level or numerical priority field of high, medium and low.

[0110] After forming a task information set, the system aggregates multiple task elements to obtain task semantic entities. Task semantic entities can be carried using a unified data structure, including, for example, task identifiers, task content fragments, constraint sets, candidate execution subject sets, source evidence sets, time expression sets, and confidence values. If multiple related task clues exist within the current time window, the system can cluster these task information fragments based on semantic similarity, object consistency, time proximity, message thread relationships, and group theme matching, thereby forming one or more task semantic entities. If different messages supplement different elements of the same task, they are merged into the same task semantic entity; if multiple parallel events exist within the same time window, multiple task semantic entities are generated separately. To prevent scattered, repetitive, or conflicting information from affecting the completeness of the task description, the system can also perform conflict resolution processing on task semantic entities. For example, when a task has two time limit expressions in different messages, the time limit in the latest confirmed message, or the time limit expression from the group administrator, project leader, or designated contact person, is used as the valid value.

[0111] Subsequently, based on the group theme information and task semantic entities, semantic fusion and aggregation processing are performed to generate a multi-dimensional task list. The group theme information provides contextual constraints and semantic correction at this stage. For example, when the group theme information indicates that the group belongs to an emergency fault response scenario, the task fusion engine will prioritize tasks related to fault location, rollback verification, and customer response, and interpret "as soon as possible" as a shorter timeframe. When the group theme information indicates that the group is a product requirements review group, the system will interpret expressions such as confirming solutions, supplementing PRDs (Product Requirements Documents), and evaluating schedules as requirement process tasks rather than immediate fault handling tasks.

[0112] Specifically, the fusion processing can include context constraint supplementation, semantic reference resolution, task content normalization, execution subject mapping, and unified priority calculation. Context constraint supplementation is used to add background information that is not explicitly present in the task semantic entity but can be inferred from the group theme information into the task entry; semantic reference resolution is used to restore pronouns or omitted expressions such as "this problem," "previous version interface," and "report" mentioned earlier to explicit objects; task content normalization is used to convert natural language tasks into a unified verb-object structure; execution subject mapping is used to map colloquial titles to responsibilities, roles, or specific personnel; and unified priority calculation quantifies the urgency expressions from different messages on the same scale.

[0113] In one possible embodiment, the multidimensional task list consists of multiple task entries, each of which contains at least one or more of the following information: task content, time constraints, executing entity, and priority. It may further include task source, collaborating personnel, supervising personnel, status flags, business tags, and confidence values.

[0114] For example, a task entry can be represented as: "Task content: Investigate the reasons for mobile login failures; time constraint: before 18:00 on the same day; executing entity: client development manager; priority: high." Another task entry can be represented as: "Task content: Organize customer feedback and simultaneously fix the affected area; time constraint: before 10:00 on the next day; executing entity: after-sales contact person; priority: medium-high." When generating a multi-dimensional task list, the system can also perform deduplication, sorting, and completeness checks on task entries. When task content is similar, the executing entity is the same, and the time constraints are close, they can be merged into the same entry; when a task lacks an executing entity, the missing field can be temporarily retained and filled in by the subsequent responsibility matching process; when a task lacks a clear time constraint but the group subject information contains an urgent attribute, a default urgency level can be assigned according to preset rules.

[0115] This step is crucial for solving the technical problem of this application. Its overall implementation principle lies in transforming unstructured task cues scattered across group messages, instructions, and documents into a computable, correctable, and distributable structured task representation through semantic analysis, keyword recognition, summary generation, cross-message aggregation, and context fusion. Traditional solutions often remain at the level of local keyword matching or static template matching, making it difficult to handle omissions, additions, and role switching in multi-turn dialogues. This application's embodiment first generates task semantic entities, then jointly processes these entities with group theme information, unifying the expression of task content, time constraints, execution subject, and priority. This achieves a semantic leap from fragmented discussion content to executable task items. Based on the above analysis, this step significantly improves the completeness, consistency, and contextual fit of task identification, reduces the probability of task fragmentation, information omissions, and semantic conflicts, and provides a reliable structured foundation for subsequent accurate allocation.

[0116] S103. Based on the preset responsibility library, send the multi-dimensional task list to the user with the corresponding responsibility.

[0117] For example, a preset responsibility database is used to provide a basis for matching user responsibilities for task distribution. Users are the recipients of the multi-dimensional task list, i.e., the task assignees corresponding to the responsibility database. The preset responsibility database can store the correspondence between users and responsibilities and is updated in conjunction with the organizational structure system, project configuration system, and permission system. Data fields in the responsibility database may include user identifier, job title, department affiliation, responsibility type, business line affiliation, responsible module, manageable task categories, collaboration relationship, supervision relationship, on-call status, and availability status. To ensure that the distribution results match the actual collaboration relationships, the responsibility database can also simultaneously store static and dynamic responsibility mappings. Static responsibility mappings correspond to job or department responsibilities, while dynamic responsibility mappings correspond to on-call schedules, project rotations, fault response personnel, current version manager, or temporary project members.

[0118] Optionally, the preset responsibility library includes PM responsibilities, RTE responsibilities, development responsibilities, and testing responsibilities. Understandably, PM responsibilities include product decisions, requirement changes, project progress, risk identification, and attention to decision milestones, with a primary focus on project progress, risks, and decision milestones. RTE responsibilities include technical tasks, development tasks, testing tasks, technical bottlenecks, and unresolved issues, with a primary focus on project progress, technical dependencies, and risks. Development responsibilities include coding tasks, code review, technical implementation, and bug fixing, with a primary focus on project progress and technical dependencies. Testing responsibilities include testing tasks, defect reporting, test coverage, and quality acceptance, with a primary focus on project progress and technical dependencies.

[0119] Based on this responsibility database, the system can match the task content, time constraints, execution subject, and priority in the multi-dimensional task list with the user's responsibilities, thereby determining the user who should receive the task.

[0120] In practice, after analyzing the multi-dimensional task list item by item, the system reads the task content, time constraints, execution subject, and priority fields of each task item. If the task item already contains a clear execution subject, the system first determines whether the execution subject is a specific user identifier. If so, it verifies the user's current valid responsibilities, on-duty status, and permission scope in the responsibility database. If the verification passes, the user is directly identified as the user with the corresponding responsibility. If the execution subject is a role name, job title, or colloquial terminology used in a group chat, such as "backend developer," "test manager," or "operations contact person," the responsibility matching engine searches the responsibility database for the user set under the corresponding responsibility role. Then, it combines the business line, project affiliation, on-duty status, and current workload to determine the final receiving user. If the task item does not contain a clear execution subject, the system reverse-engineers the responsibility based on the task content and time constraints. For example, fixing payment callback anomalies is matched to the payment service development responsibility, confirming user announcement text is matched to the operations responsibility, and compiling customer complaint details is matched to the after-sales responsibility. For high-priority and cross-departmental tasks, the system can simultaneously identify the directly responsible user, the collaborating user, and the supervising user, and generate task views with different focus areas for different users. The directly responsible user receives the complete task execution information, the collaborating user receives the assistance information, and the supervising user receives the follow-up reminder.

[0121] After identifying users with corresponding responsibilities, a multi-dimensional task list is sent to the appropriate recipients. Sending methods can include generating task cards and pushing them to specific users within the group interface of an instant messaging tool, sending task messages to user-specific chat sessions, writing task records in the task center of the workbench, triggering in-system notifications, SMS reminders, or email reminders. If the task has a high priority, the system can use a multi-channel concurrent reminder mechanism and set up receipt confirmation logic; if the task has a lower priority, a single-channel notification is used and the delivery status is recorded. In addition to the task item itself, the sent data can also include the name of the task source group, a summary of task evidence, generation time, deadline, responsibility description, collaborative relationships, and a processing entry link, so that users can quickly understand the task background and enter the execution page after receiving the task. For closed-loop management, the system can also write a task distribution log after successful sending, recording the task identifier, receiving user, sending time, sending channel, delivery result, and confirmation status, and write the task status back to the group task management table. If the user does not confirm receipt within a preset time limit, the system can automatically add reminders or escalate the distribution based on the supervisory relationships in the responsibility database.

[0122] In one possible embodiment, the preset responsibility database can also support a multi-level matching strategy. The first level performs explicit matching based on the execution subject field; the second level performs semantic matching based on the similarity between task content and responsibility tags; the third level filters currently on-duty users with responsiveness requirements based on time constraints and priorities; and the fourth level selects users with a high success rate in processing the same type of task based on historical processing records. Through multi-level matching, the system can still determine the appropriate recipient even when the execution subject is unclear, the expression is colloquial, or the current responsible person is temporarily unavailable.

[0123] Based on the above analysis, by linking a structured, multi-dimensional task list with a pre-defined responsibility database, task delivery no longer relies on simple @mention relationships or fixed job templates within groups. Instead, it comprehensively considers task content, time constraints, the executing entity, and priority to ensure tasks are sent to users with the corresponding responsibilities. This step, together with the preceding steps, forms a complete processing chain, transforming fragmented expressions in group chats into actionable distribution results tailored to the responsible parties. This improves the accuracy, timeliness, and traceability of task delivery, while reducing the costs of manual organization and secondary relaying.

[0124] This application provides a group task allocation method based on instant messaging tools. It first acquires group theme information representing the group's basic attributes, and simultaneously collects group chat and announcement information within the current time window. Then, it performs semantic parsing on the group information to extract task semantic entities. The group theme information and task semantic entities are then correlated and fused to generate a structured multi-dimensional task list. Finally, it automatically pushes tasks to users with corresponding responsibilities by matching them to a preset responsibility library. This method achieves automated identification, structured generation, and accurate allocation of group tasks. This method eliminates the need for manual sorting of scattered task information in group chats, automatically uncovers hidden task requirements within the group, and filters irrelevant content by combining group theme information, ensuring that the extracted tasks highly match the group's core positioning. The generated multi-dimensional task list clearly presents the key dimensions of the tasks, facilitating users' quick understanding of task requirements. The automatic distribution mechanism based on the preset responsibility library effectively avoids the problems of incorrect or missed task distribution, significantly improving the efficiency and accuracy of task allocation in instant messaging groups and significantly reducing the cost of manual sorting and communication coordination.

[0125] Figure 2 A flowchart illustrating a group task allocation method based on an instant messaging tool provided in this application embodiment. Figure 2 ,like Figure 2 As shown, in this embodiment... Figure 1 Based on the embodiments, a group task allocation method based on instant messaging tools is described in detail. The method includes:

[0126] S201. Obtain the group topic information; and obtain the group information within the current time window.

[0127] Among them, the group theme information represents the group's basic information.

[0128] For example, this step can be referred to step S101, and will not be described again.

[0129] In one example, the initial historical messages of a group chat on an instant messaging tool are retrieved; feature extraction is performed on the initial historical messages to obtain feature vectors; and contextual semantic analysis is performed on the feature vectors to obtain the group theme information.

[0130] Among them, the feature vector represents the semantic feature vector in the initial historical messages that reflects the group chat target, group rules and member roles.

[0131] For example, the initial history messages of a group chat are used as the raw input for the group's main message. These can typically be the first messages after the group is created, the group creation announcement, pinned content, or early conversations directly related to the purpose of creating the group.

[0132] Feature vectors are used to characterize the semantic features of the initial historical messages that reflect the group chat objectives, group rules, and member roles. They are usually obtained by jointly encoding keywords, semantic relationships, instruction expressions, and identity titles in the messages using a text encoding model.

[0133] The group theme message is used to carry the abstract results of the basic information of the group chat. Its content may include the group's business goals, discussion boundaries, role division, collaboration constraints, and focus.

[0134] For example, after a group chat is created, the system can preset a historical data collection window, call the message interface to obtain the first batch of text messages, system announcements, group nickname descriptions, and member speech records within that window, and perform word segmentation, noise reduction, entity recognition, and sentence vector encoding on the text content. For content reflecting the group chat's objectives, the system can identify task intent words, project names, business themes, and collaboration directions; for content reflecting group rules, the system can extract time limits, speech constraints, data sharing rules, or approval requirements from the group rules; for content reflecting member roles, the system can identify role markers such as person in charge, contact person, collaborator, and reviewer, and their corresponding members. After the above semantic information is fused to form a feature vector, the feature vector is subjected to contextual semantic analysis. Combined with the referential, supplementary, and constraint relationships between adjacent messages, association reasoning is performed to output the group's main theme information. This analysis can be implemented using attention encoding, semantic aggregation, or context classification models. The model parameters can be obtained by training based on historical group chat samples. In practical applications, other models can also be selected, and this embodiment does not limit this.

[0135] Through the above processing, the system can reliably extract the core message of the group from the early unstructured text of the group chat, avoiding biased judgments based on a single message, and enabling subsequent task identification, responsibility matching, and task allocation to be based on a more accurate group semantic foundation. Because the group message information integrates three semantic features—goal, rules, and member roles—it has a stronger ability to interpret subsequent group messages, reducing the probability of misjudgments caused by contextual omissions, role confusion, and unclear task boundaries, thereby improving the overall accuracy and consistency of group task processing.

[0136] S202. Encode and fuse the group information to obtain a task information set; aggregate the task information set to obtain task semantic entities.

[0137] The task information set includes at least one task information; the task information represents data records extracted from the group information that reflect the execution elements of a single task.

[0138] For example, group information is used as the input data source for task semantic extraction, and it may include group messages, group instructions, and group document content. Group messages may be received by an instant messaging client and synchronized to the server, group instructions may be explicit execution requests issued by members within the group, and group document content may be meeting minutes, requirement specifications, or supplementary materials.

[0139] The task information set is used to carry at least one task information extracted from the group information. The encoding fusion process is used to map information from different sources, different forms of expression and different time segments into a unified intermediate representation, so that subsequent aggregation can be carried out in a unified semantic space.

[0140] Task information is a data record used to characterize the execution elements of a single task. It can correspond to any one or more of the following: execution subject, execution behavior, time constraint, object scope, or constraint conditions.

[0141] Task semantic entities are used to structure the results of aggregated task information sets, so that scattered task clues can be formed into a unified semantic object that can be called upon for subsequent task list generation.

[0142] For example, the system first preprocesses the group information to unify and standardize the message body, attachment text, quoted content, and contextual content, and then vectorizes the content of different modalities. The encoding fusion process can combine textual semantic encoding with contextual feature encoding to associate related expressions in the same group chat with the same semantic representation, thus forming a task information set containing multiple task information. Each task information corresponds to a fragment of a local task element in the group information, such as a responsible person mentioned in a message, a deadline specified in a paragraph, or the execution content described in a document fragment. Subsequently, the system performs aggregation processing on the task information set, merging the scattered elements of the same task based on semantic similarity, contextual dependencies, and temporal relationships, and eliminating duplicate, conflicting, or ambiguous information, ultimately generating a task semantic entity. This task semantic entity can serve as the basic structure for subsequently generating a multidimensional task list, containing a combination of task subjects, task actions, and constraints.

[0143] In practical applications, the encoding model used for encoding fusion processing can be a pre-trained language model, a lightweight semantic encoder, or a sequence encoding network with a fused context window, depending on the deployment environment, to ensure extraction accuracy while meeting the real-time requirements of group chat scenarios. Other models can also be selected in practical applications, and this application embodiment does not limit this choice. The aggregation processing of the task information set establishes a task association graph based on message order, member reply relationships, and document reference relationships. Then, the node representations in the graph are merged and compressed to output a unified task semantic entity. Through this method, the originally discrete, scattered, and potentially omitted task content within the group can be integrated into a structured entity, thus providing stable input for subsequent task content recognition, time constraint parsing, and responsibility matching.

[0144] By adopting this implementation method, the system can transform scattered task clues in group information into unified task semantic entities, reducing information omissions caused by context jumps and incomplete expressions. Since the encoding fusion process can simultaneously absorb multi-source group information, and the aggregation process can resolve repetitive and conflicting semantics, the task recognition results are more complete and the semantic boundaries are clearer. Furthermore, this task semantic entity can provide a reliable foundation for the subsequent generation of multi-dimensional task lists, thereby improving the accuracy, continuity, and automation of group task allocation, and reducing the cost of manually organizing group messages.

[0145] Figure 3 A flowchart illustrating step S202 of a group task allocation method based on an instant messaging tool provided in this application embodiment is shown below. Figure 3 As shown, step S202 includes:

[0146] S2021. Perform semantic analysis on group messages to determine group message features; perform keyword recognition processing on group instructions to determine group instruction features; process group document content to determine group document content features; and fuse group message features, group instruction features, and group document content features to obtain a task information set.

[0147] Group information includes group messages, group instructions, and group document content. Group message features characterize semantic information related to task intent within group messages; group instruction features characterize task keyword attributes identified within group instructions; and group document content features characterize task context features extracted from group documents. For example, group message features characterize semantic components in group messages that embody task initiation, task requirements, execution constraints, or collaborative intent; group instruction features characterize keyword attributes in group instruction text related to responsible persons, execution actions, time limits, or objects; and group document content features characterize background information, supplementary conditions, and contextual information extracted from documents, minutes, or attachments. Group messages, group instructions, and group document content can originate from different types of information generated within a set time window in the same group chat session, or from historical messages, group command text, and shared documents under the same task topic.

[0148] For example, the system can first perform word segmentation, part-of-speech tagging, and syntactic dependency analysis on group messages, and then combine a pre-defined task semantic lexicon and intent recognition model to extract task trigger statements, action predicates, and constraint expressions from the messages, encoding them into vector-based group message features. For group instructions, keyword matching, entity recognition, and rule constraints can be used to identify task keywords such as responsible person, @mention objects, deadlines, and requests for execution, mapping them to structured attributes to form group instruction features. For group document content, title parsing, summary extraction, paragraph encoding, and context window aggregation can be used to extract content representations corresponding to task background, scope of application, dependencies, and supplementary explanations, forming group document content features.

[0149] After obtaining the three types of features, the system can use feature concatenation, attention weighting, or gating fusion to jointly encode the group message features, group instruction features, and group document content features, generating a unified-dimensional task information set. This task information set can preserve the task semantics, execution constraints, and contextual background scattered throughout the group chat, transforming the originally unstructured group information into an input representation that can be used for subsequent task semantic entity construction and responsibility matching. In specific model deployment, the encoding model used can be a text vector encoder, a pre-trained language model, or a lightweight semantic representation module. In practical applications, other models can also be selected for this component; this application embodiment does not limit this, but its function is to improve the semantic alignment capability of heterogeneous group information.

[0150] By extracting group messages, group instructions, and group document content into different types of features and then fusing them, we can avoid information loss and semantic bias caused by a single text source, improve the completeness of the task information set in covering the task intent and the consistency of the context, thereby providing more accurate basic data for subsequent task identification, task aggregation, and responsibility-oriented allocation, and reducing task omissions and misjudgments caused by scattered expressions or document supplementation.

[0151] In one possible implementation, the group document content is processed to determine the group document content features, including: generating a summary of the group document content based on a summarization algorithm to obtain summary information; and extracting contextual features from the summary information to determine the group document content features.

[0152] Among them, the summary information represents the summary text that summarizes the content of the group documents and contains task instruction information.

[0153] For example, the content of a group document can be meeting minutes, requirements specifications, solution documents, notices, or collaboratively edited text uploaded in a group chat. The content of the group document is used to carry descriptive information related to the task.

[0154] Group document content features are used to characterize the semantic attributes in a document that are related to the task context, and can reflect the task object, execution requirements, time constraints and associated background.

[0155] Summary information is used to compress and express the content of group documents, so that the retained summary text focuses on task instructions, key objects and constraints, so as to extract context clues from the summary later.

[0156] For example, after receiving the content of a group document, the system first segments, denoises, and performs syntactic segmentation on the document body, and then constructs the input text by combining the title, the first sentence of the body, bolded content, list content, and time information. The summarization algorithm can employ a text summarization model based on sequence coding, an extractive summarization model based on attention mechanisms, or a generative summarization model that integrates rule constraints to filter out core sentences related to the task from the full text and generate summary information. This summary information retains the task instructions in the original text, such as "to be completed within this week," "to be followed up by R&D colleagues," and "to be submitted in a revised version," so that the summary text retains task identification value while compressing its length. To improve the quality of the summary, the summary results can also be processed by removing repeated sentences, retaining key entities, and enhancing time-limited sentences, thereby avoiding the loss of execution subject or deadline information due to excessive compression. In practical applications, the summarization algorithm can also choose other models, which are not limited in this embodiment.

[0157] After obtaining the summary information, the system further performs contextual feature extraction processing. This processing, based on word vector encoding, semantic role analysis, dependency parsing, and context window matching, identifies the task subject, action predicate, recipient, time constraints, and conditional limitations in the summary, and converts them into vectorized features or structured labels. Subsequently, the system correlates and fuses the extracted semantic features with the topic tags of the group to which the group document belongs, historical task records, and the current discussion context within the group to obtain group document content features that can characterize the continuity of the task background. These features can serve as input for subsequent task semantic entity aggregation, responsibility matching, or task list generation, enabling the implicit task intent in the document to be stably identified at the summary level.

[0158] This processing method compresses long group documents into high-information-density text through a process of summarizing first and then extracting features. Context-related features are then extracted from the summary, reducing the interference of noisy sentences on task recognition. Because the summary information retains task instructions, subsequent feature extraction can more effectively capture execution requirements and constraints, thereby improving the completeness and accuracy of group document content features and providing a more reliable semantic basis for group task allocation.

[0159] S2022. Perform similarity processing on the task execution subject of each task information in the task information set to obtain a subject similarity vector; perform similarity processing on the task execution behavior of each task information in the task information set to obtain a behavior similarity vector; perform similarity processing on the task constraint conditions of each task information in the task information set to obtain a constraint similarity vector; based on the subject similarity vector, behavior similarity vector, constraint similarity vector, and preset threshold, perform aggregation processing on the task information in the task information set to obtain the task semantic entity.

[0160] The task information includes the task execution entity, task execution behavior, and task constraints.

[0161] For example, task information is used to represent a single task data record extracted from the group chat content, task execution subject is used to represent the execution object corresponding to the task, task execution behavior is used to represent the action or operation that the task needs to perform, and task constraints are used to represent the time limit, scope, dependency relationship or other restrictive requirements of the task.

[0162] The task information set is a collective representation of multiple candidate task information items within the same time window. Its aggregation is based on recognizing the semantic consistency of these task information items in terms of subject, behavior, and constraint. Subject similarity vector, behavior similarity vector, and constraint similarity vector are used to express the degree of similarity between different task information items in the three types of elements, respectively. The preset threshold is used to limit the boundaries that can be merged into the same task semantic entity.

[0163] In its implementation, the system first performs semantic encoding on the content of group messages, group instructions, and group documents to obtain several task information items. Each task information item includes at least the task execution subject, the task execution behavior, and the task constraints.

[0164] Subject similarity processing can be achieved based on vector cosine similarity, word embedding distance, or entity alias mapping results, assigning higher similarity to subjects identified as pointing to the same responsible person, the same department, or the same role.

[0165] Behavioral similarity processing can be achieved through verb standardization, synonym expression normalization, and action category mapping, such as mapping processing, repair, and investigation to similar actions.

[0166] Constraint similarity processing can uniformly encode temporal representations, priority representations, scope limitations, and dependency conditions to determine whether different task information has a consistent execution boundary. After obtaining three types of similarity vectors, the system can compare them with preset thresholds respectively, and use weighted fusion, similarity cluster generation, or hierarchical clustering to complete aggregation, thereby merging multiple task information that are semantically repetitive, semantically supplementary, or semantically scattered but point to the same task into a single task semantic entity.

[0167] This aggregation process can unite and merge the scattered execution objects, action requirements, and constraints in a group chat without relying on a single keyword hit, thus forming a unified structured result from multiple sources of the same task. Because the subject, behavior, and constraint are all involved in similarity judgment, the system effectively avoids erroneously merging information that is only partially similar but substantially different in task, while also incorporating supplementary descriptions of the same task into the same task semantic entity. Therefore, when generating a multi-dimensional task list based on the task semantic entity, more complete task content, time constraints, and responsibility assignments can be obtained, thereby improving the completeness of task identification, the accuracy of aggregation, and the executability for responsibility allocation.

[0168] S203. Perform semantic fusion processing on the group theme information and task semantic entities to obtain a task list; perform feature extraction processing on the task list to obtain a multi-dimensional task list.

[0169] The task list represents the initial set of tasks generated after the fusion group's thematic context.

[0170] For example, group theme information is used to characterize the business topic, collaborative goal, or discussion background of the group chat, while task semantic entities are used to characterize the task expression with structured attributes extracted from the group information. After the two are semantically associated, the system can unify the task clues scattered in the group messages into the same context for understanding.

[0171] The task list is used to hold the initial set of tasks after fusion. Its content includes both explicit tasks and implicit tasks completed by the group theme, so that subsequent processing is no longer limited to isolated statements, but forms a continuous task semantics around the theme within the group.

[0172] In its implementation, the system first performs semantic vectorization on the group's main information, mapping the task subject, task actions, constraints, and time expressions in the task semantic entity into a unified semantic representation. Then, based on similarity matching, contextual attention weights, or rule constraints, it completes semantic fusion to generate a task list. This task list can be stored in a structured record format, with each record corresponding to an initial task item and associated with the source message, trigger time, and contextual confidence level, in order to preserve the original context of the group chat.

[0173] After the task list is formed, the system further performs feature extraction processing to extract task content features, timeliness features, responsibility features, priority features, and dependency features, and organizes these features into a multidimensional task list. The multidimensional task list can be stored using vector representation, field tables, or graph structures for subsequent responsibility matching and task distribution.

[0174] In practical applications, the system can correct the group theme information based on pre-trained language models, domain dictionaries, and group chat history, thereby improving the consistency between the fusion results and the business context. When there is ambiguity between the task semantic entities and the group theme information, the system can rearrange or complete the task list according to the strength of contextual association, making the task items closer to the actual collaborative intent within the group. The semantic fusion module and feature extraction module used can also be configured with other models or computing power in practical applications; this application embodiment does not limit this. Through this processing method, the system can transform scattered task information in group chats into a task list with multi-dimensional attributes, thereby improving the completeness of task expression, contextual consistency, and subsequent allocation accuracy, and reducing omissions and misclassifications caused by semantic fragmentation.

[0175] S204. Generate a responsibility view based on the multidimensional task list and the preset responsibility library; based on the responsibility view, send the multidimensional task list to the users with the corresponding responsibilities.

[0176] The responsibility view indicates the mapping relationship between each task and the person directly responsible, the supervisor, and the collaborators.

[0177] For example, the multidimensional task list is a collection of tasks to be assigned formed from group chat semantic processing, used to carry information such as task content, time constraints, execution entity, and priority. The preset responsibility library stores job responsibilities, personnel affiliations, and task acceptance rules within the organization, providing a matching basis for task assignment. The responsibility view establishes a correspondence between tasks and directly responsible persons, supervisors, and collaborators, so that task distribution no longer relies solely on a single job label, but rather forms a multi-role association mapping based on task attributes.

[0178] In practical implementation, after receiving a multi-dimensional task list, the system first compares the execution subject, action attributes, constraints, and priority fields of each task with the responsibility entries in the preset responsibility library, and then calls the responsibility matching engine to generate a responsibility view. This view can be implemented by a relational mapping table, where each row corresponds to a task node and records the identifiers of the direct responsible person, supervisor, and collaborators associated with the task, as well as the corresponding department, group member number, and available terminal information. The responsibility entries in the responsibility library can be pre-configured according to the organizational structure, project division of labor, and business processes, and can be stored using database tables, key-value indexes, or graph structures to improve task retrieval and mapping speed. In practical applications, the responsibility library can also choose other storage formats, which are not limited in this application.

[0179] After generating the responsibility view, the system distributes and controls the multi-dimensional task list according to the mapping relationship in the view, sending the same task to one or more associated user terminals. Different message content can be set according to role differences when sending: the person directly responsible receives complete task information, the supervisor receives information including execution requirements and schedule constraints, and collaborators receive sub-task information related to the collaborative content, thus ensuring that users with different responsibilities receive task content matching their roles. Message distribution can be completed through enterprise instant messaging interfaces, in-site notification interfaces, or mobile push interfaces, and can include task numbers, deadlines, related group message references, and receipt entry points.

[0180] Optionally, after sending the multi-dimensional task list to the users with the corresponding responsibilities, you can continuously monitor user behavior and maintain the multi-dimensional task list. For example, when a user replies with confirmation words such as "completed," "received," or "confirmed," the progress of the multi-dimensional task list is updated; when the same task is mentioned multiple times, a new task point is added based on the current multi-dimensional task list; when modifications or supplementary explanations are made to the task content in a group, additions or modifications are made based on the current multi-dimensional task list.

[0181] Through the above method, the system associates the multi-dimensional task list with the preset responsibility database to form a responsibility view, and then completes targeted sending based on this view, ensuring that the task allocation results are consistent with the organization's responsibilities. This method can reduce manual verification and repeated forwarding, improve task flow efficiency, and reduce omissions, mis-sending, and execution deviations caused by responsibility mismatch, thereby improving the accuracy and traceability of group chat task processing.

[0182] In one example, a multidimensional task list is sent to the directly responsible user; a supervisory responsibility label is added to the task summary in the multidimensional task list, and the task summary with the supervisory responsibility label is sent to the indirectly responsible user; a collaborative task label is added to the task summary in the multidimensional task list, and the task summary with the collaborative task label is sent to the collaborative responsible user.

[0183] The responsibility view identifies directly responsible users, indirectly responsible users, and collaboratively responsible users.

[0184] For example, the responsibility view is used to represent the mapping relationship between tasks and different responsibility roles, and can indicate the direct responsible user, indirect responsible user, and collaborating responsible user for each task.

[0185] The directly responsible user receives a complete multi-dimensional task list to have overall control over the task content, time constraints, execution entity, and priority.

[0186] Indirect responsibility users are used to receive task summaries with supervisory responsibility tags. The supervisory responsibility tags are used to identify that the task corresponding to the summary has supervision, tracking or auditing attributes.

[0187] Collaborative responsible users are used to receive task summaries with collaborative task tags. The collaborative task tags are used to identify that the task corresponding to the summary requires cross-member cooperation, resource collaboration, or information linkage.

[0188] In its implementation, the system generates task summaries based on task elements in a multi-dimensional task list, determines responsibility allocation results by combining them with a preset responsibility library, and then writes the responsibility view into the task distribution module. The task summary can be compressed from task content, time requirements, responsible parties, and priority fields, facilitating the carrying of differentiated information under different responsibility roles. For task summaries requiring supervision, the system can attach a supervision responsibility tag to the summary text, metadata record, or message encapsulation header and push it to the receiving end corresponding to the indirect responsible user; for task summaries requiring collaboration, a collaboration task tag is attached before pushing it to the receiving end corresponding to the collaboration responsible user. This tag can be implemented using text markers, structured fields, or message attributes. In practical applications, other encoding methods can also be selected for this tag, which is not limited in this embodiment.

[0189] By distributing the same multi-dimensional task list in a differentiated manner according to the responsibility role through a responsibility view, users with direct responsibility receive the complete task, users with indirect responsibility receive supervisory guidance information, and users with collaborative responsibility receive collaboration guidance information. This allows different roles to quickly locate the required content based on their own responsibilities. Because supervisory responsibility tags and collaborative task tags identify task summaries by role, the system can reduce interference from irrelevant information, improve the relevance and executability of task communication, and reduce the probability of task omissions, misreceipts, and duplicate confirmations in group chat environments.

[0190] This application provides a group task allocation method based on instant messaging tools. The method first acquires group theme information representing basic group information, then simultaneously collects group interaction information within the current time window. Next, it sequentially encodes and fuses the group information to obtain a task information set, aggregates the task information set to generate task semantic entities, and semantically fuses the group theme information with the task semantic entities to generate a task list and extracts features to obtain a multi-dimensional task list. Finally, it generates a responsibility view based on the multi-dimensional task list and a preset responsibility library, and pushes tasks to users with corresponding responsibilities. This fully automated process achieves the beneficial effects of eliminating the need for manual input, seamlessly extracting and structuring tasks from natural language interactions in group chats, and automatically matching personnel responsibilities for accurate allocation. It effectively solves the problems of traditional group task allocation relying on manual sorting, low efficiency, and easy omissions and mismatches, significantly improving the task flow efficiency of team collaboration in instant messaging scenarios.

[0191] This application also provides a group task allocation method based on an instant messaging tool, which archives the data of the entire group task allocation cycle and indicates historical queries; if a task is not completed in a closed loop, it is inherited from the daily report to the weekly report, and then from the weekly report to the monthly report, so as to continuously monitor the progress status of the task.

[0192] Optionally, version records can be maintained for task semantic entities to support semantic backtracking and semantic consistency auditing.

[0193] Optionally, the effectiveness of task allocation can be evaluated. The effectiveness of task allocation is measured based on task omission rate, duplicate assignment rate, semantic confidence distribution, and user feedback rate. Understandably, when the task omission rate is higher than a threshold, the prompts in the relevant models described in the above embodiments are adjusted to enhance the recall rate of semantic extraction; when the duplicate assignment rate is higher than a threshold, the prompts in the relevant models described in the above embodiments are adjusted to enhance the accuracy of task responsibility chain identification; when the user feedback rate is higher than a threshold, the trigger frequency of feedback loops is increased; and when the semantic confidence is too low, the frequency of manual intervention for confirmation is increased.

[0194] This application's embodiment can be applied to cross-departmental collaborative project management scenarios. A software development team continuously discusses development requirements, shares requirement documents, and assigns personnel in a group chat on an instant messaging tool, focusing on a product feature optimization project. The system first obtains historical corpus from the initial historical messages of the group chat, performs feature extraction processing on the historical messages to form feature vectors representing the group chat's goals, group rules, and member roles. Then, it performs contextual semantic analysis on these feature vectors to obtain the group's main theme information, representing the group's basic information. For example, if keywords such as "optimization," "urgent," and "product feature" appear consecutively in historical messages, combined with recent messages frequently mentioning optimization matters and the distribution of member roles such as front-end developers, testers, and project managers within the project group, the main theme information is determined to be "urgent optimization of product features."

[0195] Within the current time window, the system acquires group information, including group messages, group instructions, and group document content. Group messages may include requests for the front-end team to optimize the login function, to be completed within 24 hours. The system performs semantic analysis on these messages to determine message features related to the task intent, extracting the task execution behavior of optimizing the login function, the task execution subject (the front-end team), and the task constraint of completing within 24 hours. Group instructions may include requests such as "@Zhang Gong is responsible for front-end optimization" and "@Li Gong is responsible for testing compatibility." The system performs keyword recognition processing on these instructions to determine their features, composed of task keyword attributes, identifying the subject assignment relationship, responsible actions, and the information of the members to whom they are directed. Group document content may be uploaded requirement documents. The document summary may include requests such as "Login function needs to support multi-device login and be compatible with older versions." The system performs context feature extraction processing on this summary information to determine the group document content features, including task instructions such as supporting multi-device login and compatibility with older versions.

[0196] After obtaining the characteristics of group messages, group instructions, and group document content, the system encodes and fuses these three types of features to obtain a task information set. This task information set includes at least one task, with each task reflecting the execution elements of a single task in the form of data records. These execution elements include the task execution subject, task execution behavior, and task constraints. For example, it can generate data records corresponding to: Front-end team—optimize login function—complete within 24 hours; Engineer Zhang—responsible for front-end optimization—urgent handling; Engineer Li—test compatibility—compatible with old versions; Test role—verify multi-device login—support old versions. Subsequently, the system performs similarity processing on the task execution subject of each task to obtain a subject similarity vector, on the task execution behavior to obtain a behavior similarity vector, and on the task constraints to obtain a constraint similarity vector. Based on the subject similarity vector, behavior similarity vector, constraint similarity vector, and a preset threshold, the task information in the task information set is aggregated to form a task semantic entity that describes the set of task elements in a structured form. The aggregated task semantic entities can be grouped into a unified task semantic for urgently optimizing the login function and completing compatibility testing, which includes task elements such as front-end team, testers, optimization, testing, 24 hours, multi-device login, and compatibility with old versions.

[0197] In one possible implementation, the system performs semantic fusion processing on the group theme information and task semantic entities to generate a task list that incorporates the group theme context. This allows the task description to simultaneously address both the overall goals of the group chat and the specific execution content within the current window. For example, by fusing the urgent optimization of product features with the aforementioned task semantic entities, a task list is generated that urgently optimizes the login function, supports multi-device login, is compatible with older versions, and must be completed within 24 hours. Subsequently, feature extraction processing is performed on the task list to obtain a multi-dimensional task list. This multi-dimensional task list includes dimensions such as task content, time constraints, execution subject, and priority. The task content could be "optimize the login function and complete compatibility verification," the time constraint could be "complete within 24 hours," the execution subject could be associated with front-end development and testing roles, and the priority could be set to high priority based on the "urgent" marker in the group theme.

[0198] The system then uses a pre-defined responsibility database to match responsibilities across the multi-dimensional task list, generating a responsibility view. This database can pre-record Zhang's front-end development skills, Li's testing and compatibility verification skills, and Wang's project supervision responsibilities. Based on this database, the responsibility view maps each task to its direct responsible person, supervisor, and collaborators, or can be represented as the allocation of directly responsible users, indirectly responsible users, and collaborating users. For example, for the task of optimizing the login function, Zhang is matched as the directly responsible user, Wang as the indirectly responsible user, and Li as the collaborating user. For parts involving compatibility with older versions and multi-device login verification, Li's collaborative testing relationship and Wang's supervisory relationship can be maintained in the responsibility view.

[0199] Based on this responsibility view, the system sends a multi-dimensional task list to the directly responsible user, enabling Engineer Zhang to receive a task list containing task content, a 24-hour completion deadline, the scope of login function optimization, and multi-device login support requirements. Simultaneously, the system adds a supervisory responsibility tag to the task summaries in the multi-dimensional task list and sends them to the indirectly responsible user, enabling Manager Wang to receive a task summary with the supervisory tag. Finally, the system adds a collaborative task tag to the task summaries in the multi-dimensional task list and sends them to the collaborative responsible user, enabling Engineer Li to receive a task summary with the collaborative testing tag, which includes information such as compatibility with older versions, testing compatibility, and multi-device login verification.

[0200] Figure 4 A schematic diagram of a group task allocation device based on an instant messaging tool provided in this application embodiment. Figure 1 ,like Figure 4 As shown, the group task allocation device 40 based on an instant messaging tool provided in this embodiment includes:

[0201] The acquisition module 401 is used to acquire group theme information and group information within the current time window; wherein, the group theme information represents the basic information of the group;

[0202] Processing module 402 is used to process group information to obtain task semantic entities; process group theme information and task semantic entities to obtain a multidimensional task list; wherein, the task semantic entity represents a set of task elements extracted from group information and described in a structured form; the multidimensional task list represents a multidimensional task list containing task content, time constraints, execution subject and priority.

[0203] The sending module 403 is used to send a multi-dimensional task list to the user with the corresponding responsibility based on a preset responsibility library.

[0204] This embodiment provides a group task allocation device based on an instant messaging tool, which can execute the method provided in the above-described method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0205] Figure 5 A schematic diagram of a group task allocation device based on an instant messaging tool provided in this application embodiment. Figure 2 ,like Figure 5 As shown, the group task allocation device 50 based on an instant messaging tool provided in this embodiment includes:

[0206] The acquisition module 501 is used to acquire group theme information and group information within the current time window; wherein, the group theme information represents the basic information of the group;

[0207] The processing module 502 is used to process the group information to obtain the task semantic entity; process the group theme information and the task semantic entity to obtain the multidimensional task list; wherein, the task semantic entity represents the set of task elements extracted from the group information and described in a structured form; the multidimensional task list represents a multidimensional task list containing task content, time constraints, execution subject and priority.

[0208] The sending module 503 is used to send a multi-dimensional task list to the user with the corresponding responsibility based on a preset responsibility library.

[0209] In one possible implementation, the processing module 502 includes:

[0210] The first processing module 5021 is used to encode and fuse the group information to obtain a task information set; wherein, the task information set includes at least one task information; the task information represents data records extracted from the group information that reflect the execution elements of a single task;

[0211] The second processing module 5022 is used to aggregate the task information set to obtain the task semantic entity.

[0212] In one possible implementation, group information includes group messages, group commands, and group document content; the first processing module 5021 includes:

[0213] Semantic analysis is performed on group messages to determine group message features; keyword recognition processing is performed on group instructions to determine group instruction features; and group document content is processed to determine group document content features. Among these, group message features represent the semantic information related to task intent in group messages; group instruction features represent the task keyword attributes identified in group instructions; and group document content features represent the task context features extracted from group documents.

[0214] The characteristics of group messages, group instructions, and group document content are fused to obtain a set of task information.

[0215] In one possible implementation, the group document content is processed to determine group document content characteristics, including:

[0216] Based on the summarization algorithm, the content of group documents is processed to generate a summary information; wherein, the summary information represents the summary text that summarizes the content of group documents and contains task instruction information.

[0217] Contextual feature extraction is performed on the summary information to determine the content features of the group documents.

[0218] In one possible implementation, the task information includes the task execution subject, task execution behavior, and task constraints; the second processing module 5022 includes:

[0219] For each task information in the task information set, perform similarity processing on the task execution subject to obtain the subject similarity vector;

[0220] Perform similarity processing on the task execution behavior of each task in the task information set to obtain a behavior similarity vector;

[0221] The similarity of the task constraints for each task in the task information set is processed to obtain a constraint similarity vector.

[0222] Based on the subject similarity vector, behavior similarity vector, constraint similarity vector, and preset threshold, the task information in the task information set is aggregated to obtain the task semantic entity.

[0223] In one possible implementation, the processing module 502 includes:

[0224] The third processing module 5023 is used to perform semantic fusion processing on the group theme information and the task semantic entity to obtain a task list; wherein, the task list represents the initial task set generated after fusing the group theme context.

[0225] The fourth processing module 5024 is used to perform feature extraction processing on the task list to obtain a multi-dimensional task list.

[0226] In one possible implementation, the sending module 503 includes:

[0227] Based on a multidimensional task list and a preset responsibility library, a responsibility view is generated; the responsibility view indicates the mapping relationship between each task and the direct responsible person, supervisor, and collaborator.

[0228] Based on the responsibility view, a multi-dimensional task list is sent to the user with the corresponding responsibility.

[0229] In one possible implementation, the responsibility view indicates the directly responsible user, the indirectly responsible user, and the collaborating responsible user; based on the responsibility view, a multi-dimensional task list is sent to the user with the corresponding responsibility, including:

[0230] Send the multi-dimensional task list to the directly responsible user;

[0231] Tag the task summaries in the multidimensional task list with a supervisory responsibility label, and send the task summaries with the supervisory responsibility label to the indirectly responsible users;

[0232] Tag the task summaries in the multidimensional task list with collaborative task tags, and send the task summaries with collaborative task tags to the collaborative responsible users.

[0233] In one possible implementation, the acquisition module 501 includes:

[0234] Retrieve the initial message history of a group chat on an instant messaging tool;

[0235] Feature extraction is performed on the initial historical messages to obtain feature vectors; where the feature vectors represent the semantic feature vectors in the initial historical messages that reflect the group chat objectives, group rules and member roles;

[0236] Contextual semantic analysis is performed on the feature vectors to obtain the group theme information.

[0237] This embodiment provides a group task allocation device based on an instant messaging tool, which can execute the method provided in the above-described method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0238] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 6 As shown, the electronic device 60 provided in this embodiment includes at least one processor 601 and a memory 602. Optionally, the electronic device 60 further includes a communication component 603. The processor 601, memory 602, and communication component 603 are connected via a bus 604.

[0239] In a specific implementation, at least one processor 601 executes computer execution instructions stored in memory 602, causing at least one processor 601 to perform the above-described method.

[0240] The specific implementation process of processor 601 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0241] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0242] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0243] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0244] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0245] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0246] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0247] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0248] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0249] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0250] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0251] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0252] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0253] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A group task allocation method based on instant messaging tools, characterized in that, include: Get the group's main message; And obtain the group information within the current time window; wherein, the group theme information represents the basic information of the group; The group information is processed to obtain a task semantic entity; the group theme information and the task semantic entity are processed to obtain a multidimensional task list; wherein, the task semantic entity represents a set of task elements extracted from the group information and described in a structured form; the multidimensional task list represents a multidimensional task list containing task content, time constraints, execution subject and priority. Based on the preset responsibility library, the multidimensional task list is sent to the user with the corresponding responsibility.

2. The method according to claim 1, characterized in that, The group information is processed to obtain the task semantic entity, including: The group information is encoded and fused to obtain a task information set; wherein the task information set includes at least one task information; the task information represents a data record extracted from the group information that reflects the execution elements of a single task; The task information set is aggregated to obtain task semantic entities.

3. The method according to claim 2, characterized in that, The group information includes group messages, group commands, and group document content; The group information is encoded and fused to obtain a task information set, including: Semantic analysis is performed on the group messages to determine group message features; keyword recognition processing is performed on the group instructions to determine group instruction features; the group document content is processed to determine group document content features; wherein, the group message features represent the semantic information related to the task intent in the group messages; the group instruction features represent the task keyword attributes identified in the group instructions; and the group document content features represent the task context features extracted from the group documents. The group message features, group instruction features, and group document content features are fused together to obtain a task information set.

4. The method according to claim 3, characterized in that, The group document content is processed to determine its characteristics, including: Based on the summarization algorithm, the content of group documents is processed to generate a summary, which is a summary text containing task instruction information obtained by summarizing the content of group documents. The summary information is subjected to context feature extraction processing to determine the content features of the group documents.

5. The method according to claim 2, characterized in that, The task information includes the task execution entity, task execution behavior, and task constraints. The task information set is aggregated to obtain task semantic entities, including: Perform similarity processing on the task execution subject of each task information in the task information set to obtain a subject similarity vector; The similarity of the task execution behavior of each task in the task information set is processed to obtain a behavior similarity vector. The task constraints of each task information in the task information set are processed for similarity to obtain a constraint similarity vector. Based on the subject similarity vector, the behavior similarity vector, the constraint similarity vector, and the preset threshold, the task information in the task information set is aggregated to obtain the task semantic entity.

6. The method according to any one of claims 1-5, characterized in that, The group theme information and the task semantic entity are processed to obtain a multidimensional task list, including: The group theme information and the task semantic entity are semantically fused to obtain a task list; wherein, the task list represents the initial task set generated after fusing the group theme context. The task list is subjected to feature extraction processing to obtain the multidimensional task list.

7. The method according to any one of claims 1-5, characterized in that, Based on a pre-defined responsibility database, the multi-dimensional task list is sent to users with corresponding responsibilities, including: Based on the multidimensional task list and the preset responsibility library, a responsibility view is generated; wherein, the responsibility view indicates a mapping view of the correspondence between each task and the direct responsible person, supervisor and collaborator; Based on the aforementioned responsibility view, the multidimensional task list is sent to the user with the corresponding responsibility.

8. The method according to claim 7, characterized in that, The responsibility view indicates the directly responsible user, the indirectly responsible user, and the collaborating responsible user; Based on the aforementioned responsibility view, the multidimensional task list is sent to the user with the corresponding responsibility, including: Send the multi-dimensional task list to the directly responsible user; Add a supervisory responsibility tag to the task summary in the multidimensional task list, and send the task summary with the supervisory responsibility tag to the indirectly responsible user; Tag the task summaries in the multidimensional task list with collaborative task tags, and send the task summaries with collaborative task tags to the collaborative responsible user.

9. The method according to any one of claims 1-5, characterized in that, Obtain the group's main information, including: Retrieve the initial message history of a group chat on an instant messaging tool; The initial historical messages are subjected to feature extraction processing to obtain feature vectors; wherein, the feature vectors represent the semantic feature vectors in the initial historical messages that reflect the group chat objectives, group rules and member roles; The feature vector is subjected to contextual semantic analysis to obtain the group theme information.

10. A group task allocation device based on an instant messaging tool, characterized in that, include: The acquisition module is used to retrieve group topic information; And obtain the group information within the current time window; wherein, the group theme information represents the basic information of the group; The processing module is used to process the group information to obtain a task semantic entity; and to process the group theme information and the task semantic entity to obtain a multidimensional task list; wherein, the task semantic entity represents a set of task elements extracted from the group information and described in a structured form; and the multidimensional task list represents a multidimensional task list containing task content, time constraints, execution subject, and priority. The sending module is used to send the multi-dimensional task list to the user with the corresponding responsibility based on a preset responsibility library.

11. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-9.

13. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-9.

Images